POPULARITY
Categories
157 episodes with preserving
71 episodes with preserving
25 episodes with preserving
21 episodes with preserving
21 episodes with preserving
10 episodes with preserving
24 episodes with preserving
8 episodes with preserving
13 episodes with preserving
22 episodes with preserving
11 episodes with preserving
9 episodes with preserving
7 episodes with preserving
10 episodes with preserving
8 episodes with preserving
9 episodes with preserving
12 episodes with preserving
5 episodes with preserving
16 episodes with preserving
7 episodes with preserving
5 episodes with preserving
9 episodes with preserving
6 episodes with preserving
8 episodes with preserving
5 episodes with preserving
5 episodes with preserving
7 episodes with preserving
8 episodes with preserving
8 episodes with preserving
7 episodes with preserving
4 episodes with preserving
9 episodes with preserving
5 episodes with preserving
9 episodes with preserving
7 episodes with preserving
9 episodes with preserving
7 episodes with preserving
9 episodes with preserving
3 episodes with preserving
3 episodes with preserving
7 episodes with preserving
5 episodes with preserving
6 episodes with preserving
4 episodes with preserving
6 episodes with preserving
6 episodes with preserving
5 episodes with preserving
11 episodes with preserving
4 episodes with preserving
4 episodes with preserving
4 episodes with preserving
3 episodes with preserving
5 episodes with preserving
7 episodes with preserving
4 episodes with preserving
4 episodes with preserving
3 episodes with preserving
7 episodes with preserving
6 episodes with preserving
3 episodes with preserving
3 episodes with preserving
6 episodes with preserving
6 episodes with preserving
6 episodes with preserving
4 episodes with preserving
5 episodes with preserving
4 episodes with preserving
4 episodes with preserving
3 episodes with preserving
3 episodes with preserving
5 episodes with preserving
2 episodes with preserving
2 episodes with preserving
3 episodes with preserving
5 episodes with preserving
4 episodes with preserving
9 episodes with preserving
4 episodes with preserving
2 episodes with preserving
3 episodes with preserving
3 episodes with preserving
3 episodes with preserving
3 episodes with preserving
4 episodes with preserving
5 episodes with preserving
4 episodes with preserving
5 episodes with preserving
3 episodes with preserving
2 episodes with preserving
5 episodes with preserving
3 episodes with preserving
4 episodes with preserving
5 episodes with preserving
4 episodes with preserving
3 episodes with preserving
6 episodes with preserving
6 episodes with preserving
3 episodes with preserving
5 episodes with preserving
4 episodes with preserving
4 episodes with preserving
5 episodes with preserving
2 episodes with preserving
5 episodes with preserving
2 episodes with preserving
6 episodes with preserving
2 episodes with preserving
3 episodes with preserving
4 episodes with preserving
2 episodes with preserving
6 episodes with preserving
3 episodes with preserving
4 episodes with preserving
3 episodes with preserving
3 episodes with preserving
5 episodes with preserving
3 episodes with preserving
4 episodes with preserving
4 episodes with preserving
2 episodes with preserving
4 episodes with preserving
4 episodes with preserving
2 episodes with preserving
3 episodes with preserving
4 episodes with preserving
4 episodes with preserving
Preserving culture requires more than celebrating it. The traditions also need customers, economic opportunity, and a path to the next generation. Suleiny Altamirano, founder and CEO of Tizana Mexicana, shares how moving from Guerrero, Mexico, to Seattle inspired her to build a bridge between Mexican artisans and new markets. She explains why trust matters more than technology, how immigrant founders can build a new support system, and why social impact companies still need a sustainable business model. We also discuss: • Why immigrant founders need to create a new family of mentors, supporters, and collaborators • How Tizana Mexicana works with approximately 175 artisans • Why artisan products cannot compete on price alone • How cultural traditions disappear when creators cannot earn a living • Building trust across language, banking, technology, and communication barriers • Why founders should talk about their ideas before they feel ready • Balancing social impact with profitability • Expanding beyond Washington and bringing more artisans to Seattle • The role of community in entrepreneurship and leadership Suleiny' links" LinkedIn: https://www.linkedin.com/in/suly-altamirano Website: https://www.tizanamexicana.com Instagram: @tizanamexicana Jason's CTA: Subscribe to The Jason Cavness Experience on YouTube for more conversations with founders, operators, and builders turning personal experience into meaningful companies. Jason's links: https://linktr.ee/jasoncavness
Explore how asset segregation, initial strategies, and land trusts make real estate holdings virtually uncollectible for litigators. This segment dives into handling default judgments, corporate veil piercing, and passing real estate and business assets seamlessly down to the next generation without going through probate.
In this episode, we unpack why high-stress seasons, like back-to-school transitions, often trick us into mistaking temporary exhaustion for relationship friction. Instead of getting stuck in a battle of me vs. you or striving for perfection, we explore how to shift back to us vs. the chaos. Join us for practical, realistic ways to protect your bond so your relationship isn't another item on your stress list, rather it's your safe place to land.Resources: The Gottman InstituteWe are so grateful for your support! Please share this podcast with someone who needs it and leave us review: https://podcasts.apple.com/us/podcast/positive-on-purpose/id1531548022
(This episode originally aired on February 10, 2026. Dr. Omar Lateef has since announced his intention to retire from Rush Health in June 2027.) As financial pressure mounts and the healthcare safety net continues to strain, academic medical centers are drawing on their culture of innovation to pursue better outcomes — and narrow life expectancy gaps in the communities they serve. In this episode of Radio Advisory, host Rae Woods sits down with leaders from Rush University Medical Center to explore how health systems can stay focused on results amid tightening margins, political scrutiny, and ongoing uncertainty in grant funding. Dr. Omar Lateef, President and CEO of Rush, and Dr. David Ansell, Senior Vice President for Community Health Equity, describe how Rush treats gap-reduction as a long-term operating strategy rather than a moral or messaging exercise. They share practical examples of how local partnerships, targeted investments, and day-to-day operational choices can improve outcomes while still making financial sense — and why avoiding battles over language helps keep the focus squarely on results. We're here to help: How Rush University Medical Center is addressing the root causes of social determinants of health 264: Research funding is being slashed. What's the real industry impact? How research funding cuts are impacting healthcare (and how to respond) 12 things CEOs need to know in 2026 Tool: How policy changes will impact your bottom line Who gets the chance to be healthy? | Rush The Rush Center for Community Well-Being at Sankofa Wellness Village | Rush Health disparities in Chicago and the work to solve them with Rush University Medical Center (AMA) The Anchor Strategy — A Place-Based Business Approach for Health Equity | New England Journal of Medicine Rush Signs on as First Partner for Local Laundry Service | Rush David Ansell Books – New: The Death Gap Sponsor link: Preserving iodine, protecting imaging
Summer Rewind EpisodeIn this deeply personal and politically urgent conversation, Dr. David Johns welcomes his White House family—Deesha Dyer, former Social Secretary to President Barack Obama and First Lady Michelle Obama—for a raw discussion about what we lose when power demolishes history, and what endures when community builds with intention.Deesha takes us inside the moment she saw excavators tearing down the East Wing of the White House in October 2025—the space where she orchestrated state dinners, opened doors for people who never thought they'd walk through them, and made the People's House truly belong to the people.This conversation moves from the rubble of demolished institutions to the unshakeable foundation of community organizing. Deesha shares lessons learned from Michelle Obama about grace under pressure, why she wrote Undiplomatic for Black girls who "risk it all," and what defending democracy actually looks like in practice: investing in local organizations, letting leaders lead in their expertise, documenting our work, and showing up for each other with intention.Deesha's story is a masterclass in refusing to let imposter syndrome—or anyone else—define your worth. As she reminds us: buildings can be demolished in 72 hours, but they can't demolish the memories, the legacy, or the collective power we build together.This is required listening for anyone who believes that mutual aid isn't charity—it's how we survive, build, and love each other into liberation.Website: DeeshaDyer.comInstagram: @deedyer267Email: deesha@deeshadyer.com (She answers every email!)Book: Undiplomatic: How My Attitude Created the Best Kind of Trouble(Available at libraries and independent bookstores)Become a supporter of this podcast: https://www.spreaker.com/podcast/teach-the-babies-w-dr-david-j-johns--6173854/support.
I'm excited to share another episode in our series of highly vetted companies. Today, I'm delighted to reconnect with Angelo Keely, the Co-founder and CEO of Kion, for our third conversation. I've been using essential aminos. I really leaned into them during my book launch, and I credit them with helping me maintain my muscle mass despite training at a lower volume. In today's conversation, Angelo and I discuss GLP-1s and muscle loss, exploring protein intake, the leucine threshold hypothesis, the association between declining estrogen and blunted muscle protein synthesis, and using essential aminos to build and maintain muscle. We unpack the role of sleep and recovery for muscle health, daily versus as-needed creatine monohydrate and essential amino acid supplementation, what omega-3s are, why they matter, and how to supplement. We also cover midlife women's body composition, weight loss, and muscle building, and Angelo shares his take on the disconnect between body composition changes in midlife women and how companies actively market to midlife women, knowing the challenges we face in that area. Stay tuned for today's insightful and informative discussion with Angelo Keely of Kion. IN THIS EPISODE, YOU WILL LEARN: How caloric restriction leads to fat loss but can also cause the body to break down muscle How essential aminos can help preserve our muscles as we age The essential amino formula Angelo recommends for midlife women Hormone changes may affect muscle protein synthesis The huge impact poor sleep has on whole-body protein synthesis Essential aminos vs. whey protein and how they differ Value of essential aminos for maintaining muscle during caloric deficits How aminos and creatine both need to be taken consistently, but for different physiological reasons What to look for when choosing a quality omega-3 supplement How to make daily supplementation easier to maintain Connect with Cynthia Thurlow Follow on X, Instagram & LinkedIn Check out Cynthia's website. Submit your questions to support@cynthiathurlow.com Join other like-minded women in a supportive, nurturing community: The Midlife Pause/Cynthia Thurlow. Purchase Cynthia's book, The Menopause Gut. Cynthia's Intermittent Fasting Transformation Book The Midlife Pause Supplement Line Connect with Angelo Keely Kion (Use code CYNTHIA for 20% off)
Last time we spoke about the reaction after the surprise attack on Pearl Harbor. After four years of fighting Japan almost alone, China's Chiang Kai-shek learned of Pearl Harbor with profound relief. Japan's attack, meant to neutralize American naval power, instead galvanized the entire Western world against Tokyo. Within hours, Chiang declared war on the Axis powers, transforming China from an isolated nation to a recognized Allied member. The strategic mathematics were clear to Chiang: America's industrial might would ultimately crush Japan. China simply needed to survive long enough. The United States quickly formalized this partnership, pledging military support through Lend-Lease and elevating Chiang to leadership of Allied forces in the China theater. Yet Pearl Harbor brought catastrophe for Western civilians in China. Americans and Britons, previously protected as neutral foreigners, were suddenly rounded up and interned in camps. For many, decades of presence in China ended in hasty evacuation, leaving behind lives built over generations. #216 The Battle of Hong Kong Part 1: The Approaching Storm Welcome to the Fall and Rise of China Podcast, I am your dutiful host Craig Watson. But, before we start I want to also remind you this podcast is only made possible through the efforts of Kings and Generals over at Youtube. Perhaps you want to learn more about the history of Asia? Kings and Generals have an assortment of episodes on history of asia and much more so go give them a look over on Youtube. So please subscribe to Kings and Generals over at Youtube and to continue helping us produce this content please check out www.patreon.com/kingsandgenerals. If you are still hungry for some more history related content, over on my channel, the Pacific War Channel where I cover the history of China and Japan from the 19th century until the end of the Pacific War. Japan emerged as a significant threat to British interests during the 1920s following the collapse of the Anglo-Japanese Alliance. This concern intensified dramatically after Japan's invasion of China, with the ensuing Sino-Japanese War escalating tensions further. British assets in the region came under direct assault, including attacks on Western military and commercial vessels such as HMS Ladybird. Yet despite these provocations, the British Government maintained an inconsistent stance toward Japan, responding with only feeble diplomatic and military countermeasures. This contradiction characterized London's approach: while diplomatic channels remained conciliatory, military planners viewed Japan with genuine apprehension. During the 1930s, Singapore and Malaya represented Britain's most valuable and prosperous holdings in the Far East and consequently received substantial garrison reinforcements. By contrast, Hong Kong was widely acknowledged to be indefensible—a conclusion reached repeatedly by successive military studies. Despite this consensus, the British Government proceeded with the Hong Kong Defence Scheme in 1936, which aimed to transform Hong Kong Island into a fortified redoubt while establishing delaying positions in Kowloon and the New Territories. The Hong Kong Garrison's mandate was straightforward: repel external attacks and prevent enemy use of the harbour and dry dock. The underlying strategy relied on a simple principle: hold Hong Kong Island as an impregnable fortress and await relief from the Royal Navy. Consequently, protecting the harbour became the paramount objective. Japanese strategists were correctly identified as Hong Kong's primary threat, with military planners anticipating attack from the sea rather than overland. To counter such an offensive, the defence scheme incorporated five successive layers of protection: naval mines fitted with indicator nets and anti-submarine booms formed the first barrier; coastal artillery provided the second; beach installations featuring pillboxes, wire entanglements and mines constituted the third; infantry strongpoints positioned on commanding heights guarded against amphibious landings; and finally, a mobile reserve force stood ready to counterattack and eliminate any successful enemy incursion. The 1936 plan also mandated extensive fortification of Kowloon's bottleneck approaches through bunkers, trenches and blocked passages, with a principal defensive line—resembling a miniature Maginot Line—constructed across the mountains immediately north of Kowloon. The garrison structure comprised three infantry battalions plus the Hong Kong Volunteer Defence Corps (HKVDC), a unit approximately regiment-strength equipped with armoured cars, engineers, artillery, anti-aircraft guns and infantry. These forces organized into three brigades: mainland, island, and garrison reserve. The air component was planned to be formidable: a reconnaissance squadron, a fighter-bomber squadron, two torpedo squadrons and a flight of auxiliary spotting aircraft operated by the HKVDC. Naval forces consisted of a modest destroyer contingent, an MTB flotilla, gunboats and patrol vessels. When Japan occupied Guangzhou, Hong Kong found itself effectively encircled. In response, limited conscription was implemented, though restricted to non-Chinese males for HKVDC service. Chinese personnel were largely confined to non-combat positions such as air-raid precautions and medical duties. The expanding Japanese threat prompted a substantial expansion of the Hong Kong Garrison to four regular infantry regiments—two British units (one specializing in medium machine-gun support) and two Indian regiments. Supporting forces included two coastal artillery regiments, a field artillery regiment-plus, and an anti-aircraft regiment. The Royal Navy expanded to include the 2nd MTB flotilla (eight 60-foot, 29-knot motor torpedo boats armed with twin torpedoes and eight Lewis guns each, plus two smaller 55-foot vessels), four shallow-draught river gunboats (HMS Cicala, Tern, Moth and Robin), four aging World War I S-class destroyers (HMS Thracian, Thanet, Scout and Tenedos—the last transferred to Singapore in 1939), HMS Cornflower sloop, and various auxiliary vessels including tugs, boom defence ships and minelayers. The Fleet Air Arm deployed three Supermarine Walruses—single-engine reconnaissance amphibians—at Kai Tak Airport. The Royal Air Force remained the weakest service, equipped with only four Vickers Wildebeest torpedo-bombers and a handful of HKVDC aircraft: one Avro Tutor, two Hornet Moths and two Cadet biplanes. When World War II erupted in Europe, Britain's focus necessarily shifted to national survival, inevitably leaving distant imperial outposts like Hong Kong neglected. Rather than receive reinforcements, Hong Kong actually lost capable personnel; many veteran troops were repatriated to Europe, leaving behind a less proficient garrison. Peacetime duties had proven a poor preparation for combat—officers and men had passed their time socializing rather than maintaining combat readiness. Governor Sir Geoffrey Northcote persistently urged London to adopt a firmer stance against relentless Japanese provocations: aircraft incursions, sinking of junks and fishing vessels, and infiltration by Taiwanese fifth columnists (Japan had colonized Taiwan in 1895). Recognizing the hopelessness of his position, Northcote recommended withdrawal in October 1940, arguing that defending the colony would inevitably result in mass civilian casualties and widespread destruction—a pragmatic assessment given Hong Kong's indefensible state. London categorically rejected this proposal, reasoning that withdrawal would demoralize China, embolden Japan and undermine American confidence in British resolve. Underlying this decision was an additional consideration: the long-standing dispute over Hong Kong's sovereignty. Since Britain acquired Hong Kong in 1841, the prospect of surrendering it, even temporarily, raised fears about reclaiming it afterward. Preserving the appearance of British determination was deemed crucial to demonstrating to Chiang Kai-shek, Nationalist China's leader, that Britain intended to retain the territory. Despite these political calculations, London consistently refused fresh troop deployments. Major-General A. E. Grasett, the China theatre commander, requested reinforcements only to face rejection—London offered only Indian or other colonial forces. Air Chief Marshal Sir Robert Brooke-Popham, appointed British commander in the Far East in 1941, likewise pressed for garrison strengthening, to no avail. The situation shifted dramatically after Grasett, a retired Canadian who had recently left his China post, travelled to Britain via Canada. Upon arriving in Canada, he independently approached his former classmate Major-General Crerar, Chief of the Canadian General Staff, and advocated for reinforcement. Grasett argued persuasively that deploying two or more additional battalions would enable the Hong Kong garrison to withstand an extended siege. When Grasett reached London in early September 1941, he presented this case directly to the British Chiefs of Staff, simultaneously proposing that Canada could provide the extra units. His arguments proved persuasive; the Chiefs of Staff altered course and convinced a reluctant Churchill to reconsider. In light of accelerating Japanese pressure, Britain substantially enlarged the Malaysia/Singapore garrison from nine to 32 battalions and despatched two capital ships—HMS Prince of Wales and Repulse—to the region, reasoning that this show of force would deter Japanese aggression against British Far Eastern interests, including Hong Kong. The original defence strategy envisaged holding Hong Kong Island exclusively while denying the harbour to attackers. A single battalion (the Punjabis) would execute delaying actions across the New Territories, followed by a three-battalion defensive stand on the Gin Drinkers Line. This brief mainland resistance would allow time for demolishing stores, power installations, docks and wharves, clearing food stocks and removing vital matériel before withdrawing to Hong Kong Island for a protracted resistance awaiting relief from Singapore or American forces from the Philippines. The Gin Drinkers Line—styled after the French Maginot Line in miniature—ran approximately 18 kilometres south of the Sino-Hong Kong border. Beginning near Gin Drinkers Bay west of Kowloon, it extended 17 kilometres across the Kowloon Peninsula hills toward its eastern extremity. The strategic expectation was that this line would delay the Japanese assault for three weeks or longer, permitting battalion withdrawal to Hong Kong Island for sustained resistance while mainland forces completed evacuation procedures. The 1937 war plan stipulated that four battalions would be necessary for an effective delaying action and seven for an adequate Gin Drinkers Line defence—figures which excluded the aircraft and naval support essential to success. With only a single army brigade available and virtually no air capability, planners abandoned the mainland defence concept, leaving defensive works incomplete and deteriorating under tropical conditions. Chinese personnel were conspicuously absent from initial war preparations and strategic planning, only being incorporated late in the process. The largest Chinese component was the volunteer contingent serving with the HKVDC. On the regular army side, approximately 150 Chinese recruits were enlisted in mid-1941 for the 5th Anti-Aircraft Regiment. The Hong Kong Chinese Regiment, hastily constituted on 3 November 1941, comprised officers borrowed from British and Indian units alongside 52 Chinese junior NCOs and other ranks with minimal training. A Royal Engineers troop staffed entirely by Chinese soldiers saw service, and some Chinese personnel served with the Royal Navy and Dockyard Defence Corps. Chinese participation increased substantially in auxiliary services—air-raid precautions, auxiliary police, nursing, St John Ambulance and fire brigades. This reluctance to fully integrate Chinese forces likely reflected prevailing attitudes of the era. Pre-war British propaganda had frequently mocked Japanese military capability, portraying officers and soldiers as myopic and incompetent. Japanese combat prowess was systematically underestimated; their victories in China were attributed to facing inferior opposition rather than evidence of genuine military strength. Japanese weaponry was dismissed as inferior to European standards. Consequently, British confidence remained high that the colony could hold indefinitely if adequately reinforced. In September 1941, Hong Kong received heartening news: two additional Canadian battalions would be deployed to the garrison. Brigadier John Maltby promptly revived the pre-war plan for defending the Gin Drinkers Line, developing a strategy centred on three mainland battalions (2nd Royal Scots, 5/7 Rajputs and 2/14 Punjab) supported by HKVDC elements and four artillery troop detachments, while three island battalions (1st Middlesex, Royal Rifles of Canada and Winnipeg Grenadiers) bolstered by the bulk of the HKVDC would hold Hong Kong Island. Command responsibility was divided between two brigades: the Island Brigade (Hong Kong Infantry Brigade) under Brigadier Lawson, a recently arrived Canadian, and the mainland Kowloon Infantry Brigade under Brigadier Wallis. The Canadian regiments arrived on 16 November 1941. Maltby was commissioned into the Indian Army in 1911, and saw extensive service across multiple campaigns: World War I, the Persian Gulf (1913–14), and the North-West Frontier (1923–24). His early promise earned him selection for accelerated advancement through prestigious staff colleges at Quetta and RAF Andover. By 1937, he was again serving on the North-West Frontier, and by 1939 commanded both the 3rd Jhelum Brigade (later the Calcutta Brigade) and the 19th Infantry Brigade in Deccan. In 1940, posted to China, Maltby orchestrated the closure of the North China Command by withdrawing two infantry battalions from Shanghai. Promoted to major-general as GOC China in August 1941, he had only three months to prepare the colony's defenses, a race against time he could not win. The Gin Drinkers Line occupied a naturally dominant position atop mountains spanning northern Kowloon's breadth, yet possessed inherent weaknesses that would prove fatal. The line lacked defensive depth and remained vulnerable to flanking manoeuvres, with two sectors representing particular danger: Customs Pass and the gap between Golden Hill and Laichikok Peninsula near Gin Drinkers Bay. Given the extended front, each battalion arranged its forces in a string of platoon positions with gaps covered by daytime fire and nighttime patrols. One company per battalion could be held in reserve within prepared positions guarding the most dangerous sectors. The reserve company of the centre battalion (2/14 Punjab) initially advanced as forward troops, screening demolition operations and delaying the enemy's initial thrust. The hasty decision to establish mainland defences meant that essential support systems—communication networks, artillery registrations, mortar calibrations—remained incomplete when combat commenced. Ammunition shortages compounded the problem. For instance, the 2/14 Punjab conducted only a single 3-inch mortar practice session before the attack; adequate ammunition arrived only in November and totalled a mere 70 rounds per battalion for both training and combat. These mortars were registered for the first time in their battle positions, with combat beginning just 12 hours later. The 2-inch mortar situation was even more desperate: troops received live ammunition only as fighting started, making their baptism of fire an actual combat situation. Transportation deficiencies—trucks and animals in short supply—forced manual ammunition transport across Hong Kong's challenging terrain, imposing severe physical demands on already stretched personnel. On 6 November 1941, Imperial Japanese Headquarters ordered its Commander-in-Chief in China to prepare plans for capturing Hong Kong. The 23rd Army, part of the China Southern Expeditionary Army Group and commanded by Lt. Gen. Sakai Takashi with Maj. Gen. Kuribayashi Tadamichi serving as Chief of Staff, would form the invasion's backbone. All preparations were to be completed by the end of November 1941. The operation, designated 'Operation C', sought to 'seize Hong Kong within ten days'. Takahashi was the son of a factory worker, and rose through the ranks to become a veteran of China campaigns. He was tasked with capturing Hong Kong using the 38th Division normally assigned to the Southern Expeditionary Army Group. Despite setbacks and humiliation over requests for additional resources—particularly following intelligence reports of tanks (actually Bren carriers) in Hong Kong—Sakai remained determined to succeed. He served as Governor of Hong Kong from 26 December 1941 to 20 February 1942. Kuribayashi, best remembered as the Japanese commander at Iwo Jima, graduated from the Army Academy in 1914 specializing in cavalry. He served as deputy military attaché to Washington and Canada and studied at Harvard University. In December 1941, he held the position of Chief of Staff to the 23rd Army during the Hong Kong operation. Operation C followed a straightforward strategic framework. A blocking force would prevent Chinese interference from the rear, while the main invasion force would spearhead the assault. The 23rd Army, comprising four divisions plus a mixed brigade and two infantry regiments, received the assignment to attack Hong Kong. Of these four divisions, only the 38th played a substantive role in the invasion itself. The attack employed a combined-arms approach. Naval forces would establish a blockade and deliver bombardment while aircraft targeted key installations. Ground forces from the 38th Division—specifically the 228th, 229th, and 230th Regiments—would execute the land invasion using a three-pronged strategy. The right column would sweep westward in a wide arc, clearing Castle Peak Road and targeting the western flank of the Gin Drinkers Line near Laichikok Peninsula. The centre force would press directly through the line's middle, while the left column would cross Tidal Cove (Tolo Harbour) toward Kai Tak Airport and eastward toward Devil Peak, a crucial defensive position. The objective was to annihilate British forces on the mainland and compel Hong Kong's surrender through sustained bombardment. Should British forces resist, the island phase of the campaign would commence. Securing the rear against the Chinese National Army of the 7th Military District and Communist guerrillas fell to the 66th Regiment, normally part of the 51st Division but attached to the 38th for this operation. The 66th would simultaneously occupy and secure rear areas throughout the New Territories and Kowloon while the three assault regiments engaged the main defensive forces. This rearguard operation was organised into columns named after their commanding officers: Kitazawa, Kobayashi, Sato, and Araki. The Japanese selected the narrowest approach between Kowloon and Hong Kong Island, targeting the eastern coast near Lyemun Passage where only 410 metres of water separated the mainland from the island. Upon landing, forces would rapidly advance inland to seize the critical junction of Wongneichong Gap, then pivot westward and southward to capture Victoria City and the remainder of Hong Kong. The naval contingent divided into two groups: the Bombardment Group and the Attack Group. Approximately 300 personnel from the Special Naval Landing Force (SNLF) were attached to the invasion force. Air support came from the First Air Brigade under Col. Habu Hideharu, which contributed 56 light bombers and fighters. Their mission was to neutralise the Royal Air Force and Royal Navy, thereby securing aerial and maritime superiority for the ground campaign. On the eve of war, the British garrison numbered 13,981 all ranks, including nursing staff and St John Ambulance personnel. Of these, 8,919 were British and Canadian, with 4,402 Indian and Chinese. The core infantry force comprised 5,422 regular soldiers, complemented by approximately 6,000 additional personnel across artillery, engineering, naval, and Hong Kong Volunteer Defence Corps (HKVDC) units. By 1941, the Royal Navy had become a shadow of its former self as the once-dominant China Squadron. Of the three World War I destroyers present on 8 December, two had been sent to Singapore—the third undergoing repairs—by 1930. This redeployment stemmed partly from the futility of defending the colony and partly from a commitment by senior US Navy officers: Britain would reinforce Singapore with Hong Kong's fleet in exchange for four American destroyers at Singapore. This American guarantee, however, never materialized. The defensive infrastructure focused primarily on countering a seaborne attack: twenty-nine coastal guns (9.2in., 6in., 4.7in., and 4in. calibres), ten 18- and 2-pounder beach defense guns, and twenty-eight field guns (60-pounder, 6in., 4.5in., and 3.7in. calibres). The Hong Kong Garrison's infantry comprised four regular army battalions—two British and two Indian—bolstered by five HKVDC infantry companies, each averaging 100 men. The 1st Middlesex Regiment, deployed in August 1937, served as a machine-gun battalion equipped with forty-eight Vickers medium machine guns. Having been abroad since 1931, the regiment had deteriorated through protracted garrison service, Hong Kong being among the softest postings available. From September 1939 onward, all regiments—including the 5/7 Rajputs (arrived June 1937) and the 2/14 Punjab—experienced systematic loss of experienced officers and NCOs to other theaters. The Royal Scots proved particularly hard hit, suffering disciplinary problems and an unusually high court-martial rate. Units like the Hong Kong and Singapore Royal Artillery suffered critical leadership gaps that went unaddressed. By December 1941, many formations fielded inexperienced replacement troops, significantly undermining combat readiness. The Punjabis exemplified these deficiencies. Stationed in Hong Kong since November 1940, they received their vehicles and mortars only in August 1941—three months before combat. Compounding this, forty percent of their strength consisted of raw recruits arrived just in October. When war erupted, artillery regiments found themselves relying on junior officers in senior roles, with British commanders unfamiliar with Urdu directing Indian troops, a handicap that severely compromised effectiveness. Deficient equipment compounded leadership shortages. Shells dating from 1918 detonated prematurely upon discharge, while ammunition shortages crippled operations. On 8 December, all field guns deployed with merely 100 rounds apiece, with 200 rounds held in reserve for the entire campaign, and coastal guns received only 25 rounds each. Insufficient cross-training meant many infantrymen failed to understand the artillery "clock face" targeting system, causing them to miss their objectives. Malaria devastated European troops, particularly the Royal Scots. Of approximately 770 personnel, at least 180 suffered recurring malarial episodes from the Shing Mun Redoubt area's notoriously endemic conditions. Staff shortages forced nearly every unit to promote junior officers to positions beyond their rank, exacerbating structural weaknesses. Although the Canadians' arrival bolstered the garrison, significant handicaps undermined their effectiveness. Both battalions carried a C-class designation—deemed unfit for foreign service—having spent years in soft garrison postings without rigorous training. The deficit was stark: each Canadian soldier received only thirty-five rifle practice rounds. The two battalions incorporated 115 men with fewer than sixteen weeks' training, while sixty-two of seventy-five recruits joining the Winnipeg Grenadiers in June came directly from a training depot lacking even basic rifles. Twelve percent of the Canadian force remained untrained. Equipment deficiencies proved even more damaging. A standard 1941 Canadian infantry battalion required 102 motor vehicles—including twenty-two Universal Carriers, thirty-seven ¾-ton trucks, thirteen 1-ton trucks, thirty-one bicycles, and twenty-two Boys anti-tank rifles. When the Canadians landed, logistical failures left them with only six Bren carriers, two water tankers, twelve ¾-ton trucks, and a single Boys anti-tank rifle. Although they carried 3in. mortars, neither ammunition nor radios accompanied them. Canada possessed only 300 3-inch mortar bombs in its entire inventory at that time, making any allocation to Hong Kong impossible. A second vessel, the SS Don Jose, was dispatched with replacement vehicles and equipment but never reached Hong Kong before 8 December. The Japanese seized it en route to Manila. Despite these shortcomings, the force departed on 27 October with 1,975 men and arrived on 16 November. Intelligence estimates regarding Japanese available forces varied widely, ranging from two to four divisions. On 6 December evening, Chinese reports confirmed the previous day's arrival of three Japanese divisions at Buji township, located eight miles from the border in what is now Shenzhen. Earlier, around 1 November, a Japanese deserter provided intelligence of significant troop concentrations north of the frontier, including large-calibre artillery indicative of at least divisional strength. The Hong Kong invasion force comprised three service branches: the Imperial Japanese Army (IJA) deployed the 38th Infantry Division, the Imperial Japanese Navy (IJN) committed the Second China Fleet, and the Imperial Japanese Air Force (IJAF) contributed the 1st Air Group. Command of the ground assault rested with Lieutenant-General Sano Tadayoshi of the 38th Division, supported by Major-General Ito Takeo heading the infantry group, while supporting arms including artillery, engineers, armor, and transport fell under Sano's authority. The main assault force consisted of three infantry regiments from the 38th Division, each led by a colonel: the 228th under Doi Teihichi, the 229th under Tanaka Ryosaburo, and the 230th under Shoji Toshishige. The 38th Division itself was a standard 1940s B-class infantry division of approximately 20,000 troops (ranging from 18,000–21,000 depending on operational requirements), organized into three regiments with three battalions each. Reflecting the Japanese army's reliance on animal and human transport, each regiment fielded roughly 3,800 troops supported by some 700 horses. Battalion strength averaged around 1,071 personnel, with companies numbering approximately 180 men. Japanese infantry battalions of this period featured a distinctive organization: rifle platoons consisted of four sections (three equipped with rifles and light machine guns, one with light mortars and rifles), complemented by an additional heavy machine-gun company fielding twelve Type 92 weapons and a troop of light artillery—likely Type 92 70mm infantry guns. The divisional field artillery regiment operated 75mm Type 41 mountain guns organized into three 688-man battalions, each with two troops of two guns. Mountain artillery units, numbering 36 guns total, carried additional personnel (3,400) and animals (1,400) to navigate the terrain complexities of Hong Kong operations. Recognizing the need for overwhelming firepower to compel Hong Kong's surrender, the IJA augmented the invasion force with heavy artillery—15cm and 24cm howitzers commanded by Major-General Kitashima Kineo. News of the Canadian reinforcements arriving in Hong Kong forced a reassessment. Commander-in-Chief Sakai Takashi of the 23rd Army informed his superiors that the original ten-day timeline was unrealistic and requested additional troops. Although most formations were already committed to other operations, Imperial Headquarters eventually authorized two depot regiments from Japan to support the campaign. Intended for second-line duties rather than frontline combat, these units were to be returned promptly after freeing experienced personnel for Hong Kong operations. This arrangement allowed the 19th and 20th mixed brigades—approximately 6,000 men organized in five infantry battalions, a field artillery battalion, and an engineer battalion—to be released for Hong Kong. However, Sakai's request for reinforcements met with disapproval from Imperial Headquarters, and he received an unmistakable message: a swift, decisive campaign was essential. His situation worsened when Japanese agents mistakenly identified British Bren carriers as tanks, forcing Sakai to improvise armor by raiding the 1st Reserve Tank Battalion and tankette units from infantry depot divisions. In attempting to secure additional armor from the 11th Army's 9th Reconnaissance Regiment, Sakai encountered resistance. General Hata Shunroku lodged sustained complaints with senior officials, and Sakai received a sharp rebuke—further requests would not be entertained, and his replacement would follow if necessary. Stung by this loss of face, Sakai abandoned all additional requests and became even more determined to achieve success. Vice-Admiral Niimi Masaichi, commanding the 2nd China Expeditionary Fleet, oversaw the entire naval element, divided into bombardment and attack groups. His flagship, the light cruiser Isuzu, carried Captain Ura Koichi, who simultaneously commanded the bombardment group. This naval force assembled ships from various China Expeditionary Fleet units, particularly from the Canton Special Base Force, 15 Squadron, and 11 Torpedo Boat Division at Guangzhou. Reflecting Niimi's lack of confidence in the army's air support capabilities, the seaplane tender IJN Kamikawa Maru joined the operation on his insistence, providing organic naval aviation against the threat posed by British Wildebeests. Niimi also petitioned for destroyers IJN Wakaba and Yugure to counter the three British destroyers, though this request was denied. The Special Naval Landing Forces (SNLF) were assigned to conduct diversionary attacks south of Hong Kong Island, intended to distract British forces from the main land invasion thrust. The IJAF's 1st Air Brigade, commanded by Colonel Habu Hideharu, centered on the 45th Air Regiment with 34 Kawasaki Ki-32 Type 98 'Mary' light bombers. This composite force drew aircraft and units from airfields across China and Manchuria—Beijing, Shanghai, Taiwan, and Qiqihar—all assembled at Baiyun Aerodrome in Guangzhou on 7 December 1941. Supporting the 1st Air Brigade were the 10th Independent Squadron under Captain Takatsuki Akira with 13 Nakajima Ki-27 Type 97 'Nate' fighters, 3 Mitsubishi Ki-15 Type 97 'Babs' command reconnaissance aircraft from the 18th Independent Reconnaissance Squadron, and the 44th Independent Squadron fielding 6 Tachikawa Ki-36 Type 98 'Ida' observation aircraft. Years before the invasion, the Japanese cultivated an extensive network of agents and sleeper operatives in Hong Kong, methodically gathering intelligence and mapping military installations. This sophisticated network proved remarkably effective: invasion maps carried by IJA officers were actually Hong Kong Government and British military maps, overprinted with Japanese annotations and interpretations. British officers later discovered in captivity that the Peninsula Hotel's regular barber was a Japanese Navy lieutenant-commander who had exploited the intimate setting to extract valuable information about British force dispositions. Intelligence operatives infiltrated every level of Hong Kong society—the garrison tailor and barber both served as spies. This groundwork enabled exceptional accuracy in artillery fire and bombing campaigns. The Japanese further leveraged criminal networks, recruiting gangsters throughout Hong Kong and Guangdong province with money and arms, turning them into irregular auxiliaries supporting the invasion. On Sunday, 7 December, seven hundred members of the 2nd Royal Scots and Middlesex Regiment marched to St John's Cathedral for the morning service. As Major-General Maltby read a passage from Matthew, his aide-de-camp, Lieutenant I. MacGregor, delivered an urgent message that would change everything. Preparations for conflict had already begun. On 5 December, the HKVDC had mobilized fully, yet Hong Kong's broader population remained unmoved—the Happy Valley racetrack reported record attendance that very day. The contrast was stark: while one institution prepared for war, another celebrated as though nothing threatened. Japanese forces had been positioned advantageously since autumn 1941. The 23rd Army was already stationed west of Guangzhou, with contingents scattered across Foshan Town, Zhuhai, and other locations on the Pearl River delta. Upon orders for war on 1 December, they faced minimal distances to cover. The army regrouped at Humen, Shilong Town, and Guangzhou itself—staging areas for the invasion. The first visible signs arrived on 4–5 December when Japanese forces landed at Mirs Bay; the 228th Regiment moved into its staging area at Buji Town, now part of Shenzhen City. Maltby received these reports and immediately departed the cathedral service to issue urgent stand-to orders. Yet skepticism persisted within military circles. Some dismissed the threat, dismissing such intelligence as "astonishingly erroneous." Meanwhile, the Imperial Japanese Navy massed in the waters beyond Hong Kong's horizon. The Kowloon Infantry Brigade held a stretched defensive line: the 5/7 Rajputs anchored the centre, the 2/14 Punjab the east, and the 2nd Royal Scots the west. Forward positions extended far—C Company, 2/14 Punjab, under Major S. Burns ranged along Castle Peak Road toward Yuen Long. The brigade screen, C Company of the 2/14 Punjab, held Shueng Shui and Taipo Market with four Bren carriers and two armoured cars from the HKVDC, backed by engineers and the 22nd Fortress Company. The Hong Kong Infantry Brigade held Hong Kong Island itself. The Winnipeg Grenadiers occupied Victoria City on the north-west coast, while the Royal Rifles of Canada stretched along the north-east shores from Saukiwan to Lyemun. Brigadier Lawson commanded from his headquarters at Wongneichong Gap in the island's centre. At 0355 hours on 8 December, the code word "Blossom Blossom" was transmitted. Eleven minutes later, Lieutenant-General Sakai ordered the invasion to commence. Fifty minutes after that, at 0445 hours, Major Charles Boxer—a fluent Japanese speaker formerly of the Lincolnshire Regiment—was roused by Radio Tokyo announcing that war was imminent. Governor Young and Major-General Maltby were informed at once. By 0500 hours, the forward demolition points had been blown by Major Gray, OC of C Company 2/14 Punjab, and the HKVDC engineers under Major J. H. Bottomley. At 0645 hours, the garrison received formal notification: Britain and Japan were at war. Seventy-five minutes later, air-raid sirens wailed. Within moments, smoke billowed from Kai Tak Airport as Japanese aircraft destroyed virtually the entire air garrison. The Wildebeests, Walruses, HKVDC biplanes, and several civilian planes were engulfed in flames. Only two Ju-52s of the Eurasia Corporation and one CNAC T-32 Condor escaped destruction—dispersal bays had never been built due to lack of funds. The Japanese then turned to secondary targets, bombing Shamshuipo Camp and the neighbouring police station. The Canadians suffered few large-scale casualties in the initial strikes, though Sergeant Routledge and Signalman Fairley of the signal platoon were wounded—the first Canadian soldiers injured in World War II. Remarkably, Hong Kong Island's theatres, cinemas, and restaurants continued operating as though the island faced no peril. Japanese demolition efforts proved only minor obstacles. Bridges blown were quickly replaced. The invasion proceeded on three fronts: the 230th Regiment advanced westward along the northern route toward Yuen Long and Castle Peak, targeting the northern slope of Tai Mo Mountain. The 228th Regiment drove down the centre, following the Sheung Shui–Taipo–Kowloon road toward Grassy Hill north of the Shing Mun Redoubt. The 229th Regiment, crossing at Sha Tau Kok and by boat at Starling Inlet, proceeded via Taipo toward Tai Shui Hang and Ma On Mountain. Whenever confronted with strongly defended positions, the Japanese bypassed them entirely rather than engage. The first day brought scattered but spirited resistance. Around 1300 hours, the Punjabis inflicted the first serious check, drawing blood from advancing Japanese forces. At 1830 hours, they ambushed and decimated several Japanese platoons south of Taipo market town. The HKVDC's armoured cars and Bren carriers struck similar successes nearby. The 2 Royal Scots' reconnaissance platoon clashed with elements of the 229th Regiment around Yuen Long. Despite these minor victories, Japanese momentum continued. Armed with stolen British maps and guided by fifth columnists, they pushed forward relentlessly. By late afternoon on the 8th, the Punjabis withdrew toward Grassy Hill to avoid being outflanked. The artillerymen of the 12th Battery engaged an IJN destroyer and Japanese forces around Taipo, inflicting heavy casualties. Farther south near Shatin, on the Taipo–Kowloon highway, rapidly advancing Japanese troops surprised the defenders and severed the fuse at an HKVDC demolition point. Their triumph proved momentary; the engineers completed the backup circuit and detonated the bridge, destroying it along with the Japanese soldiers crossing it. By early morning on 9 December, orders came to withdraw all mainland units to the Gin Drinkers Line. The Punjabis, already at Shatin Wai overlooking Tide Cove, held their position under pressure and continuous shelling for the next two days. The HKVDC withdrew to Fotan, a village south of Shatin. Maltby ordered D Company, 5/7 Rajputs, under Captain H. R. Newtons to Smuggler's Ridge to close the gap between the 2 Royal Scots and 2/14 Punjab. HKVDC armoured cars and Bren carriers continued patrols along Castle Peak Road while HMS Cicala, having survived two air attacks, maintained station in Castle Peak Bay. Despite an IJN blockade, the destroyers HMS Thanet and Scout slipped anchor and escaped to Singapore around 2130 hours. The surviving CNAC and Eurasia aircraft took off with high-value evacuees: Dr. Sun Yat-Sen's widow and her two equally renowned sisters—the Soong sisters—along with Chinese Finance Minister H. H. Kung. Also departing was Lieutenant-Colonel H. Owen-Hughes of the HKVDC, who carried orders to coordinate with the Chinese Nationalist Army for a rear-guard attack on Japanese forces to relieve the pressure on Hong Kong. I would like to take this time to remind you all that this podcast is only made possible through the efforts of Kings and Generals over at Youtube. Please go subscribe to Kings and Generals over at Youtube and to continue helping us produce this content please check out www.patreon.com/kingsandgenerals. If you are still hungry after that, give my personal channel a look over at The Pacific War Channel at Youtube, it would mean a lot to me. On 8 December 1941, Japan attacked Hong Kong with the 38th Division, supported by naval and air forces. Despite British defenses manned by undertrained, under-equipped troops, Japanese forces advanced rapidly across three fronts. Within hours, the RAF was destroyed, demolitions proved futile, and British units withdrew toward the Gin Drinkers Line facing overwhelming odds.
Given on the Thirteenth Sunday after Pentecost, 2026.
My friend Kari Bates (mother of six, lives in Elko NV, life coach) and Kimberly Bates (HR professional living in NYC) joins us to talk about their new podcast called “Coffee and Cocoa”. Their podcast focused on how Kari and Kimberly reframed/rebuilt their relationship on five core principles after Kimbery left the Church. They talk about the early years, the mistakes they made, and how they learned and implemented new tools/perspectives to preserve their relationship. And how their relationship is now beyond preserving the relationship but thriving. It is a beautiful family love story. And a story that they felt impressed to share with others through their new awesome podcast. If you are looking for a real-life story and actionable principles of preserving/growing a family relationship where there are differences in religious and/or political beliefs, I encourage you to listen/connect with their podcast. Thank you, Kari and Kimberly, for being on the podcast and your much-needed work in our community to build bridges. You two are awesome! Links: Kari's Instagram: @karibatesfamilycoach Coffee and Cocoa Podcast: https://www.karibates.com/coffee-cocoa-podcast
Acts of terrorism driven by hateful ideologies continue to injure, harm and kill thousands of innocent people each year.For Hassan Ndugwa – a survivor of the 2010 terrorist attack targeting a screening of the football World Cup Final in Kampala, Uganda – the experience became a call for action.Angered by the misrepresentation of Islam in the aftermath of the attacks, Mr. Ndugwa and a fellow survivor founded the Uganda Muslim Youth Development Forum to prevent violent extremism and promote peace through religious and community engagement.In an interview with UN News's Assumpta Massoi, he emphasised the importance of survivor voices and the role of remembrance in healing.
When's the last time a Philadelphian showed you a random act of kindness? We're asking because the Philly Money Hunt Instagram account depicts someone hiding what appears to be wads of cash in random places around the city and films those who find the money and those who are too late. Host Trenae Nuri, creative producer Abby Fritz and host of the Streets Dept. Podcast, Conrad Benner, talk about other ways strangers have been kind in the city. Plus, they chat about how the police department is planning to use artificial intelligence for 911 calls and whether our city murals should be preserved as historic landmarks. Our newsletter has Philly news & events in your inbox every weekday morning. Call or text us: 215-259-8170 Instagram: @citycastphilly Support our show and get great perks as a City Cast Philly Neighbor. Sign up here. Advertise on the podcast or in the newsletter: citycast.fm/advertise Learn more about the sponsors of this episode: Fitler Club
Bleeding is what kills people after injury. In this companion conversation to their earlier episode, host and vascular surgeon Dr. Wayne Causey asks Dr. John Pavlus, Chief of Interventional Radiology at Brooke Army Medical Center, to do the thing most medical conversations skip. He walks step by step through exactly how a bleeding trauma patient is treated without major surgery. The tools are small. A needle, a short hollow tube called a sheath placed in the artery at the groin, wires thinner than a strand of spaghetti, and catheters steered by live X-ray to the one vessel that is leaking. The patient leaves with a bandage instead of an incision. The decisions behind those tools are what make the difference. It starts with the CT scan. Contrast is injected and images are captured at three different moments, and the timing of those pictures decides what the doctor believes he is looking at. A scan done for a different purpose at an outside hospital can make a patient look like an arterial bleeder when the bleeding is coming from a vein instead, and veins are not something a catheter can easily fix. Getting the timing right is the difference between the right treatment and the wrong one. From there the conversation turns to the system. At Brooke Army Medical Center, a trauma activation commits the interventional team to having a needle in the artery within sixty minutes of the call, at any hour. That standard was not bought with equipment. It was built on years of trust with the trauma surgeons, to the point that when a trauma surgeon calls a bleed, nobody argues about the pictures. Everyone moves, including anesthesia. Then come the organs. The liver is complicated because it carries two separate blood supplies, and one of them cannot be reached easily from the inside. The spleen is the favorite, shut down with a metal coil placed at a precise landmark, sometimes in fifteen minutes. And the conversation closes on thrombin, a clotting agent injected through the skin under ultrasound, no X-ray required. It is cheap, it is simple, and it is the one tool a military interventional radiologist would want in his pack if told to deploy tomorrow. The thread running through all of it is not equipment. It is repetition. Do the same thing the same way every time, and the mind is free to solve the problem that actually matters. Chapters (01:12-07:15) What Endovascular Care Actually Is and What the CT Scan Shows (07:15-12:23) The Sixty-Minute Clock and Activating the Trauma Interventional Radiology Pathway (12:23-20:49) A Bleeding Liver with Two Blood Supplies and Why Access Comes First (20:49-28:59) A Bleeding Spleen with Microcatheters Coils and Knowing When Good Enough Is Enough (28:59-35:51) Thrombin and the Simple Tool Worth Carrying Far Forward Chapter Summaries (01:12-07:15) What Endovascular Care Actually Is and What the CT Scan Shows Dr. Pavlus defines his specialty in the plain language he uses with patients. Minimally invasive, image guided procedures done through pinholes in the skin, either plugging up an artery that is bleeding or lining the inside of an injured one with a small tube. The discussion then turns to the CT scan, where contrast dye is imaged at three separate moments, and how the timing of those pictures determines whether the bleeding is arterial, venous, or a contained pocket of blood called a pseudoaneurysm. (07:15-12:23) The Sixty Minute Clock and Activating the Trauma Interventional Radiology Pathway A trauma surgeon standing at the scanner calls a bleed and the pathway fires. A single alert reaches the interventional radiologist, the nurse, the technologist, and the resident at the same time, and everyone drives in. The standard is a needle in the artery within sixty minutes of the call, and the guest is direct that the only way to hold that standard is to remove every point of debate from the process. Anesthesia is activated at the same moment, because these patients are rarely stable enough for anything less. (12:23-20:49) A Bleeding Liver with Two Blood Supplies and Why Access Comes First The liver is harder than most people assume because it carries two separate incoming blood supplies, and the second one cannot be reached quickly from inside a catheter. That is why a certain grade of liver injury belongs in the operating room with a surgeon rather than in the radiology suite. The guest then walks through his access routine in detail, from ultrasound guided puncture of the artery at the groin to the specific wire and catheter he uses every single time, and explains why keeping the hole in the artery as small as possible matters in a patient who may receive thirty units of blood. (20:49-28:59) A Bleeding Spleen with Microcatheters Coils and Knowing When Good Enough Is Enough Splenic bleeding can be shut down with a metal coil placed at a precise landmark between two small pancreatic arteries. Dr. Pavlus explains why he abandoned one widely used technique after it tore an artery early in his career, and why he now threads a much smaller catheter inside his working catheter to reach the target safely. He is also candid that in an unstable patient at two in the morning, the goal is not a perfect result. It is a live patient who can be handed back to the trauma team. (28:59-35:51) Thrombin and the Simple Tool Worth Carrying Far Forward Thrombin is a clotting agent injected directly through the skin with a needle, guided by ultrasound rather than X-ray. It is the standard repair for a pseudoaneurysm in the groin, but the guest has extended it to bleeding inside solid organs and small vessels in soft tissue that would be difficult or impossible to reach with a catheter. Because it requires no X-ray suite and almost no equipment, he names it as the single technique he would most want available in a far forward combat setting. The episode closes on consistency, repetition, and adapting a fixed base technique to whatever the patient in front of you presents. Take Home Messages Timing of the Contrast Changes the Answer: A CT scan is not one picture. Contrast dye is imaged before it arrives, as it fills the arteries, and again after it has spread, and comparing those three moments is what separates arterial bleeding from venous bleeding from an old finding that was never bleeding at all. A scan ordered for a different purpose at an outside hospital can point a team toward the wrong treatment entirely. Trust Is Built Long Before the Emergency: The sixty minute standard from phone call to needle in the artery is not achieved with faster equipment. It is achieved by removing every point of debate from the pathway, which only happens after years of a trauma service and a radiology service learning to rely on each other. When the trauma surgeon calls a bleed, nobody re-argues the pictures. Everyone moves. Access Is the Whole Game: You can perform the most elegant procedure in the world inside a patient, and if the puncture in the artery is mishandled, that is the only part anyone will remember. Ultrasound guidance takes no meaningful extra time, and keeping the opening as small as possible protects a patient who may go on to receive massive amounts of blood. Perfect Is the Enemy of Alive: In a stable patient with a low grade injury there is time to chase an ideal result. In a crashing patient at two in the morning there is not. Placing a coil in a good enough position and stopping high flow bleeding so the trauma team can move on is a legitimate and often correct decision, and knowing which situation you are in is a clinical skill of its own. The Simplest Tool May Be the Most Deployable: Thrombin injection needs a needle, an ultrasound probe, and a vial. No X-ray suite, no power injector, no shelf of catheters. That is exactly why it stands out as the technique most likely to work far forward, where the equipment, the imaging, and the logistics that a modern hospital takes for granted simply are not there. Episode Keywords interventional radiology, military medicine, trauma interventional radiology, embolization, splenic artery embolization, liver embolization, solid organ injury, thrombin injection, pseudoaneurysm repair, endovascular hemorrhage control, non compressible torso hemorrhage, angiography, microcatheter, coil embolization, Brooke Army Medical Center, combat casualty care, far forward surgical care, vascular surgery, WarDocs podcast, military trauma care, hemorrhage control, John Pavlus, Wayne Causey Hashtags #MilitaryMedicine, #InterventionalRadiology, #TraumaCare, #HemorrhageControl, #CombatCasualtyCare, #VascularSurgery, #WarDocs, #MilitaryHealth Honoring the Legacy and Preserving the History of Military Medicine The WarDocs Mission- WarDocs exists to honor the legacy of Military Medicine, preserve its history, and inspire every generation — across all Services, Corps, and Ranks — to serve with excellence and pride. Through mentorship, coaching, and education, we equip those considering, entering, and serving in military medicine with the knowledge, connections, and community they need to thrive. We celebrate Who we are, What we do, and, most importantly, How we serve Our Patients, the DoW, and Our Nation. Find out more and join Team WarDocs at https://www.wardocspodcast.com/ Check our list of previous guest episodes at https://www.wardocspodcast.com/our-guests Subscribe and Like our Videos on our YouTube Channel: https://www.youtube.com/@wardocspodcast Listen to the “What We Are For” Episode 47. https://bit.ly/3r87Afm WarDocs- The Military Medicine Podcast is a Non-Profit, Tax-exempt-501(c)(3) Veteran Run Organization run by volunteers. All donations are tax-deductible and go to honoring and preserving the history, experiences, successes, and lessons learned in Military Medicine. A tax receipt will be sent to you. WARDOCS documents the experiences, contributions, and innovations of all military medicine Services, ranks, and Corps who are affectionately called “Docs” as a sign of respect, trust, and confidence on and off the battlefield, demonstrating dedication to the medical care of fellow comrades in arms. Follow Us on Social Media Twitter: @wardocspodcast Facebook: WarDocs Podcast Instagram: @wardocspodcast LinkedIn: WarDocs-The Military Medicine Podcast YouTube Channel: https://www.youtube.com/@wardocspodcast
(0:00) Introduction(0:15) Preserving public games(22:39) $5/$5/$10 session(22:51) A4cc on T43cxcQx2x(24:56) KK all-in preflop(25:35) 55 on A75r(26:56) 77 on J98r9r7(28:29) AKhh on Q94cch(30:19) JJ on J92ddxYour correspondent discusses how best to balance private games in card rooms with existing public games. He argues that public games (which can often run with less overhead and less friction than private ones) should be protected — both for the players, so that they can come and go as they please, and, perhaps, for the card rooms themselves.FORUM DISCUSSIONEpisode NotesTwitterCrush Live Poker
Great healthcare design is invisible. You feel calmer, you find your way, your family has a place to rest, and you never notice the hundred decisions that made it happen. Kristin Green, a healthcare architect at Earl Swensson Associates (ESa) in Nashville, builds her whole practice around that idea. In this episode, she walks through what it actually takes to design hospitals that balance operational efficiency with genuine, patient-centered care. We get into the human details and the hard tradeoffs. Why she starts every project by leaving her ego at the door. How ESa levels the room so the nurse's voice carries as much weight as the CEO's. What she'd change about sterile, all-white healthcare spaces. And how AI has earned a real place in her workflow, from generating exterior concepts to building a data-backed space program in hours instead of weeks. This one is for architects, healthcare and AEC leaders, and anyone curious about where design and technology are heading. It's a thoughtful look at building spaces that serve people first and still hold up for the next 30 years. About Kristin Green Kristin Green is a healthcare architect at Earl Swensson Associates (ESa) in Nashville. Known for her collaborative approach, she works closely with design teams, clients, end users, and consultants to guide projects from concept to completion while preserving the integrity of the design vision. A University of Cincinnati graduate, she is skilled in 3D software and driven by the belief that great design should be functional yet beautiful, inspiring, and timeless. Her process starts with intentional listening, so that the spaces she creates reflect the people and stories behind them. What We Cover Introduction and Kristin's path from Cincinnati's co-op program to healthcare design Why ESa's single-office culture in Nashville shapes how the work gets done Intentional listening and the visioning session that starts every project Leveling the room so end users, not just leadership, shape the design Balancing emotional patient experience with operational efficiency The small details that change everything, from daylight to washers and dryers on the floor Emerging trends: hospital robots, pneumatic systems, and AI-enabled patient monitoring How ESa thoughtfully integrates AI, from concept generation to data-backed programming Preserving design vision across years, consultants, and team turnover The Arthur M. Blank Hospital and designing for timelessness over trends Where to find Kristin and closing thoughts Key Takeaways Listening beats telling. The best healthcare projects start with a visioning session where the design team leaves its ego at the door and hears what actually matters to the people who will use the space. Warmth is a design decision. A facility can be cutting edge and still feel human, through daylight, color, and details that fight the sterile, institutional default. AI is a curator's tool, not a replacement. Kristin uses it to generate thousands of concepts and to build data-backed space programs, then applies human judgment to what's real, buildable, and right for the client. The first design is rarely the final one. Collaboration and early contractor pricing often produce a better, more cost-effective result without losing the intended look. Design for decades. Timeless choices matter most in buildings that may not change for 20 or 30 years. Resources + Links ESa website: https://esarch.com https://www.linkedin.com/in/kristin-plummer/ https://www.linkedin.com/company/earl-swensson-associates/ Career Collective: https://www.mycareercollective.com
This episode of Hustle Inspires Hustle features Alex Quin and Richard Wang, the 25-year-old entrepreneur behind Northern Dumpling Kitchen, as they unpack what it really means to step into a family business, carry forward your parents' legacy, and build something bigger than yourself.Richard grew up in the restaurant industry after his parents immigrated to Canada and spent decades working their way up from difficult beginnings. Although Richard studied finance, his long-term plan was always to return to the family business. Today, he's helping operate the restaurant while working toward one of his biggest goals: helping his father work less and eventually retire.The conversation dives into Richard's decision to document the journey publicly, including the renovations, cooking, long hours, language barriers, frustrations, and behind-the-scenes realities of running a restaurant. After roughly 200 days of consistent posting, his content began reaching audiences far outside his local community, with his Instagram following jumping from around 8,000 to 50,000 in approximately two weeks.Alex breaks down why Richard's content works so well and how founder-led storytelling can create an emotional connection that traditional advertising often struggles to replicate. They also get tactical about restaurant growth, discussing OpenTable, Google Business, Toast, email and SMS marketing, customer retention, experiences, and the systems restaurants can use to generate more revenue without massive advertising budgets.Beyond marketing, Alex and Richard discuss money, mentorship, real estate, investing, learning from older entrepreneurs, documenting family recipes, improving communication across generations, and staying grounded while building a business. Ultimately, the episode is about family, entrepreneurship, and what happens when a new generation decides not only to preserve what their parents built, but to take it further.Episode Outline:[00:00] Intro and how Alex discovered Richard's story[03:00] Growing up in the restaurant industry and choosing the family business[06:30] Taking over responsibilities, leadership, and language barriers[10:30] Why Richard's vulnerable storytelling connected with people[14:00] His parents' immigrant story and building a life in Canada[17:30] Family legacy, helping his dad retire, and financial goals[21:00] Money, investing, mentorship, and lessons for young entrepreneurs[23:30] Restaurant growth strategies: OpenTable and Experiences[27:00] Google Business, Toast, email, SMS, and customer retention[30:00] Cooking, communication, and improving his Mandarin[32:00] Going from 8K to 50K followers and founder-led content[34:15] Preserving family recipes and building for the next generation[35:30] Richard's message to immigrants and young entrepreneurs[37:05] Closing thoughts and outroWisdom Nuggets:Your Story Can Be More Powerful Than Your Product: Richard's audience didn't grow because he pretended the restaurant was perfect. People became invested because they watched him struggle, learn, improve, and fight to build something meaningful for his family.Consistency Doesn't Always Pay Off Immediately: Richard spent roughly 200 days building his audience before experiencing a major breakout. His growth from around 8,000 to 50,000 followers happened rapidly, but it was built on months of consistent work.Founder-Led Content Creates Emotional Investment: When customers understand who is behind a business and why they're building it, the relationship becomes deeper than a transaction. People begin rooting for the person as much as the company.Small Systems Can Create Big Revenue: Restaurants don't always need massive advertising budgets. Reservation platforms, Google Business, customer data, email, SMS, and better retention systems can create meaningful incremental revenue.Learn the Business From the Inside Out: Richard isn't only trying to market the restaurant. He's learning to cook, communicate with staff, improve operations, understand the finances, and master the details that make the business work.Protect the Knowledge Your Family Built: Recipes, techniques, processes, and lessons that exist only inside someone's head can disappear. Documenting them turns years of experience into something the next generation can preserve and improve.Diversification Creates Stability: Business, real estate, and stocks can perform differently at different times. Building across multiple areas can create more stability than relying entirely on one source of income.Legacy Is Bigger Than Money: Helping your parents doesn't necessarily mean forcing them to stop working. It's about building a situation where they work because they want to, not because they have to.Power Quotes“Founder-led content drives people, and it plugs at the heartstrings.”— Alex Quin“I'm doing some big things right now, but this is really only the beginning.”— Richard WangConnect with Richard Wang:Instagram: https://www.instagram.com/northern_dumplingkitchenTikTok: https://www.instagram.com/@thedumplingprinceConnect With the Podcast Host Alex Quin:Instagram: https://www.instagram.com/alexquinTwitter: https://twitter.com/mralexquinLinkedIn: https://www.linkedin.com/in/mralexquinWebsite: https://alexquin.comTikTok: https://www.tiktok.com/@mralexquinBooks Mentioned:How to Market Your Restaurant Online — Alex QuinProfit First — Mike MichalowiczOur CommunityInstagram: https://www.instagram.com/hustleinspireshustleTwitter: https://twitter.com/HustleInspiresLinkedIn: https://www.linkedin.com/company/hustle-inspires-hustleWebsite: https://hustleinspireshustle.comThis page may contain affiliate links or sponsored content. When you click on these links or engage with the sponsored content and make a purchase or take some other action, we may receive a commission or compensation at no additional cost to you. We only promote products or services that we genuinely believe will add value to our readers & listeners.See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.
Manufactured housing assets and mobile home communities, along with single-room occupancy motels, and multifamily properties, are key sources of Naturally Occurring Affordable Housing. On this episode, Proactive Realty Group's Canaan Williams (ImpactAlpha Edge) tells how he has acquired and renovated nearly two dozen distressed, condemned or nearly abandoned assets and rehabilitated them, including with energy upgrades, while preserving below-market rents. Dr. Van, as he is known, knows housing from growing up in a working-class Bay Area family that relied on Section 8 subsidies. South Carolina-based Proactive's fixed-income strategy promises monthly payments from the rental stream.
Some local businesses see the value in preserving trades that are still done by hand. Today, we hear about a hand-painted-sign renaissance in San Francisco. Then, a multimedia exhibit that's a celebration of resistance by Trans and queer communities.
Preserving season can be incredibly rewarding, but when the garden is producing faster than you can keep up, it can also become overwhelming fast!Over the years, we've learned plenty of lessons the hard way about what makes harvest and food preservation season easier... and what makes it much harder than it needs to be.In this episode, we're sharing some of the biggest food preservation mistakes we've made and the practical habits that now help us manage the abundance without burning ourselves out.We'll talk about why you shouldn't wait until the produce piles up, how to choose the best preservation method for each food, preparing ahead for a preserving day, and why we always think about dinner before we start!We also discuss something that's easy to overlook: whether you're preserving food your family will actually eat and whether doing it yourself is truly saving you money.Whether you're putting up a few jars from the garden or preserving a year's worth of food, we hope these lessons make your harvest season a little easier.Check out the blog post for more information and any details mentioned: https://homesteadingfamily.com/tips-for-a-busy-preservation-day/Links Mentioned:- School of Traditional Skills Annual FREE Summit: Keep an eye out in future podcast episodes for a link to sign up for the free summit! We'll share it as soon as sign-ups are open!- Homestead Kitchen Magazine: https://homesteadingfamily.com/magazine- Carla Emery's Encyclopedia of Country Living: https://amzn.to/4fXywalTime Stamps:0:00 - Introduction1:30 - Chit Chat8:42 - Homestead Kitchen Magazine10:21 - Main Topic~~~~~~~~~~~~~~~MORE ABOUT US!WELCOME! We're so glad you're here! We are Josh and Carolyn Thomas. Together with our eleven children, we are The Homesteading Family where we're living a self-sustainable life in beautiful North Idaho. Let us welcome you and show you a bit about us here: http://bit.ly/HFWelcomeVideoGrow, Preserve & Thrive with us!Visit us on our blog: https://www.homesteadingfamily.comFacebook at https://www.facebook.com/homesteadingfamilyInstagram: https://instagram.com/homesteadingfamilyRumble: https://rumble.com/HomesteadingFamilyA few highlights you don't want to miss are our FREEBIES!!Healthy Healing at Home – Learn how to confidently use herbal medicine in your home with this FREE 4 video workshop: https://homesteadingfamily.com/HHHytYour Best Loaf – A Free 4 video workshop teaching you how to make great bread at home, every time, regardless of the recipe you are using: https://homesteadingfamily.com/free-bread-workshopBecome a master canner in just a few days with my FREE video series where together we are going to stock your shelves with amazing food the whole family will love.
Hey there, Family Brand OGs and newcomers alike! This week, we're joined by our friend Wayde Elliott, who we had the privilege of meeting this summer while cruising through Croatia with our families. We got to know Wayde and his daughter, Avery, on that trip, and some of the conversations we had together made us realize he had a perspective we haven't shared enough of here on the Family Brand Podcast: what does it look like to intentionally build a strong family when your family doesn't necessarily look the way you once imagined it would? Wayde's story has included some massive changes. After 20 successful years as a dentist, a spinal fusion surgery suddenly took him out of dentistry at nearly the same time he was going through a divorce. He eventually reinvented himself professionally through self-storage and real estate development, but in this conversation, we really wanted to talk about what was happening at home. When Wayde realized his marriage was ending, he made a conscious decision that the divorce didn't have to become destructive. Instead of approaching it from a place of winning or losing, he asked his ex-wife what she wanted and committed to creating the best possible outcome for her, himself, and most importantly, their kids. That decision wasn't always easy, but more than a decade later, Wayde says they still have a good relationship and have been able to remain flexible and supportive as co-parents. We also get into the parts of co-parenting that aren't so simple. What happens when you and your ex have completely different ideas about parenting? How do you respond when a decision in the other household undermines something you're trying to teach in yours? Wayde shares why he's learned to create space before reacting and to pay attention to the meaning he's assigning to an experience. We also talk about the difference between living in what Wayde calls a "suffering state" versus a "beautiful state," and why he's committed to continually becoming a better version of himself. That leads us into one of our favorite parts of the conversation: personal and family core values. Wayde shares his own values—love, learning, abundance, fitness, family, and contribution—and how he uses them as an alignment filter for his life. But maybe the most powerful moment of this episode happened before we ever hit record. During our time together in Croatia, Wayde and Avery had a conversation that changed the way he thought about showing up as her dad. Like a lot of entrepreneurs, Wayde is wired to solve problems—but Avery didn't need him to solve hers. She needed her dad to listen. She told him she wanted more one-on-one time with him, and instead of immediately trying to fix the situation, Wayde sat with her and listened. Since coming home, he's intentionally created more time that's just for the two of them. It's such a simple reminder, but one we all need: sometimes the people we love don't need our solutions. They need our presence. Hit play to hear the full conversation with Wayde! We talk about divorce, co-parenting, fatherhood, ego, forgiveness, personal growth, core values, and what it can look like to choose a beautiful family life even when the original plan changes. P.S. We're here to help you build your own Family Brand, one episode at a time. So go ahead, hit play, and let's grow together! Or, if you are ready to dive into our community, follow the links below! LINKS: All Links Family Brand! stan.store/familybrand familybrand.com/quiz familybrand.com/retreats. Links For This Episode: https://www.instagram.com/waydeelliottdmd?igsh=MTc0emRjMDBxcG1mZA%3D%3D https://www.linkedin.com/in/waydeelliottdmd Episode Minute By Minute: 00:00 Meet Wayde Elliott 01:00 Why Wayde Chose the Family Trip to Croatia 02:00 Losing His Career While Going Through Divorce 03:00 From Dentistry to a Completely New Career 04:18 Choosing an Intentional Divorce 05:00 Asking, "What Does Success Look Like?" 07:00 Creating a Win-Win for the Entire Family 08:00 Preserving a Relationship With Your Ex 09:00 Why Communication Matters in Co-Parenting 10:00 Parenting Differently in Two Households 12:00 The Meaning You Give an Experience 13:00 Love and Forgiveness Over Criticism 14:00 Deciding What Your Family Will Look Like 15:00 Interested vs. Committed 17:00 Why Ego Can Destroy Relationships 18:00 Responding Instead of Reacting 19:00 Living in a "Beautiful State" 20:00 Wayde's Personal Core Values 22:00 The Conversation With His Daughter That Changed Him 24:00 Your Kids Don't Always Need You to Fix It 26:00 Staying Present With Your Kids After Divorce 28:00 Wayde's Advice for Parents Going Through Divorce 30:00 Choosing the Meaning You Give Hard Experiences
Billy Pratt is a distinguished Hawaiian waterman, community leader, and dedicated guardian of a cultural legacy that reaches far beyond the islands. For decades, Billy has worked to preserve and perpetuate the values embodied by one of Hawaii's greatest icons, Duke Kahanamoku; Olympic champion, legendary waterman, and Hawaii's beloved Ambassador of Aloha. As president and longtime director of the Outrigger Duke Kahanamoku Foundation, Billy has helped advance opportunities for generations of young people through scholarships, athletic support, water safety, and community service. As a co-founder and chairman of the Hawaii Waterman Hall of Fame, he has helped ensure that Hawaiʻi's extraordinary watermen and waterwomen, and the traditions they represent, are honored, remembered, and celebrated. His commitment also helped bring Duke Kahanamoku's remarkable story to audiences around the world through the acclaimed documentary Waterman, on which Billy served as an associate producer. But Billy Pratt's legacy is about more than accomplishments. It is about stewardship, protecting Hawaiʻi's history, honoring those who came before, and inspiring those who will follow. He does through his service to numerous boards, committees and organizations all focused on making an impact. Billy represents a powerful Hawaiian principle: that true leadership is not measured by recognition, but by service, responsibility, and what we leave for the next generation; and he does it all with incredible passion, purpose and “whacking it ‘til the wheels come off.” 0:00 – Introducing Billy & His Story 9:20 – Billy's Background & Early Years 19:45 – Finding His Path & Early Influences 31:10 – Major Turning Points Along the Journey 42:35 – Challenges, Setbacks & Lessons Learned 54:20 – Building Success Through Experience 1:06:05 – Mindset, Discipline & Personal Growth 1:18:30 – Stories That Shaped His Perspective 1:30:15 – Relationships, Community & Giving Back 1:42:40 – Lessons From a Life of Experience 1:54:25 – What Billy Values Most Today 2:05:10 – Final Reflections & Closing Thoughts
The preservation of habitat is extremely important for Wisconsin wildlife. Nate Fore sat down with me. He's a Wetland Easement Biologist with Pheasants Forever. He tells me that they're primarily concerned with grasslands.See omnystudio.com/listener for privacy information.
Preserving the Dome and the Dip: Secrets to Mastering the Dandie Dinmont Standard This episode of the Dandie Dinmont Terrier Diaries is part of Pure Dog Talk's Clubhouse. The series features members of the Club Partnership Program who are working to preserve the knowledge of their legacy breeders. Host Laura Reeves is joined by the legendary Donna Francis of Elfwish Kennels. With a storied career spanning decades as an AKC-licensed professional handler, skilled groomer and breeder, Donna shares her rich history in the sport. She recounts her evolution from early days with Beagles, Siberian Huskies, and Pembroke Welsh Corgis to her deep-dive into Dandie Dinmonts. Throughout the conversation, Donna provides master-class insights on the critical mechanics of canine structure, the realities of grooming a three-textured terrier coat, and why "type-to-type" line breeding remains the most reliable path to preserving the breed's authentic hallmarks. Whether you are a newcomer looking for mentorship, an aspiring judge or a seasoned breeder, Donna's candid advice offers a rare and valuable masterclass in breed preservation. From a Sold Saddle to Siberian Huskies: Donna's Early Years The $35 Beagle: Donna shares how she sold her horse saddle at age 11 to buy her first purebred Beagle puppy, who went on to become a champion at her very first dog show.Becoming a Professional Handler: After breeding Siberian Huskies in Olmsted Falls, Ohio, Donna worked hard to earn her AKC handling license, committing to perfection and developing an expert "eye for a dog".The Birth of Elfwish: How a chance meeting with Sandy Wolfskill led to co-owning the number-one Pembroke Welsh Corgi bitch in the country and creating the "Elfwish" kennel name, inspired by a Corgi named Wishin and a small elf statue. Falling in Love with the Dandie Dinmont Meeting Elsbeth: Initially finding them "funny looking," Donna was won over by Elsbeth, a Dandie Dinmont breeder Betty-Anne Stenmark selected for them. Elsbeth went on to win the specialty that weekend and even charmed actor Orlando Bloom in first class on the flight home.Crufts Success with Kevin:Donna reflects on her "heart dog," Kevin, who won second at the Terrier Group at Crufts and became a cornerstone stud dog for their breeding program, siring national winners for years. Deep Dive into Structure, Type, and Movement The Danger of Over-Sizing: Donna warns that many modern Dandies are being bred too large, explaining that a head that is too big or a dog with incorrect proportions cannot properly enter and exit a burrow to do its historical job.The Anatomy of a Working Terrier:The Rib Cage: A long rib cage is vital to protect the dog's internal organs from badgers, paired with a slightly shorter loin.The Topline: Dandies have a highly specific, slight topline rise that is easily bred incorrectly, warning against "roached" backs or rear-high structures.Fronts and Feet: A correct front assembly requires a "wrap around" and strong, tight feet and toes so the dog can dig. Movement on the Go: Donna describes the ideal Dandie gate as a smooth, gliding reach and drive with a stable topline, cautioning against short, choppy steps that signal a flawed front assembly. The Art and Challenge of Grooming Three Coat Textures: Grooming a Dandie is notoriously difficult because of their three distinct textures (the harsh outer coat, soft medium coat, and inner coat).Stripping and Scissoring: Donna discusses the necessity of stripping dead hair weekly to prevent the dog's proportions from looking distorted, emphasizing that a bad grooming job can obscure a great dog, while expert grooming can make a poor structure look passable to untrained eyes. Hard-Hitting Advice for Breed Preservation Type-to-Type Breeding: Donna strongly advocates for breeding "type-to-type" and line breeding, choosing pairs with highly similar physical styles rather than chasing superficial "all-breed winner" studs, which often results in a "mishmash" litter.The Value of Mentorship: Her primary advice for newcomers is to find a trusted mentor, stick with them, and ignore political gossip. She urges established breeders to open their arms and support new exhibitors rather than pushing them away.Temperament is Non-Negotiable: Donna explains that showmanship is "90% of it," requiring a bold, tail-wagging dog with a stable, friendly temperament. She urges socialization through puppy classes and cautions against placing unstable temperaments in companion homes.Dandie Reproduction: Insights on the difficulty of breeding Dandies, including Donna's success with side-by-side artificial insemination (AI) on a table, the necessity of pre-whelping x-rays and ultrasounds, and managing selective C-sections. For more information on Dandie Dinmont Terriers.
Bleeding is what kills people after trauma. That single fact sits at the center of this WarDocs episode, in which host Dr. Wayne Causey, a vascular surgeon, sits down with two military interventional radiologists — Dr. John Pavlus of Brooke Army Medical Center and Dr. Jonathan Schutt, an interventional radiology resident at Yale — to examine one of the fastest-moving areas in modern medicine and what it could mean for the wounded service member. Endovascular care, as they describe it, is deceptively simple to explain and remarkably hard to field: a small stick in the groin or the wrist, image guidance instead of an incision, and wires and catheters small enough to be called straws, threaded through the vascular tree to block a bleeding artery or reline an injured one. As one guest puts it, the patient goes home with a band-aid. The conversation moves quickly from definition to system. At Brooke Army Medical Center, a trauma activation commits the interventional team to needle-stick access within sixty minutes of the call, day or night. That standard was not bought with equipment. It was built on years of bi-directional trust with the trauma surgeons, to the point that the team now responds without stopping to relitigate the imaging. Both guests are blunt that ownership is the price of admission: if interventional radiology wants a seat on the trauma team, it has to show up at two in the morning for cases that are neither lucrative nor glamorous. The harder question is how far forward this capability can go. REBOA is scaled today at Role 2, and stent graft and embolization cases in Role 3 remain largely case-reportable events performed by clinicians who brought their own equipment. The limiting factor, both guests argue, is not technique — it is imaging, logistics, and institutional will. Meanwhile, Israeli teams transition to bunker operations within twenty-four hours, and Ukrainian experience with drone-driven injury patterns is already reshaping assumptions about REBOA and embolization that the United States has not yet tested. The episode closes on people rather than platforms: the case for a military interventional community that crosses Service lines and partners with surgical colleagues, the argument for a skill identifier that lets the system find the right clinician, and a practical inventory of what one interventional radiologist would carry in a backpack if told to deploy tomorrow. Chapters (01:11-06:26) Two Pathways Into Military Interventional Radiology (06:26-10:15) Endovascular Care Explained: A Lot Through a Pinhole (10:15-17:02) The Sixty-Minute Trauma Activation at Brooke Army Medical Center (17:02-26:11) Forward Capability: REBOA, Stent Grafts, and the Role 2 and Role 3 Gap (26:11-35:54) Silos, Superpowers, and the Real Cost Equation (35:54-49:56) Allied Lessons, a Military IR Community, and What Fits in a Backpack Chapter Summaries (01:11-06:26) Two Pathways Into Military Interventional Radiology Both guests trace how they arrived at interventional radiology and at military service — one from the Air Force Academy and a fighter pilot track redirected by a day shadowing an orthopedic surgeon, the other from a childhood spent in a pararescue uncle's uniform and an HPSP commissioning. Each was pulled toward endovascular work by the same realization: that the future of the specialty was in doing more through less. Their training routes differ, one through diagnostic radiology and fellowship, the other through an integrated residency pathway. (06:26-10:15) Endovascular Care Explained: A Lot Through a Pinhole The guests define endovascular care in the language they use with patients: a small poke in the groin or the wrist, image guidance rather than an open field, and catheters threaded through the vascular tree like a plumber working pipes. Roughly ninety-five percent of the work is image guided, most often with fluoroscopy. The host adds the surgeon's framing — always ask what can be fixed through the blood vessel before opening a chest or an abdomen. (10:15-17:02) The Sixty-Minute Trauma Activation at Brooke Army Medical Center A blunt trauma patient arrives, CT shows active extravasation from a high-grade splenic injury, and the trauma activation commits the interventional team to needle-stick access within sixty minutes. The guests describe how that pathway was built on bi-directional trust rather than debate over each scan, and why the team now launches without relitigating the imaging. Both stress that owning trauma call — unglamorous, poorly reimbursed, and at all hours — is what earns interventional radiology its place on the team. (17:02-26:11) Forward Capability: REBOA, Stent Grafts, and the Role 2 and Role 3 Gap The conversation turns to what exists downrange. REBOA is scaled today at Role 2, and endovascular hemorrhage control at Role 3 remains largely a set of case reportable events performed with clinician-supplied equipment. The guests explain stent grafts as simultaneous hemorrhage control and reconstruction, and identify imaging, transport, and packaging — not procedural skill — as the true limiting factors on projecting this capability forward. (26:11-35:54) Silos, Superpowers, and the Real Cost Equation One guest argues that interventional radiology has been siloed by civilian incentives the military has no reason to copy, and that the specialty's real advantage is the fusion of diagnostic reading and procedural skill he calls a superpower. The host and guests weigh the higher up-front cost of advanced imaging and devices against the dramatically lower recovery burden of a pinhole procedure. The biggest hurdle, one guest says flatly, is people — convincing decision makers the capability is worth funding. (35:54-49:56) Allied Lessons, a Military IR Community, and What Fits in a Backpack Israeli teams shifting hospitals to bunker operations within twenty-four hours and Ukrainian experience with drone-driven injury patterns are held up as evidence the United States is playing catch-up. The guests describe the effort to build a military interventional radiology community across Services and to partner with the American College of Surgeons military chapter. The episode closes with a practical deployment loadout — ultrasound, micropuncture kits, sheaths, a base catheter, coils, and wire — and a walk through current training pathways into the specialty. Take Home Messages Bleeding is the mission. The immediate cause of preventable death after trauma is hemorrhage, which is why endovascular capability belongs in the operational conversation at all. Every argument for pushing this capability forward reduces to stopping the bleeding fast enough, and doing it without creating a second catastrophe. Framing the specialty this way makes its military relevance impossible to dismiss. Trust is the system, not the equipment. A sixty-minute call-to-stick standard at a level one trauma center was not purchased — it was built over years of bi-directional trust between the trauma team and the interventional service. Once that trust exists, the activation launches without relitigating the imaging, and everything else falls into motion. Any unit trying to replicate the capability should build the relationship before it buys the gear. Ownership earns the seat. Trauma call is unglamorous, poorly reimbursed, and inconvenient, which is exactly why some centers have written interventional radiology out of the pathway entirely. Showing up at two in the morning, reviewing imaging alongside the trauma team, and taking responsibility for the patient is what secures a permanent place on that team. Presence before the activation is what makes the activation work. The limiting factor is logistics, not technique. Everything done at a level one trauma center is technically achievable far forward — the constraint is diagnostic imaging, fluoroscopy, packaging, and airlift, not procedural skill. Progress therefore depends on investment decisions and institutional will rather than on new procedures. Convincing leaders that the capability is valuable is the hurdle, and funding follows conviction. Allies are already ahead, and the injury patterns are changing. Israeli teams move a hospital into bunker operations within twenty-four hours, and Ukrainian experience with drone-driven wounding is already reshaping assumptions about balloon occlusion and embolization. Planning for the last war is the fastest way to arrive unprepared for the next one. Learning from partner nations now is cheaper than relearning under fire. Episode Keywords military medicine, interventional radiology, endovascular care, WarDocs podcast, non compressible torso hemorrhage, REBOA, stent graft, embolization, hemorrhage control, combat casualty care, Brooke Army Medical Center, trauma activation, expeditionary interventional radiology, Role 2 care, Role 3 care, military trauma system, vascular surgery, image guided procedures, John Pavlus, Jonathan Schutt, Air Force medicine, Army medicine, military health system, battlefield medicine, damage control #WarDocs, #MilitaryMedicine, #InterventionalRadiology, #EndovascularCare, #CombatCasualtyCare, #HemorrhageControl, #TraumaCare, #MilitaryHealthSystem Honoring the Legacy and Preserving the History of Military Medicine The WarDocs Mission- WarDocs exists to honor the legacy of Military Medicine, preserve its history, and inspire every generation — across all Services, Corps, and Ranks — to serve with excellence and pride. Through mentorship, coaching, and education, we equip those considering, entering, and serving in military medicine with the knowledge, connections, and community they need to thrive. We celebrate Who we are, What we do, and, most importantly, How we serve Our Patients, the DoW, and Our Nation. Find out more and join Team WarDocs at https://www.wardocspodcast.com/ Check our list of previous guest episodes at https://www.wardocspodcast.com/our-guests Subscribe and Like our Videos on our YouTube Channel: https://www.youtube.com/@wardocspodcast Listen to the “What We Are For” Episode 47. https://bit.ly/3r87Afm WarDocs- The Military Medicine Podcast is a Non-Profit, Tax-exempt-501(c)(3) Veteran Run Organization run by volunteers. All donations are tax-deductible and go to honoring and preserving the history, experiences, successes, and lessons learned in Military Medicine. A tax receipt will be sent to you. WARDOCS documents the experiences, contributions, and innovations of all military medicine Services, ranks, and Corps who are affectionately called “Docs” as a sign of respect, trust, and confidence on and off the battlefield, demonstrating dedication to the medical care of fellow comrades in arms. Follow Us on Social Media Twitter: @wardocspodcast Facebook: WarDocs Podcast Instagram: @wardocspodcast LinkedIn: WarDocs-The Military Medicine Podcast YouTube Channel: https://www.youtube.com/@wardocspodcast
Hosts Heather and Matthew welcome Dr. Gene Brown, a physician leader at Charleston ENT & Allergy, to discuss the challenges and opportunities facing independent physician practices. As the leader of South Carolina's largest private Ear, Nose, & Throat practice, Dr. Brown shares insight into the growth of Charleston ENT & Allergy and the strategies that have helped the practice remain independent. We also explore the financial pressures and changing regulations impacting independent practices, as well as how allergy symptoms and treatments have evolved since they opened their doors in 1997. Tune in now!
Short description David Brown and Paul Murrell explore the forces reshaping the automotive world, from Porsche's restructuring and Honda's compact electric city car to Hyundai's 40-year journey in Australia. They celebrate engineering heritage, review the Bentley Continental GTC plug-in hybrid and the Mazda MX-5 RF, and examine how open-source software is changing in-car technology. Along the way, they reflect on classic cars, transport planning, vehicle design and why driving enjoyment still matters as the industry evolves. Episode Breakdown • Porsche restructures and EV reality — 00:01:23 • Honda's compact electric city car — 00:08:07 • Preserving engineering heritage — 00:17:42 • Hyundai marks 40 years in Australia — 00:27:06 • Bentley Continental GTC review — 00:32:34 • Motoring memories and classics — 00:38:21 • Android Open Source Project in cars — 00:42:08 • Mazda MX-5 RF road test — 00:55:10 Full summary Porsche restructures and EV reality Porsche's planned workforce reductions prompt discussion about changing global demand, slowing EV uptake and the importance of balancing electric, hybrid and petrol models. The conversation also explores premium pricing, tariffs and why iconic models such as the 911 continue to underpin the brand. Honda's compact electric city car Honda's new small electric hatch sparks a broader look at Japan's kei car heritage, urban mobility and practical vehicle sizing. David and Paul argue that compact cars could ease congestion, improve parking efficiency and better match everyday driving needs. Preserving engineering heritage Recognition for long-time Pyrmont Bridge custodian David Glasson leads to a discussion about maintaining historic infrastructure and passing specialist engineering skills to future generations. The pair draw parallels with classic car restoration and the value of keeping mechanical knowledge alive. Hyundai marks 40 years in Australia Hyundai's evolution from budget newcomer to mainstream leader highlights changing perceptions of Korean vehicles. The discussion covers drive-away pricing, longer warranties and the brand's role in reshaping Australian consumer expectations. Bentley Continental GTC review Paul reviews Bentley's plug-in hybrid convertible, praising its refinement, effortless performance and exceptional comfort. Despite its extraordinary engineering and luxury, the high purchase price and ownership practicality remain part of the conversation. Motoring memories and classics Stories from Cars and Coffee gatherings, historic Chargers, Minis and other classics celebrate Australia's motoring culture. The segment reflects on how enthusiast communities preserve automotive history through shared experiences and restoration. Android Open Source Project in cars Alan Zervis explains how Android Open Source Project underpins many modern vehicle infotainment systems while remaining separate from Android Auto. The discussion examines software-defined vehicles, over-the-air updates, interface design and why intuitive controls remain essential for safe driving. Mazda MX-5 RF road test David Brown and Paul Murrell celebrate the enduring appeal of the Mazda MX-5 RF, using The Unbearable Lightness of Being as a metaphor for the car's lightweight philosophy. They praise its 1,100kg kerb weight, six-speed manual gearbox, naturally aspirated engine and retractable fastback roof, arguing that involvement, balance and driver engagement matter more than outright speed. The discussion explores how the MX-5 has remained true to its original concept since 1989, becoming the world's best-selling two-seat sports car while resisting the trend towards heavier, more powerful vehicles. They conclude that its modest performance is usable in everyday driving, making it one of the most rewarding sports cars on Australian roads. Program Links and Credits Overdrive is produced by Driven Media. Website: carstransportculture.com
Parham Parastaran is the founder of Left Lane Auto, an automotive service company operating 40 brands across 90 locations in 20 states. His career began in his family's Car-X shop while he attended the University of Illinois. He later expanded the business into a 17-location portfolio that included independent tire stores.After selling the company he had built over 24 years, Parastaran watched nearly every employee leave within a year. That experience shaped his approach to employee retention after acquisition: preserve the local identity, protect established working arrangements, and earn the team's support before introducing change.In this episode…Multi-location operators buy shops for their revenue, reputation, and experienced teams. The first push toward standardization often puts those assets at risk. Pay changes alter household income. Schedule changes disrupt family routines. A fast rebrand removes a familiar name that employees and customers already trust.Each additional location increases the pressure to impose a single operating model. That is the central challenge behind employee retention after acquisition. Left Lane Auto protects continuity while its leaders learn how each shop works. Operational changes begin after the local team understands the reason and supports the direction.Parastaran also explains how this philosophy shapes conversations with sellers. Owners receive flexibility in how they exit, remain involved, or retain a financial interest. The business follows strong shops and structures the transition around what keeps each operation stable.Here's a glimpse of what you'll learn: [01:03] Parham Parastaran and Left Lane Auto[01:37] Building an automotive career from one family shop[07:40] Preserving local businesses after an acquisition[13:16] Lessons from losing a long-standing team[21:19] Balancing tire sales with mechanical service[23:27] Evaluating shops and speaking with sellers[26:10] Structuring flexible transitions for former owners[27:38] Expanding deal options through Bertram CapitalResources mentioned in this episode:Parham Parastaran on LinkedInLeft Lane Auto LLC WebsiteTread PartnersGain Traction Podcast on YouTubeGain Traction Podcast WebsiteMike Edge on LinkedInQuotable Moments:“It's the blessing of having nothing, so you have no choice.”“You know, you're going to go backwards when you lose people.”“The things that I learned from the failures was that I don't ever want to put ourselves in a cash position where we're going back.”“We'll follow good stores.”“We'll find a way to say yes.”Action Steps:Build an employee retention checklist after an acquisition that records each person's pay plan, regular schedule, tenure, and responsibilities before the handoff.Pause proposed compensation and scheduling changes until their effects on employees' lives and store performance are documented.Meet with the store manager and longest-tenured employees to identify the routines, relationships, and local practices customers depend on.Keep the acquired shop's local name while reviewing its customer recognition and community value with the existing team.Write the former owner's post-sale role, decision authority, and exit timeline into the transition plan before announcing the acquisition.
Why This Episode MattersItalian food isn't one cuisine. Vincenzo Prosperi explains why Italy's regional traditions matter—and what gets lost when everything is reduced to pizza, pasta and red sauce.Authenticity and innovation can coexist. The question isn't whether chefs should create new dishes, but whether classics should be changed beyond recognition and still carry the old name.Social media is changing restaurants. Vincenzo, Mark and Francis dig into influencer culture, food designed for photographs and the pressure on restaurants to pay for attention.Preserving tradition means supporting the people still cooking it. Vincenzo is shifting his platform toward regional chefs, family restaurants and food traditions that might otherwise disappear.BanterMark Pascal and Francis Schott talk about why their restaurant events have never come from a marketing plan. They come from things they, their staff, and their guests can all get excited about.They look ahead to their annual Big Night dinner and revisit the early days of their table-to-farm dinners, including one memorable attempt that involved hauling restaurant tables and a wood-burning grill to a farm... just in time for the rain.The ConversationVincenzo Prosperi grew up in Abruzzo surrounded by his grandparents, great-grandparents and the food traditions that came with them. After moving to Australia, he saw the distance between the Italian cooking he knew and much of the “Italian” food being served around the world. His wife suggested he turn that frustration—and his passion—into what became Vincenzo's Plate.Vincenzo talks with Mark and Francis about the reaction videos that helped build his audience, from cream-laden carbonara to celebrity chefs taking liberties with Italian classics. But as reaction content became increasingly negative and crowded, he changed direction. Today, his focus is on teaching, ranking Italian products and using his audience to introduce viewers to regional Italian chefs and traditions.The conversation expands into a bigger question: What happens to traditional food in a culture built around clicks, trends and influencers?Vincenzo worries that regional Italian cooking is being replaced by a more standardized version of fine dining while small family restaurants struggle for attention. Mark and Francis see a similar problem in restaurants designed around what photographs well rather than what tastes good.They also tackle the line between preservation and creativity. Vincenzo isn't against innovation—quite the opposite. But before you riff on a cuisine, you should understand what you're riffing on. A new dish can be anything you want it to be. Carbonara, however, has a history.And yes, peas remain under suspicion.Guest BioVincenzo Prosperi is the creator of Vincenzo's Plate, devoted to authentic Italian cooking and the traditions behind it. Born in Pescara in Italy's Abruzzo region and now based in Australia, Vincenzo has built a worldwide audience through recipes, food education, product rankings and collaborations with chefs throughout Italy.Timestamps00:00 Great restaurant events are about creating memorable experiences09:30 Meet Vincenzo Prosperi of Vincenzo's Plate11:30 Growing up with Italian food traditions in Abruzzo18:00 Celebrity chefs, reaction videos and crimes against carbonara25:00 Why Vincenzo moved beyond reaction content28:00 Trust, sponsorships and building an audience without selling out35:00 Is fine dining erasing Italy's regional food traditions?38:00 Influencers, restaurant economics and food made for Instagram43:00 Innovation vs. tradition: when should a classic be left alone?47:00 Mark and Francis remember an old-school trattoria in Italy InfoVincenzo's Plate https://www.vincenzosplate.com/Vincenzo's Plate on YouTube https://www.youtube.com/@vincenzosplateEvents @ The Guys' Places https://www.stageleft.com/events/Subscribe: Restaurant Guys' Regularhttps://restaurantguysregulars.buzzsprout.com/Magyar Bankhttps://www.magbank.com/Stage Left Wine Shophttps://www.stageleftwineshop.com/Our PlacesStage Left Steakhttps://www.stageleft.com/Catherine Lombardi Restauranthttps://www.catherinelombardi.com/Stage Left Wineshophttps://www.stageleftwineshop.com/Reach Out to The Guys!TheGuys@restaurantguyspodcast.com
Greg Peterson says the only place that lack lives is between our ears — and if you've ever stood in your kitchen staring at 50 pounds of tomatoes, you know he's right. We talk about what to do with all of it. Greg hosts The Urban Farm Podcast (he just passed episode 1,000), and this one runs on both our feeds. We get into the harvest mindset, the one mistake that turns abundance into panic, and a long list of ways to put tomatoes, zucchini, and basil up for winter — including a few I'd never tried until this year. If you're new here: I'm Wendi Bergin, and I believe preparedness brings peace. Not fear. Peace. Chapters 00:00 — Welcome, and why we recorded this for both our audiences 01:20 — "The only place lack lives is between our ears" 02:10 — 40 tomato plants, worm composting, and volunteer tomatoes everywhere 03:30 — Seed starting and seed saving: what a grex is and why you'd want one 04:45 — Wendi's confession: she won't eat a fresh tomato 06:15 — The biggest mistake people make when harvest season starts 07:20 — Freeze first, can later — handling the early trickle 08:15 — Jars, lids, and buying in bulk before you're desperate 10:10 — Pint or quart? Wide mouth or regular? 11:30 — Why Wendi stopped canning seasoned sauce and cans it plain 13:40 — Why tomatoes need added lemon juice now 14:10 — Water bath vs. pressure canning, explained simply 16:00 — Tested recipes, rebel canning, and "when in doubt, throw it out" 19:30 — Bone broth and canning elderberry juice 21:10 — Fixing watery salsa, and the steam juicer trick 22:30 — Freezing tomatoes to release the water 23:40 — Don't throw out tomato water: stews, chili, and cooking pasta in it 25:10 — Roasted tomato sauce, skins on 26:00 — The two separate steps of canning: the recipe and the preserving 26:50 — Azure Standard, and what a bulk buying club actually is 29:00 — Freshly milled grain, Sue Becker, and the sourdough question 31:45 — Grain mills: NutriMill, WonderMill, MockMill 34:00 — Tomato powder and everything you can do with it 36:00 — Too much zucchini: relish, pickles, fritters, mock pineapple 39:00 — Zucchini lasagna with no pasta at all 40:20 — Preparedness brings peace 43:20 — Basil: pesto, basil salt, basil sugar 44:30 — Greg's 1,000 episodes and where to find us both A few things worth writing down The number one harvest-season mistake is walking in unprepared. Count your jars and your lids before the tomatoes come in. Today's tomatoes are less acidic than the ones our mothers canned. Added lemon juice is what makes water bath canning safe for them. Water bath handles high-acid foods — fruits, pickles, relishes, jams, acidified tomatoes. Pressure canning is required for meats and vegetables. Canning plain and seasoning later is simpler, safer, and more versatile. The canning process can throw off the flavor of herbs and spices. Follow tested recipes. USDA guidelines, the Ball Blue Book, your canner's manual. When in doubt, throw it out — the jar is worth more than the beans. Links Greg Peterson, The Urban Farm Podcast — https://www.urbanfarm.org Azure Standard — https://www.azurestandard.com Bread Beckers (Sue Becker) — https://amzn.to/4wXJ7cM Ball Blue Book of Preserving - https://amzn.to/4whgqpW USDA Complete Guide to Home Canning — https://nchfp.uga.edu Bill McDorman — seed saving, monthly seed chat on The Urban Farm Podcast Come find me Homesteading Made Simple / Joyfully Prepared — wherever you listen, and now on YouTube. If this one helped, share it with the friend whose counter is currently buried in zucchini. Joyfully Prepared Library -https://www.joyfullypreparedlibrary.com Meal Planning Worksheet - https://joyfulprep.kartra.com/page/planning Joyfully Prepared Realtor - https://joyfullyprepared.com/realestate/ Website - https://joyfullyprepared.com Instagram - https://instagram.com/joyfulprepper Facebook - https://www.joyfulprepgroup.com
Dekada nang dinodokumento ng music photographer na si Niña Sandejas ang ilan sa pinakamahahalagang artists, banda, gigs, at eksena sa Filipino music at art. Pero para sa kanya, hindi lang ito tungkol sa pagkuha ng magagandang litrato—tungkol ito sa pagdodokumento ng kasaysayan at pagpreserba ng kultura.Sa episode na ito ng The Linya-Linya Show, pinag-usapan namin ang kanyang buhay at trabaho bilang music photographer, ang halaga ng documentation, at ang pagtatayo niya ng sariling publishing house na alt164. Kinuwento rin niya ang realities ng pag-publish ng sariling libro, at kung bakit mahalagang magkaroon tayo ng mga aklat at archives na nagkukuwento ng visual history ng Filipino music.At siyempre, pinag-usapan namin ang kanyang Greyhoundz book—at kung paano ang pagsuporta sa isang libro ay nakakatulong din para makagawa pa ng mas maraming libro tungkol sa iba't ibang artists, photographers, at eksena sa Pilipinas.Isang insightful na kuwentuhan tungkol sa photography, publishing, music, history, at kung bakit mahalagang idokumento at pangalagaan ang kulturang binubuo at isinasabuhay natin ngayon.Listen up, yo.
It's easy to pick up a camera and photograph someone, but helping them preserve their legacy takes more thought and connection.Today on the show we're talking all about legacy, and we get into gardening, Gen Z heart hands, grandmas with ipad cameras, and so much more with my guest, Nichole Babiez.----Sponsor of the show: Sofi HYSAOpen a free High Yield Savings Account today and get a bonus $25 deposited in your account, just for signing up.allheartphoto.com/sofi----Connect with Nichole:website: https://nicholebabiez.com/instagram: https://www.instagram.com/nicholebabiezphotography/----Book More Weddings Summit Aug 17-21, 2026Free Ticket for youhttps://go.bookmoreweddingssummit.com/registration-page?am_id=johnmansfield3759----Follow the showPatreon: https://patreon.com/wittpodWebsite: https://podcast.allheartphoto.comInstagram: https://instagram.com/witt.podYouTube: https://www.youtube.com/@wisdominthetangents
Memphis is made vibrant by its green spaces and the wildlife that inhabit them. Those spaces exist, and wildlife populations persist thanks in part to the organizations dedicated to protecting and preserving them, including today's guests from Ducks Unlimited and Wolf River Conservancy. From Ducks Unlimited's work conserving more than 20 million acres of waterfowl habitat across North America to Wolf River Conservancy's work planting more than 60,000 trees and protecting upwards of 20,000 acres of land, there's a lot to unpack in today's conversation. Joining us from Ducks Unlimited is Chief Operating Officer Karen Waldrop, and joining from Wolf River Conservancy is Chief Conservation Officer Ryan Hall.
What happens when a man diagnosed with prostate cancer decides not to undergo surgery or radiation?In this episode of the Intellectual Medicine podcast, Dr. Stephen Petteruti shares the true story of a 51-year-old patient diagnosed with biopsy-confirmed prostate cancer who chose a different path. After years of active surveillance, repeated MRIs, and multiple biopsies, the patient sought an alternative approach designed to monitor his cancer without additional biopsies while focusing on preserving his health and quality of life.Dr. Petteruti explains the rationale behind his non-biopsy monitoring protocol, which combines advanced imaging, lifestyle modification, nutritional strategies, and carefully selected repurposed medications. He discusses how the patient's MRI findings improved dramatically over time and why those results should be interpreted cautiously rather than viewed as evidence of a cure.Throughout the episode, Dr. Petteruti also explores the limitations of PSA testing, the role of PSMA PET imaging, and the importance of understanding all available treatment options before making life-changing decisions. Prostate cancer management should aim not only to control the disease but also to preserve vitality, function, and quality of life.Timestamps(00:00) Diagnosed with prostate cancer at age 51(01:15) Active surveillance and the burden of repeated biopsies(02:40) Introducing a non-biopsy monitoring approach(04:10) Building a protocol: repurposed drugs, lifestyle, and nutrition(05:45) MRI changes from PI-RADS 4 to PI-RADS 1(07:00) Why this is not evidence of a cure(07:45) PSMA PET scan: looking beyond the prostate(08:25) Why PSA alone doesn't tell the whole story(09:20) Preserving vitality while managing prostate cancer(10:25) Why patients deserve more than two treatment optionsEnjoy the podcast? Subscribe and leave a 5-star review on your favorite platforms.Dr. Stephen Petteruti is a board-certified physician specializing in longevity-focused, integrative medicine. He works with men navigating prostate cancer, testosterone, and hormone health, aging, and performance using proactive, evidence-informed strategies grounded in real clinical practice. His approach prioritizes preserving function, strength, and quality of life while helping patients make clear, informed decisions beyond reactive, fear-driven care.
New York Times bestselling author and former U.S. Attorney Barbara McQuade offers an exposé on what she frames as an escalating threat of far-right politics to both national security and American democracy. In The Fix: Saving America from the Corruption of a Mob-Style Government, McQuade draws on her decades of legal experience to argue how systems of organized crime and political opportunism exploit the levers of power — using corruption, cruelty, and chaos as tools to dominate institutions and eliminate accountability. McQuade exposes government tactics like information warfare, aggressive retribution, conformism enforced by fear, and pervasive dismantling of legal checks and balances, thereby attacking public interest and undermining justice in the process. Weaving together courtroom stories, political analysis, and cautionary lessons from history, McQuade makes the case that the threats we face are not merely possibilities — they're already here. The Fix is not just a warning, however. The book is also a call to action. McQuade offers reforms and strategies that she believes can reclaim the rule of law and recenter democracy with the power of the people. The Fix is for everyone concerned about America and for those prepared to take a stand. Barbara McQuade is a professor from practice at the University of Michigan Law School, her alma mater, where she teaches courses in criminal law, criminal procedure, national security, and data privacy. She is also a legal analyst for NBC News and MSNBC, and co-host of the #SistersInLaw podcast. From 2010 to 2017, McQuade served as the US Attorney for the Eastern District of Michigan. She was appointed by President Barack Obama and was the first woman to serve in her position. Earlier in her career, she worked as a sportswriter and copy editor, a judicial law clerk, an associate in private practice, and an assistant US attorney. McQuade is the author of the national bestseller, Attack from Within. She and her husband have four children and live in Ann Arbor, Michigan. Debora Juarez has built a 35-year career focused on legal advocacy and economic development for the most marginalized communities in our state. Debora is an enrolled member of the Blackfeet Nation. After growing up on the reservation, she became the first member of her family to go to college. She completed her undergraduate degree at Western Washington University before attending the Seattle University School of Law. After five years as a public defender, she began working at Evergreen Legal Services (currently The Northwest Justice Project). She served for two years as a pro-tem judge for the King County Superior Court and City of Seattle Municipal Court. Debora was later appointed to serve as a full-time King County Superior Court Judge by Washington State Governor Mike Lowry and later by Governor Gary Locke, as Executive Director of the Governor's Office of Indian Affairs. Eventually, Debora joined a major Wall Street investment firm and created a first-rate Tribal Practice Group. and ultimately partnered with the Williams Kastner law firm, who welcomed her vision of focusing economic empowerment and development beyond tribal lands. Buy the Book The Fix: Saving America from the Corruption of a Mob-Style Government (Hardcover) Elliott Bay Book Company
BNLO is an independent show/opinion piece/podcast about Colorado Springs. Support it on Patreon at https://www.patreon.com/cw/TheBNLOShow Mesa Road has history! Thanks to Judy Beerbaum, chair of the Historic Mesa Road Action Committee, for chatting with me. There's so much we didn't get to. Check out their site at https://www.zeffy.com/en-US/donation-form/donate-to-make-a-difference-21529https://youtube.com/shorts/3FBePQQGrkg
Tez's House Club Classics 009: The Journey to Peak Time Step back into the sound of a classic 90s club night—from the first warm vocal house records welcoming you through the door to the trance, techno and rave anthems that took over when the dancefloor reached peak time. Episode 009 begins with uplifting house and unmistakable voices from Jaki Graham, Tony Di-Bart, Loveland and Baby D. As the room fills, the mix moves deeper through progressive house, tribal rhythms and harder-edged club records before breaking into the euphoric rush of Energy 52, Quench, Drax, Ultra-Sonic and Jens. This is more than a collection of 90s dance classics. It is the full arc of a night out: the anticipation, the first anthem, the rising energy and the moment the entire room becomes completely lost in the music. Featuring classic house, trance, techno and rave from Jaki Graham, Tony Di-Bart, Loveland, Baby D, Mory Kanté, Hyperlogic, Energy 52, Paul van Dyk, Quench, Drax II Ltd, Ultra-Sonic and more. The music matters. The memories matter more. Preserving the soundtrack of the dancefloor. For the complete tracklist and more from the Tez's House Club Classics archive, visit Mixcloud: https://www.mixcloud.com/Tezshouse/ The music matters. The memories matter more. Preserving the soundtrack of the dancefloor.
Area scientists and preservationists are calling on Albany officials to help preserve the James Hall Lab building in Lincoln Park, a designated National Historic Landmark since 1976. Hall was a national prominent geologist and paleontologist in the later 1800s. The building has been poorly served since it was taken by the Albany School system and made part of the (also abandoned) Sunshine School. It suffered a recent fire and outsiders often trespass inside. At a minimum, the building needs to be made secure. Dr. Ed Landing, NYS Paleontologist, emeritus, and Lynne Jackson talk with Mark Dunlea for Hudson Mohawk Magazine.
Dr. Mario Espinoza-Kulick speaks with volunteers from the Surfrider Foundation San Luis Obispo Chapter about what beach cleanups reveal about the health of our coastline. we will explore how community volunteers, local businesses, public agencies, and residents all play a role in protecting the Central Coast's beaches.Listen live and call in Thursdays from 1-2p.m. on KCBX
What does it take to keep wounded Soldiers alive when the next war means hours — not minutes — to the operating table? In this special collaboration between WarDocs and OP MED TV, recorded at the Defense Strategies Institute OP MED Symposium, COL Shaun Brown, MD, Commander of the U.S. Army Institute of Surgical Research, lays out the research agenda that will decide whether the wounded of the next conflict survive. COL Brown's path into Army medicine began with a rejection. Poor eyesight closed the door on the U.S. Naval Academy, so he pursued pre-med as a civilian undergraduate. When 9/11 happened, it solidified both his commitment to medicine and his decision to serve, and he took an Army HPSP scholarship in medical school. He trained in general surgery at William Beaumont Army Medical Center — a program with a long, quiet relationship with the special operations community — where attendings would vanish overnight for operational requirements. He chose colorectal surgery as a fellowship to add technical range he could use in civilian practice and on the battlefield, then joined Joint Special Operations Command at Fort Bragg after a year on staff in El Paso. Now commanding the Army's premier combat casualty care research enterprise, COL Brown is candid about what excites him and what worries him. He is most energized by the Organ Support and Automated Technology department, and he uses a widely shared video of a Ukrainian amputee evacuated by unmanned ground system to make his point: autonomous evacuation without autonomous medical support only moves the walking wounded. Ventilators that read changing physiology and adjust themselves, autonomous IV pumps, and en-route support are the missing half of the capability. His concern is combat wounds. In large-scale combat operations with prolonged evacuation timelines, Dr. Brown expects most damage control surgery to be done for sepsis rather than hemorrhage — the patients who cannot be evacuated become septic extremities. He also walks through the blood problem: whole blood is the standard, low-titer O supply will not be sufficient at scale, and the answer is shelf-stable components, including freeze-dried and spray-dried plasma, freeze-dried platelets, and freeze-dried red cells. The conversation closes on people. Brown details how the Army sustains surgical readiness through a diversified platform of military treatment facilities, civilian partnerships, and untapped Veterans Affairs relationships; how he coordinates with the Reserve consultant to pair complementary skill sets on deploying teams; and why, quoting the Army War College, he still serves: you train for the known and you educate for the unknown. Chapters (00:50-02:16) From Naval Academy Dreams to Army Medicine (02:16-04:33) Colorectal Surgery and the Road to Joint Special Operations Command (04:33-08:06) Commanding the ISR: Autonomous En-Route Care and the Combat Wound Gap (08:06-11:59) Blood, Plasma, and the Shelf-Stable Future of Resuscitation (11:59-18:16) Burn Care, Surgical Readiness, and Partnerships Across Components (18:16-22:22) Forward Surgical Capability Gaps and Why He Still Serves Chapter Summaries (00:50-02:16) From Naval Academy Dreams to Army Medicine Brown describes how bad eyesight ended his plan to attend the U.S. Naval Academy and sent him to a civilian undergraduate program and pre-med coursework. September 11th solidified his decision to pursue both medicine and military service, and he applied for the HPSP scholarship in medical school, choosing the Army's four-year award over the Air Force's three-year option. (02:16-04:33) Colorectal Surgery and the Road to Joint Special Operations Command He explains why he chose colorectal surgery — additional technical skill he could use in civilian practice and on the battlefield — and how residency at William Beaumont Army Medical Center exposed him early to the special operations world. He recounts getting the recruiting call while loading a moving van in New Orleans, then completing assessment and selection before moving to Fort Bragg. (04:33-08:06) Commanding the ISR: Autonomous En-Route Care and the Combat Wound Gap COL Brown identifies the Organ Support and Automated Technology department as the work he is most excited about, using a Ukrainian unmanned-ground-system evacuation video to argue that autonomous platforms without autonomous medical support can only move the walking wounded. He then names his chief concern: combat wound research funding, and his expectation that in large-scale combat operations most damage control surgery will be done for sepsis rather than hemorrhage. (08:06-11:59) Blood, Plasma, and the Shelf-Stable Future of Resuscitation The discussion turns to the evolution from component therapy to 1:1:1 ratios to whole blood, and Dr. Brown's assessment that low-titer O will not be available in sufficient quantity for large-scale combat operations. He details work on freeze-dried and spray-dried plasma, freeze-dried platelets, and freeze-dried red cells, noting that spray-drying is faster, cheaper, and uses equipment roughly the size of two ATMs — a major advantage for distributed manufacturing. (11:59-18:16) Burn Care, Surgical Readiness, and Partnerships Across Components COL Brown addresses the burn casualty problem in a future fight: forward Class VIII resuscitation supply, scaling the Advanced Burn Life Support course for deploying units, and partnerships with civilian burn centers, including placing Army burn surgeons in MILCIV sites. He then lays out the diversified surgical platform — military treatment facilities, community hospitals, and underused VA partnerships — and how he works with the Reserve consultant to pair complementary skill sets on deploying units. (18:16-22:22) Forward Surgical Capability Gaps and Why He Still Serves Asked what a forward surgical team still needs, COL Brown points to an off-the-shelf, infection-resistant biologic vascular conduit as a potential game changer over shunts — with the training investment that would require. He closes with his why, quoting the Army War College maxim that you train for the known and educate for the unknown, and asking who will be left to educate the next generation if experienced leaders walk away during the interwar period. Take Home Messages Autonomous evacuation without autonomous care only moves the walking wounded: Unmanned ground and air systems can pull a casualty off the battlefield, but a platform alone does not sustain a patient who needs a ventilator, a pump, or a transfusion en route. The medical community must be in the ground-maneuver conversation early, because a small design change can turn a logistics platform into a casualty evacuation platform. Autonomous ventilators that read changing physiology and adjust themselves are the missing half of that capability. In the next war, sepsis may drive damage control surgery more than hemorrhage: Prolonged evacuation timelines change the casualty population that reaches a surgeon. Patients in uncontrolled hemorrhage far from a surgical team frequently do not survive the wait, while patients with survivable wounds that cannot be evacuated arrive septic days later. Combat wound research and combat wound solutions deserve renewed funding priority for large-scale combat operations. Shelf-stable blood components are the answer to a cold chain that will not hold: Warm whole blood remains the standard of care, but low-titer O will not be available in the quantities a large-scale conflict demands, and cold chain storage is a logistical vulnerability. Freeze-dried and spray-dried plasma, freeze-dried platelets, and freeze-dried red cells are all in the research pipeline. Spray-drying offers a particular advantage: it is faster, cheaper, and the equipment footprint is small enough to support distributed manufacturing forward. Burn readiness is a supply problem, a training problem, and a partnership problem: Thermal weapons, lasers, fuel, and explosions could produce burn casualty volumes the system has not seen in decades. Resuscitation depends on adequate crystalloid and plasma forward, on teams trained to calculate burn surface area correctly, and on scaling the Advanced Burn Life Support course to deploying units. Long-term capacity also depends on formal relationships with civilian burn centers and on placing military burn surgeons inside those partnerships. Surgical readiness comes from a diversified platform, not from one hospital: Military treatment facility volume alone will not sustain a surgeon's skills, so readiness now depends on layering community hospital partnerships and Veterans Affairs relationships on top of the military caseload. Functional VA hospitals near large installations without strong academic affiliations are ripe for preferred referral partnerships. What surgeons need most is not trauma volume but complexity, which older patients with more complex medical conditions reliably provide. Episode Keywords military medicine, combat casualty care, US Army Institute of Surgical Research, ISR, Shaun Brown, WarDocs, OP MED TV, Army surgeon, damage control surgery, LSCO, large-scale combat operations, whole blood, freeze dried plasma, spray dried plasma, blood products, burn care, Army Burn Center, prolonged casualty care, en route care, autonomous medical systems, trauma surgery, military health system, surgical readiness, Army medicine Hashtags #MilitaryMedicine, #WarDocs, #CombatCasualtyCare, #ArmyMedicine, #TraumaSurgery, #LSCO, #MilitaryHealth, #BurnCare Honoring the Legacy and Preserving the History of Military Medicine The WarDocs Mission: WarDocs exists to honor the legacy of Military Medicine, preserve its history, and inspire every generation — across all Services, Corps, and Ranks — to serve with excellence and pride. Through mentorship, coaching, and education, we equip those considering, entering, and serving in military medicine with the knowledge, connections, and community they need to thrive. We celebrate Who we are, What we do, and, most importantly, How we serve Our Patients, the DoW, and Our Nation. Find out more and join Team WarDocs at https://www.wardocspodcast.com/ Check our list of previous guest episodes at https://www.wardocspodcast.com/our-guests Subscribe and Like our Videos on our YouTube Channel: https://www.youtube.com/@wardocspodcast Listen to the “What We Are For” Episode 47. https://bit.ly/3r87Afm WarDocs- The Military Medicine Podcast is a Non-Profit, Tax-exempt-501(c)(3) Veteran Run Organization run by volunteers. All donations are tax-deductible and go to honoring and preserving the history, experiences, successes, and lessons learned in Military Medicine. A tax receipt will be sent to you. WARDOCS documents the experiences, contributions, and innovations of all military medicine Services, ranks, and Corps who are affectionately called “Docs” as a sign of respect, trust, and confidence on and off the battlefield, demonstrating dedication to the medical care of fellow comrades in arms. Follow Us on Social Media Twitter: @wardocspodcast Facebook: WarDocs Podcast Instagram: @wardocspodcast LinkedIn: WarDocs-The Military Medicine Podcast YouTube Channel: https://www.youtube.com/@wardocspodcast
What makes a vacation rental business attractive to a buyer?Door count may get attention, but it does not tell the full story. Sustainable growth, clean financials, a trusted local brand, strong employees, and direct booking performance can all influence how a business is evaluated.In this episode, Alex & Annie sit down with Igor Simkin, CEO of Monarch Collective, for a candid look at what owners should understand before exploring a sale. Igor shares what Monarch looks for in a company, where sellers are often unprepared, and why the conversation should begin long before an owner is ready to step away.They also explore what happens after the transaction, from protecting the local brand and retaining employees to helping founders choose the role they want in the company's next chapter.Episode Chapters: 05:22 - Why Monarch Collective was created and the opportunity it saw in the market 08:07 - Preserving local brands, teams, and community relationships 11:12 - Anchor acquisitions versus tuck-in acquisitions 11:51 - The qualities Monarch looks for in an established business 13:57 - When owners should begin preparing for a potential sale 16:14 - Creating a transition plan around the founder's goals 22:16 - The performance indicators buyers consider 25:03 - Financial readiness, valuation, and common seller blind spots 28:04 - Building a shared data layer across different technology platforms 33:43 - How Monarch is approaching AI across operations and marketing 38:13 - What an acquisition can mean for employees and local leadership 46:48 - How owners can get feedback before they are ready to sellConnect with Igor:Email: isimkin@gomonarch.comLinkedIn: https://www.linkedin.com/in/igor-simkin-4902993/ Connect with Monarch Collective:Website: https://gomonarch.com/Thinking about the next chapter for your vacation rental business?Connect with Monarch Collective to explore what your options could look like and how to prepare before the time comes.
What makes a vacation rental business attractive to a buyer?Door count may get attention, but it does not tell the full story. Sustainable growth, clean financials, a trusted local brand, strong employees, and direct booking performance can all influence how a business is evaluated.In this episode, Alex & Annie sit down with Igor Simkin, CEO of Monarch Collective, for a candid look at what owners should understand before exploring a sale. Igor shares what Monarch looks for in a company, where sellers are often unprepared, and why the conversation should begin long before an owner is ready to step away.They also explore what happens after the transaction, from protecting the local brand and retaining employees to helping founders choose the role they want in the company's next chapter.Episode Chapters: 05:22 - Why Monarch Collective was created and the opportunity it saw in the market 08:07 - Preserving local brands, teams, and community relationships 11:12 - Anchor acquisitions versus tuck-in acquisitions 11:51 - The qualities Monarch looks for in an established business 13:57 - When owners should begin preparing for a potential sale 16:14 - Creating a transition plan around the founder's goals 22:16 - The performance indicators buyers consider 25:03 - Financial readiness, valuation, and common seller blind spots 28:04 - Building a shared data layer across different technology platforms 33:43 - How Monarch is approaching AI across operations and marketing 38:13 - What an acquisition can mean for employees and local leadership 46:48 - How owners can get feedback before they are ready to sellConnect with Igor:Email: isimkin@gomonarch.comLinkedIn: https://www.linkedin.com/in/igor-simkin-4902993/ Connect with Monarch Collective:Website: https://gomonarch.com/Thinking about the next chapter for your vacation rental business?Connect with Monarch Collective to explore what your options could look like and how to prepare before the time comes.
Feeling overwhelmed by canning and preserving your garden harvest? Wondering if you have to preserve everything you grow to make gardening worthwhile? In this episode, I share how my approach to food preservation has changed over the last few years and how I've found a healthier balance between growing, preserving, and enjoying life. You'll hear the lessons I learned after stepping back from doing it all, how I'm simplifying my preserving routine, and practical ideas you can use whether you have a large garden or just a few extra tomatoes. Free Download: Harvest to Jar Estimator Not sure what to do with 5, 10, or 25 pounds of tomatoes? The free Harvest to Jar Estimator shows you what you can realistically make from different harvest sizes so you can preserve with confidence. https://journeywithjill.net/tomato-harvest-to-jar Key Takeaways Learn how to simplify food preservation without giving it up completely. Discover how to prioritize the crops that matter most to your family. See practical ways to reduce preserving stress while still stocking your pantry. Find freezer, canning, and storage tips that save both time and energy. Create a preserving plan that fits your current season of life. Chapters 00:00 – Why my preserving approach changed 01:52 – My original preserving goals 04:14 – The setback that changed everything 06:33 – Taking a break from tomato canning 08:55 – Small-batch tomato preserving lessons 10:18 – My 2026 preserving plan 13:03 – Simplifying tomato preservation 17:44 – Scheduling preserving without burnout 20:05 – Deciding what not to preserve 24:20 – Prioritizing what matters most 28:02 – Freezer organization tips 31:22 – Steam canning advantages 33:07 – Garden planning that reduces preserving 35:20 – Finding a sustainable balance 36:15 – Free Harvest to Jar Estimator Resource Links Harvest to Jar Estimator (Free) https://journeywithjill.net/tomato-harvest-to-jar Friday Emails (Newsletter) https://journeywithjill.net/gardensignup Recommended Brands & Products https://journeywithjill.net/recommended-brands-and-products/ Amazon Storefront https://www.amazon.com/shop/thebeginnersgarden Complete Garden Planner https://shop.journeywithjill.net/ Disclaimer Gardening advice shared in this podcast is based on my own experience in Zone 8a (Arkansas) and from the feedback I receive from others in different gardening contexts. Your results may differ depending on your location, climate, and growing conditions. Always check your local extension service or trusted resources for region-specific guidance. Some links mentioned may be affiliate links, which means I earn a small commission at no extra cost to you if you make a purchase. As an Amazon Associate, I earn from qualifying purchases.
Honest feedback, and deliberate practice are the foundations of real self improvement. Episode Summary In this episode, Big Keith and I chop it up with Jon Dufresne of Kinetic Consulting. We explore his ever-evolving approach to firearms training, dive into the challenges and realities of night vision for both civilians and law enforcement, and get his take on gear like ballistic helmets, ear protection mounts, and weapon lights. Plus, we discuss current events around drones, surveillance, and the impact of gun control legislation in New York and New Jersey. We also share some thoughts on the importance of setting goals, practicing regularly, and balancing technology and privacy in today's world. Call to Action 1. Join our mailing list: Thegunexperiment.com 2. Subscribe and leave us a comment on Apple or Spotify 3. Follow us on all of our social media: Instagram Youtube 4. Grab some cool TGE merch 6. Ask us anything at AskMikeandKeith@gmail.com 5. Be sure to support our show sponsors; they are a big part of making the show possible. Show Sponsors HSM Ammunition – Consistency, reliability, and American craftsmanship in every round. Check them out at hsamunition.com or look for the orange HSM logo at your local gun shop. Thunder Ranch – World-class, reality-based firearm training under Jack Daniels, offering courses for every level. Get more info at traintr.com. Kings River Custom – Artisan-quality, bespoke firearms and knives. Discover their craftsmanship at KingsRiverCustom.com Key Takeaways Jon's training philosophy treats curriculum as a living document, always updated with improvements and student feedback. "See The Night," Jon's night vision company, focuses on not just supplying night vision gear, but educating users with hands-on integration and white-glove service—mainly for law enforcement, but open to civilians. Deliberate, regular practice (even dry fire for 10 minutes a night) is essential—once-a-year training isn't enough, especially when it comes to high-stress situations like home defense or active shooter response. PCCs (Pistol Caliber Carbines) have limitations—reliability and performance at distance can be serious issues for both civilians and law enforcement. Weaponized Geometry is Jon's class tailored for solo, non-team room clearing, with emphasis on realistic expectations and skill development. Including your kids in training (even in dry fire!) builds healthy lifelong habits. Gear matters: things like helmet covers, counterweights, helmet-mounted lights, and ear protection mounts each have specific pros and cons—Jon breaks it all down. Drones, AI, and surveillance are a double-edged sword, increasing both convenience and risks to personal privacy. Legislative news: New Jersey's assault weapon and magazine ban overturned, and federal efforts push for American-made weapons to equip the military. Guest Information Jon Dufresne Owner of Kinetic Consulting Former 75th Ranger Regiment Specialist in night vision integration and tactical training owner of See The Night Check them out for night vision gear and education Keywords Firearms training, Night vision, Ballistic helmets, Jon Dufresne, Kinetic Consulting, Law enforcement training, Self-defense, Weaponized Geometry, Dry fire practice, PCC reliability, HSM Ammunition, Thunder Ranch, Kings River Custom, Gun laws New York, NJ assault weapon ban, Drone surveillance, AI robotics, Civilian Marksmanship Program, Tactical gear reviews, Ear pro mounts, Home defense, American-made guns
The image of Rosie the Riveter flexing her bicep beneath the words "We Can Do It!" has become an enduring symbol. The iconic poster represents the millions of women who stepped into factories, shipyards and assembly lines during World War II, including Minnesota-raised Jeanne Gibson.Gibson, who died on July 29 at age 100, grew up in Minneapolis before heading to Seattle to help build Navy destroyers as a welder during the war. She is believed to have been one of the last surviving Rosie the Riveters. In the decades that followed, she dedicated herself to making sure the stories like hers would not be forgotten, volunteering at the Rosie the Riveter World War II Home Front National Historical Park in Richmond, Calif.Tammy Brumley is a fellow volunteer known as the park's "Rosie Wrangler.” She was also a close friend of Gibson's in her final years of life. Brumley said the women known as “Rosies” were essential to victory in World War II when they took on jobs that had traditionally been for men after millions left to serve in the military."They stepped up," Brumley told Minnesota Now guest host Kelly Gordon. "Whether it was typical riveters or welders, or people making parachute pins or K-rations, anything for the war effort.”That history becomes more urgent to preserve with each passing year, Brumley said. While a handful of Rosies are still alive, their numbers continue to dwindle.For Jeanne Gibson, the path to becoming a Rosie began in Minneapolis. She was a student at Southwest High School when Japan attacked Pearl Harbor. As a teen, she hoped to become a veterinarian but the nearest veterinary school rejected her application because she was a woman. Gibson briefly enrolled in the University of Minnesota's Cadet Nurse Corps program before deciding nursing wasn't the right fit for her. She and a friend headed west to Seattle, where they both became welders at Todd Pacific Shipyards.Brumley said Gibson had a gift for explaining welding and even found it fun."She describes welding kind of like sewing," Brumley said. "You would tack the item first, which is like basting, and then you would do the welding, which is like sewing. You would just coax that molten metal along."Working union jobs during the war exposed Gibson to something many women had never experienced before: equal pay for equal work. But after the war ended, she quickly encountered discrimination again.Brumley said Gibson left one job after discovering she was earning less than a male file clerk because she was a woman. At another position, she was told she wouldn't advance because she was expected to get married, have children and leave the workforce.But Gibson didn't accept those limitations. She quit those jobs and continued her education, earning a master's and doctorate degree from the University of California, Berkeley.Gibson went on to spend her career as a teacher, and spent her retirement volunteering every Friday to tell her story as a Rosie at Rosie the Riveter World War II Home Front National Historical Park.Brumley said that even in her final days, Gibson remained committed to telling the Rosie story. Just a week before her death, she appeared before before a crowd for the halftime of a professional soccer game to continue advocating for the women who helped reshape the American workforce.Gibson viewed those appearances as an extension of her decades-long career as a teacher, according to Brumley."It gave her another opportunity to teach Americans their history and to help young girls understand the message to advocate for themselves and never take no for an answer," Brumley said.Gibson received many accolades for her work, including a Congressional Gold Medal in 2024, the National WWII Museum's American Spirit Award in 2026 and just last month, a banner of her own at the Rosie the Riveter World War II Home Front National Historical Park.Gibson often joked that she would die with her boots on. In the end, Brumley said, she did. Gibson died wearing her red and white polka-dot tennis shoes, a nod to the scarf depicted in the iconic Rosie the Riveter poster.Use the audio player above to listen to the full conversation with Tammy Brumley and MPR News host Kelly Gordon.Subscribe to the Minnesota Now podcast on Apple Podcasts, Spotify or wherever you get your podcasts.
Watch the full episode on YouTube:We first covered Baseten last year when DeepSeek mania was at peak hype. Now they have raised a monster $13B round and become one of the new cohort of AI Infra decacorns that are (with Nvidia, Intel, and the semis complex) chief beneficiaries of the Inference Inflection. We return to Baseten at the peak of the 2026 edition of Open Weights debate. Ali has published a viral breakdown of Kimi K3:And since you last saw him, Philip has spoken at AI Engineer and written the definitive book on Inference Engineering spotted all over SF:Three years ago, inference engineering barely existed as a category.Today, it is one of the most critical disciplines in AI. Inference engineering inherently tackles a different question than standard model training: “How do you turn those weights from training into a product that is fast, reliable, and affordable at scale?” Focusing on these creates an entirely new optimization problem.In one recent GLM-5.2 experiment, quantizing more of the model actually preserved its benchmark quality while increasing throughput by 20%, because the errors introduced in different layers could cancel each other out.Inference is no longer just the final step after training. It is becoming its own engineering discipline, with its own research problems, infrastructure, and increasingly specialized roles.In this episode, Baseten's Philip Kiely and Ali Taha join swyx and Vibhu to explain what actually happens after a new open model is released and what it takes to turn “we generated a token” into a fast, reliable, production-ready API.We go deep on cache-aware routing, disaggregated prefill and decode, quantization, speculative decoding, KV-cache movement, model parallelism, GPU kernels, and the race to make frontier models up to 10× faster. Philip and Ali explain why inference optimizations can still produce gains of 20%, 100%, or even 200%; how quantization errors can cancel one another out; why identical weights can behave differently across clusters; and how Baseten grafted a Kimi vision encoder onto GLM-5.2 without changing the underlying language model.The conversation then expands beyond LLMs into NVIDIA Dynamo, mega kernels, Rubin, AI-specific chips, local inference, video generation, diffusion versus autoregressive models, and the enormous compute barrier to generating coherent long-form video. Finally, we explore the convergence of training and inference, continual learning through persistent KV cache, and the emerging loop where models help optimize the infrastructure that runs them.We discuss:* What happens when a 200,000-token request enters an inference system* Cache-aware routing and reusing previously computed KV cache* Why prefill and decode are increasingly handled by different GPUs* When dedicated deployments become cheaper and more reliable than shared APIs* How speculative decoding uses a smaller model to accelerate a larger one* Tool calling, structured outputs, and what LLMs actually do* What it takes to support a new open model on day zero* Grafting Kimi's vision encoder onto GLM-5.2* Retrofitting inefficient model layers with components from other architectures* Why models sometimes collapse into repeating the same token* How hardware, kernels, and race conditions create nondeterministic failures* Preserving model fidelity while making inference faster* How quantization errors can cancel each other out* Why inference optimizations still deliver gains of 20%, 100%, and 200%* How optimized serving can make a model up to 10× faster* NVIDIA Dynamo, KV-aware routing, and distributed model serving* Speculative decoding the speculative decoder* Why local AI is about making models less dumb while data-center AI is about making them less slow* Tensor, expert, and pipeline parallelism across GPUs* Hardware-aware model design, auto-tuning, and the case against mega kernels* Rubin and why inference is becoming a systems problem* Whether modern GPUs are evolving into programmable AI ASICs* Why enormous models like Kimi K3 require GB300-class hardware* Why open-source video generation still trails Veo, Kling, and other closed models* The quadratic attention bottleneck behind long-form AI video* Autoregressive video, real-time generation, and compounding quality drift* Why future video systems may combine autoregressive and diffusion architectures* Training for inference and inference for training* Continuous post-training, deployment, evaluation, and improvement loops* How GLM-5.2 helped optimize the kernels serving GLM-5.2 itself* Why faster networking could unlock dramatically faster decoding* Continual learning, KV-cache compaction, and persistent model memoryShow Notes* How to build a day-0 API for Kimi K3* 22580: From GPT2 to Kimi3, ExplainedPhilip Kiely* LinkedIn: https://www.linkedin.com/in/philipkiely* X: https://x.com/philipkiely* Inference Engineering: https://www.baseten.co/inference-engineering/Ali Taha* LinkedIn: https://www.linkedin.com/in/aliestaha/* X: https://x.com/waterloointernTimestamps00:00:00 Introduction and the 200K-Token Prompt00:03:18 Dedicated Deployments, Speculative Decoding, and Tool Calling00:11:26 Launching Production-Ready Open Models00:19:06 Model Retrofits, Failure Modes, and Nondeterminism00:28:22 Quantization and Canceling Errors00:32:15 The Race to 10× Faster Inference00:40:48 Dynamo, Speculation, and Local vs. Data-Center AI00:50:18 Model Parallelism, Auto-Tuning, and Mega Kernels01:00:55 Rubin, GPUs vs. ASICs, and Custom AI Chips01:10:03 Giant Models and the Limits of GPU Memory01:12:42 AI Video, Quadratic Attention, and Autoregressive Generation01:21:47 Audio, Images, and Diffusion Models01:27:32 Training, Self-Optimizing Models, and Continual Learning01:40:06 Closing ThoughtsTranscriptIntroduction: Baseten, Waterloo Intern, and Inference EngineeringSwyx [00:00:00]: Okay, we're here in the studio with Philip, old friend from Inference Engineering, the book, as well as Baseten and everything that you've done, you and I have done before, as well as Ali. Welcome.Ali [00:00:15]: Pleasure to meet you.Swyx [00:00:15]: Waterloo intern.Ali [00:00:16]: Waterloo intern, always.Swyx [00:00:17]: When did you get “Waterloo intern” as a handle?Ali [00:00:19]: As a handle? Oh.Ali [00:00:20]: I think the rebranding happened mid-March. When I saw it was open, I was like, “I have to take it. Up for grabs.”Philip [00:00:26]: The problem is that Ali is really good at his job and is not gonna be an intern much longer.Philip [00:00:30]: So we have to figure out who's gonna get the handle.Ali [00:00:33]: Well, I'll pass the torch over to the next intern.Swyx [00:00:34]: Oh, okay. It can be, like, you just pass it to another Waterloo grad.Ali [00:00:37]: To another Waterloo intern. No, bruh.Philip [00:00:39]: Yeah.Ali [00:00:39]: Intern.Swyx [00:00:40]: Intern, yeah.Ali [00:00:40]: And no.Philip [00:00:41]: You gotta get an intern from Waterloo.Ali [00:00:42]: Yeah, I've gotta get an intern from Waterloo.Swyx [00:00:44]: Right.Ali [00:00:44]: But they have to follow the path.Swyx [00:00:45]: Oh, it could, but it could come from Baseten, so it's like whoever Baseten gets from Waterloo.Ali [00:00:48]: Right.Swyx [00:00:49]: Has the title of Waterloo.Ali [00:00:50]: It stays in the ecosystem.Philip [00:00:51]: Exactly.Ali [00:00:52]: Halfway through the internship, you either get it or you're out.Philip [00:00:55]: You should also do, like, a big graduation ceremony where you change the handle.Ali [00:00:59]: Just say it.Philip [00:00:59]: For everybody.Swyx [00:01:00]: You guys are good at ceremonies, clearly. We had a nice launch of the book, very successful. But before we get into all that, I wanna start off with a fun question for you. Okay, you're an expert inference engineer. What happens when I send a long query, say two hundred thousand tokens into Baseten's inference? What's the process of query through GPU model routing, balancing, all that? What is all the stuff that we don't think about?Long Context Requests, KV Cache, and Cache-Aware RoutingPhilip [00:01:26]: With a long query specifically, the first thing that I'm gonna ask is, “Have you sent me this query before, or at least part of it?” and I really hope you have, because it's gonna be a lot easier for me and a lot cheaper for you. So the first thing that we're gonna look at is some cache-aware routing, where we're going to see, we probably have a number of instances, a number of replicas up serving whatever model you're hitting. We want to send this one to something with, number one, available prefill workers, and number two, ideally some cached input already there so that we can skip prefill on at least part of these two hundred thousand tokens. If you're doing two hundred thousand tokens, it's probably coding or a multi-turn agent or something where you would expect to have that cached. If you don't, we're gonna have to send it to a prefill worker. We've at least on certain models disaggregated prefill and decode, so you're going to have one set of GPUs that's solely going to process the input, create the KV cache, and get you your first token, and then that's going to be passed over to a separate set of GPUs, which is going to run decode. We're going to iteratively make those tokens. We're probably going to have some speculator model in front of that. I'm going to assume that you're doing coding, and because of that, our speculator model, which assumes you're doing coding, is gonna have a high draft token acceptance rate. If I'm wrong and you're asking me to summarize every Harry Potter book, it's gonna be slower. And then we stream that output to you and account for it, charge you, a couple of pennies and say, “Hey, would you like to send another one?”Swyx [00:03:04]: Except Baseten doesn't charge by pennies.Philip [00:03:07]: Well, yeah, we charge. I'm assuming that we're talking about the public model APIs. If you are setting up a dedicated deployment, then yeah, it's not pennies.Public APIs vs. Dedicated DeploymentsSwyx [00:03:18]: Yeah, one of the key differentiators when I was talking with Baseten initially was that people who want very high volume just need to rent by the box, ‘cause then it's up to you to figure out how to saturate the box.Ali [00:03:31]: And more often than not, it's, like, way cheaper if you're pushing, like, millions of tokens per hour, if you just pay per hour instead of pay per token.Philip [00:03:37]: Yeah, they do. I think that we've increasingly seen a lot of demand for the pay per token APIs, just because everyone wants to try open models, and then once they find a use case that's really sticky, then they move over to dedicated.Swyx [00:03:51]: Is there a best practice on when it's time to swap over?Philip [00:03:54]: Couple reasons. Yeah, reliability, that's a big one, right?Ali [00:03:57]: Like, if they have a very specific use case, they want you to train something specifically for them, like they want their own spec dec, for instance, for their own traffic.Swyx [00:04:04]: Spec dec is speculative decoding.Speculative Decoding and Custom SpeculatorsAli [00:04:05]: Speculative decoding, yeah.Swyx [00:04:07]: You have to explain.Ali [00:04:07]: Sorry. Like, speculative decoding is like, if you have a huge model, right? And so the model is going to be generating one token at a time every single turn, every single forward pass. So we attach, like, this little, like, parasite, like this layer that goes on top of the model, and this model just has to predict. It does three very fast autoregressive forward passes, and it will predict, like, three certain tokens, and then you do one forward stage over the entire original model in order to see if those predictions were correct or not, and then you accept them or you reject them. Now, this draft model is traffic specific, so if you, like, Philip said, if you're summarizing Harry Potter books, I can train exclusively that draft model on Harry Potter books, and I can guarantee you that I'm gonna accept the three tokens every single time. And so with that case, I increase your decode speed. I wouldn't be able to provide this to you if you're a shared endpointSwyx [00:04:53]: YeahAli [00:04:53]: ‘cause I have no idea if you're doing Harry Potter, if you're doing coding, if you're doing English. We don't know. Also, there was a thing in the book that mentioned that if they really cared about a specific threshold, chapter four, I think. Do you remember that?Philip [00:05:06]: Yeah. The things that you can do is you can set a specific, like, batch sizing, a specific, like, parallelism strategy if you're trying to optimize for, like, throughput versus latency. You can. Maybe a NVFP4 quant doesn't pass your benchmarks and you wanna run a model at higher precision, you could do that. There's just a bunch of reasons why you might wanna have your own endpoint and the biggest one, of course, just being, like, you don't have to deal with someone else doing a hundred million tokens of benchmarking traffic at the endpoint when you happen to be trying to serve your users.Swyx [00:05:40]: Yeah. I think one thing that is. That is a classic journey. Like, it's people is asking the, what happens when you type Google into the browser. Tool calling, is that just, you're generating JSON or is there more complication beyond that?Tool Calling, JSON, and Structured OutputsAli [00:05:58]: Certain customers that we have, they have their own post-trained models, and so they demand a tool calling that's not just, like parse a file or go find the weather. It's something that's very specific and you have to do post-training on this. And if the post-training on the model is not good or if the quantization after the post-training to get the inference to be fast, the model will struggle reading the JSON file and reading the tool calling. But it doesn't require its own like sandbox. It's not like it's going to use that tool calling to like escape a sandbox or like it doesn't have to be contained. It can just be a normal dedicated deployment. The challenge with tool calling more and more seems to be that the companies want certain tool calling which is a very sensitive thing to train. And because you're dealing with all of the JSON outputs, if it doesn't like close the end of the request in a very certain manner, you end up with a model that did the tool calling and like the thinking and so as a result of that, it didn't see the result and just hallucinated the result as it decoded. That seems to be the most challenging thing with tool calling, not really the sandboxes model.Philip [00:06:56]: Yeah, that's a challenge on the training side and then on the inference side, there's work that you can do to scope the possible output. So we published this at this point close to two years ago, the solution to this problem which is you make a state machine and you use that to constrain the output to a specific format. So this is the structured output problem. If you remember backSwyx [00:07:27]: Yeah, the specific grammar is,Philip [00:07:29]: Yeah, exactlySwyx [00:07:30]: GML had this thing.Philip [00:07:31]: Yeah. So it's like the old-school “make sure this is only JSON”, return only JSON orSwyx [00:07:38]: YeahPhilip [00:07:38]: Grandma's gonna die type of prompts.Swyx [00:07:39]: Is it BNF grammar? At some point OpenAI had released a thing that was like, yeah, if you want to constrain your output, write BNF grammar, back as NOR.Philip [00:07:47]: In our inference system, it's just a specified output format. And you get the guarantee that your output's gonna be structured along that format. And so applying that to tool calls can like help cut down on. You can still call the wrong tool or call no tool. It doesn't solve the certainty problem but it at least solves the output structuring problemSwyx [00:08:10]: YeahPhilip [00:08:10]: Within tool calls.Swyx [00:08:12]: And MCP is just another form of tool, right.Philip [00:08:14]: Yeah, exactly.Swyx [00:08:15]: As far as there's no special thing there.Philip [00:08:16]: The thing I'm always like explaining to people is the LLM is not capable of doing anything. It's only capable of making suggestions of what to do and then if those suggestions are formatted in a certain way and applied to a system that knows what to do with them, then an action occurs.Swyx [00:08:32]: Yeah. Part of the fun stuff is, this is solved outside of tool calling too. Like in an agent loop if the output is not correct or you're right, like reasoning, tool calling was done in the reasoning trace, just be like, “Oh, I don't know what to do. Let me just try again.” And it might get there after a few tries. And on your point of training, sometimes this is harder in smaller models, so you don't have the same exact quality outputAli [00:08:56]: Right.Swyx [00:08:57]: When you just swap from a big model, right?Ali [00:08:59]: Yeah. I will say that, before, I think we need to go back to inference engineering proper.Ali [00:09:04]: But, I had expected that something would replace JSON because it's hard to stream JSON ‘cause JSON must be complete and you must have open and close brackets and everything. So it's hard to parse something or validate something while it's being streamed. So people invented all sorts of things that are like, I forget the name of some of these alternatives, but it's something like TOML, something like YAML. But JSON seems to be dominant still.Philip [00:09:30]: The JSON outputs aren't that long, right? Like you could have a long-- ‘cause tool calls also contain the arguments in them and perhaps for a certain tool you might pass like a very long argument. But my impression of the median tool call is that it's a relatively small number of tokens, right? So I would expect that speculators are generally fairly good at something as formatted as JSON. And so you would have like a pretty fast decode step there and that the streaming wouldn't be as valuable, but maybe I'm wrong about that.Ali [00:10:02]: I think you're also bounded by the software or that the model is gonna integrate with if the software is built with JSON for the tool calls or if the company that you'- if your customer says that this is how our software works and our tools are interfaced with JSON, you can ask them to like, change their software and say like, “Yeah, this is gonna be better for the model.” but like with the right training shouldn't be that much of a difference. Also more profitable if it outputs more tokens probably.Swyx [00:10:25]: Depends on your business model.Swyx [00:10:27]: It really depends. But I will say that, as a writer with like experience a lot with generated output, I do try to move from text to JSON text which is very long JSON, right? Like there's paragraphs in every field because I'm trying to structure it, right?Philip [00:10:44]: Right.Swyx [00:10:44]: I want you to first make factual statements, then make opinions then make bullet point summaries, have dates, have entity references have your sources for references, all these things. Anyway, so these are things that like I think people who really experiment with structural output have to really care about. But, let's, let's recurse up the stack a little bit. Before we started recording, you mentioned something really cool, which is that there's a lot of engineering that-- inference engineering that goes on when a new model provider releases a new model, right? So let's call it GLM-5.2, Kimi K3. I had previously assumed, especially if it's like, well, GLM 5 to 5.1 to GLM-5.2, like that you've supported them before. Is it that much work?What It Takes to Support a New Open ModelAli [00:11:26]: It's a lot of work.Swyx [00:11:28]: Yeah. Okay. So like, a lot of people, all you guys, right whenever a new model launch like, people rush to say like, “Oh, Hugging Face supports this, Fireworks supports this, Spacetime supports this,” and I'm like, “Yeah, of course we support it.” But what goes into that? What goes intoPhilip [00:11:40]: I think it's more than just support it too, right? It benefits the consumer a lot. Like I think it was with Kimi K2.5 or GLM-5.2 the latest, there was an inference war, right? X provider is at 90 tokens a second. The next day we're at 150. The nextSwyx [00:11:55]: I kinda kicked that off with the GLM-5.2.Swyx [00:11:58]: I wrote a Twitter article about. It got like half a million views,Ali [00:12:02]: Based on being numberSwyx [00:12:03]: YeahAli [00:12:04]: Or it's for something else.Swyx [00:12:05]: Yeah. Which,Ali [00:12:06]: Oh my GodSwyx [00:12:07]: Which then got everyone really excited about, hey, how can we, bend tracks a little bit further and,Philip [00:12:14]: There's a difference between support the model, as in I can make a token out of this model, and support a model, as in I have a production-ready API from this model.Philip [00:12:26]: Getting to the point of I can make a token out of this model is not that hard because generally the, open source inference engines, vLLM, SGLang of the world oftentimes even receive weights ahead of time, maintainers do, or the people making the model merge PRs to ensure support. So you generally can, just get it working on the standard open source stack without too much pain in most cases. The challenge is, every inference company is gonna have own proprietary stack. Some open source components, some in-house stuff. And for any arbitrary model, there's going to be some new stuff. Sometimes you get lucky, like K, two five to two six was, like, pretty similar.Quantization, Speculators, and Production ReadinessAli [00:13:16]: Yeah. It was pure continued post-trainingPhilip [00:13:18]: YeahAli [00:13:18]: If I remember correctly.Philip [00:13:19]: Even in those cases, there's still stuff you have to do. You have to redo the quantization work. You're taking the model from. Generally, these models are not released in NVFP4, and we want them to be in NVFP4 for maximum Blackwell compatibility. So we have to perform that quantization, and, calibrate the quantization to make sure that we're not causing any regression in the model's intelligence. And then we also have to train the speculator, as we've talked about. Generally, we have. We have ZDR, zero data retention on our model APIs, so we don't know exactly the traffic that people are sending us, but we know what's popular. We know that coding use cases are popular. We know that agents, agentic use cases are popular. So we can get public data sets that are representative of that traffic and train general speculators. Now, with speculators today, you need to train the speculator using the base model itself because you're getting hidden states out of the model from running inference on these specific prompts, and that is the training data you use to create the speculator. So there's that process which you need the real model weights for. And then there's of course just the process of, standing up all the infrastructure behind it, loading all this stuff, testing it. And then when there's a new model with a newer architecture, I think that, like, the DeepSeek models tend to be the most challenging as they have, like, the most novel architectural stuff going on, model after model. But every new model has something. Kimi K2 had. Oh, sorry, GLM-5.2 hadAli [00:14:53]: Sparse attention.Philip [00:14:54]: Yeah,Ali [00:14:54]: YeahPhilip [00:14:54]: the DSA.Ali [00:14:55]: Right. Which is brought from DeepSeek.Philip [00:14:57]: Yeah. AndAli [00:14:59]: So you can copy-paste then?Philip [00:15:01]: It kindAli [00:15:01]: I don't know how this works.Philip [00:15:02]: So, like we had to, like, build support for that into our runtime. And you're right, like it is really interesting the way that all of these open source labs borrow from each other. For example, like GLM-5.2 doesn't have vision. So something that, Haley, a guy on our team, if we could take a look at this, he, like, grafted the Kimi vision encoder onto GLM-5.2.Retrofitting Vision into GLM-5.2Ali [00:15:27]: We'll be training the projector.Philip [00:15:28]: Exactly. So if you think about, like, the encoder, there's the encoder, which is the part that looks at the image and turns it into latent information, and then there's the projector which likeAli [00:15:38]: You can say latent space. It's okay.Philip [00:15:41]: And then there's the projector that maps it onto, the model itself, and then there's the model weights. You don't wanna mess with the model weights because you run a chance of making the model dumber at something else for the purpose of giving it vision. So instead, Haley started with just a projector, which is only a handful of millions of parameters.Ali [00:16:02]: That would be, yeah.Philip [00:16:02]: Yeah.Ali [00:16:03]: Can you show the training one?Ali [00:16:04]: Like the way it groksPhilip [00:16:05]: YeahAli [00:16:06]: Very interesting.Philip [00:16:06]: And maybeAli [00:16:07]: That right therePhilip [00:16:07]: Maybe Ali, you should take it from here. You've got a betterAli [00:16:10]: Ooh, double the sandPhilip [00:16:11]: Understanding of this than I do.Ali [00:16:11]: Yeah. You can see, like, he. The way he trained this is really cool. At the beginning, he was training it using just like, “Here's a picture of a mountain. Can you describe what's in this mountain?” And that caused it just like the first, learning walls. Like here you can see this all we're trying to teach it is to translate the encoded. Like it's already taken the encoder from Kimi K. It's taken the image. It'Philip [00:16:31]: Yeah. FrozenAli [00:16:31]: FrozenPhilip [00:16:32]: With adapter.Ali [00:16:32]: Exactly.Philip [00:16:33]: Yeah.Ali [00:16:33]: So the brain is frozen and the eyes are frozen. It's just we're tryingPhilip [00:16:37]: AlignAli [00:16:38]: Interconnect between the eye and the brain, right? So the projector. And so you take the tokens and then he's like, “Oh, can you describe what's in this image?” And he's like, “Oh, it's a mountain,” or it's a person or it's a human, whatever the case is. But that didn't cause complete understanding. So he changed it such that every image was associated with a data set of questions. Like, does this image have a white male? Does this image have birds in the top corner? Does this image have a scientist in it? All of that stuff. And it would have to answer questions correctly. And using not just training on describing an image, but being able to answer question, another question, answer over time. Like you can see the grokking, which is like genuinely insane, that retrofitting vision into a large LLM can learn to that extent. And even for images that it doesn't perform well on, for instance, if you ask it a picture of like Stephen Hawking, “Who is this?” Maybe it doesn't get it, but it will say something like, “This is Albert Einstein.” Like it still understandsPhilip [00:17:25]: Close enoughAli [00:17:26]: That this is a scientist who is a man who has, some significant achievements, all that stuff. So that's like really cool.Philip [00:17:32]: Yeah. So, we've covered Hao Tian before, who the author of the LLaVA paper that did this, a while ago. And I think that's very foundational work for anyone who hasn't done vision work before.Ali [00:17:41]: Same with the CLIP and MetaCLIP, where you go from just captioning to building out questionsPhilip [00:17:47]: RightAli [00:17:47]: Off the image and how much better you can get performance.Philip [00:17:50]: Right. Right. Right. Yeah. But what's, what's so exciting about this is if you look at a model like this. Now, this is a little bit more of a research project. It's not. It got to 56% on MMLU Pro, I think. So not quite frontier. But if you're running this model, you haven't suffered any loss on your GLM-5.2 quality. If you don't have an image, it'll just behave exactly the way it used to. And ultimatelyAli [00:18:14]: Which in the inference code you literally do not include the other part, right?Philip [00:18:18]: Yeah. You would just skip the encoder if you don't have an image input.Ali [00:18:22]: Okay.Philip [00:18:22]: Just confirming.Philip [00:18:23]: YeahAli [00:18:23]: Does it affect a lot on the overall inference side? Like you're not adding much, you're adding a very small vision encoder. These are typically likePhilip [00:18:30]: They're super fineAli [00:18:31]: Less than a billion parameters, right?Philip [00:18:32]: Yeah. It's, - There's a little bit less standardization among vision encodersSwyx [00:18:37]: YeahPhilip [00:18:37]: So the support matrix can be a little bit, sparser. But overall, yeah, it's a pretty, it's a pretty minor component of the overall system. And ultimately what you get out of the system is all of a sudden you have Kimi Vision, GLM weights, and DeepSeek attention all in one model.Open Source Model Grafting and Franken-MergesPhilip [00:18:56]: And that's, I think, a lot of the power and beauty of open source, is that you can take all of these different components and combine them together into a system that's better than anyoneSwyx [00:19:05]: YeahPhilip [00:19:05]: Can be individually.Swyx [00:19:06]: People used to say that you would also do Franken-merges where you would take likePhilip [00:19:10]: YeahSwyx [00:19:10]: Layers from each model.Swyx [00:19:11]: Does anyone do that anymore?Ali [00:19:13]: Well, to your point previously when you were mentioning like, the work that goes into supporting a model when it first comes out, like GLM-5.2 or MiniMax M3 or whatever the case is. Sometimes you do have to like, you do have to switch out some things. Like, for instance, the MiniMax M3 head uses full attention, and with full attention you end up with this like insane bottleneck in spec dec ‘cause you're doing auto-regressive token generation for three tokens, and you're doing this like N squared over all of the tokens that are in your sequence. Your KV cache is like very large because it's not sparse, it's not top K. So we find it better to like, okay, we're gonna replace this, we're gonna replace this layer with a layer from another model that's using like GQA, for instance. And then just with the right training, you can get it to have the same acceptance rate. So it is very possible to retrofit layers from other models and very much needed. If a layer is like inefficient, the training just becomes the challenge, like how do you ensure that you train it properly? Which again to your earlier point is like the mesh between training and inference. As in like you need very good training in order to do fast inference. That's like, I feel like more and more becoming true.Swyx [00:20:21]: Yeah. Anything else on the support side when you say like get it to fully production ready?Loop Detection, Race Conditions, and Non-DeterminismPhilip [00:20:26]: Yeah. I think that there's also a question of just, we can test a model to a pretty extensive degree, but we're trying to get it out quickly and then you see a bunch of other people test it and you get interesting results. There was an issue with, GLM briefly where we had some like mode collapses where it would just output the same token over and over again for certain prompts on certain temperatures. Like once you expose an endpoint to the real world, there's going to be, so many more varieties of things given to it that you're able to, discover and patch things. So it's not just a, day zero process, it's then like for the first week, for the first month, if a model remains popular, like how do you both fix bugs and then continue to push the envelope on performance?Ali [00:21:21]: What do you mean you don't want your model outputting S?Swyx [00:21:24]: Is there loop detection on that stuff, by the way? It still happens like quite a lot, which is surprising.Ali [00:21:30]: We have like we, in our endpoint, like if a model was to output the same token like four plus times, we just cut the generation. We say like, “Oh, sorry, this-- Like try again,” or like we will reprocess the request. ‘Cause we know then, like if it, like if, yeah, it's four times the same token, it's probably collapsed.Swyx [00:21:45]: Yeah. Is there a way to opt out in case I really want that?Ali [00:21:48]: You want that?Ali [00:21:50]: I think there's a way that we have to handle it. I'm not exactly certain, but I feel like in certain models, like when they output something like you can imagine, like a table for instance, and so they want, they wanna draw like 12 dashes and 12 dashes. Yeah, I think there's a way for that to happen. I think we only do it on certain tokens. Like we exclude certain special characters.Swyx [00:22:07]: Yeah.Ali [00:22:07]: So we only do it on like certain like S is the most common almost. GLM-5.2Swyx [00:22:11]: OhAli [00:22:11]: And I think it was DSV 4 as well. Like you'd just have like looping issues where like you literallySwyx [00:22:17]: ItAli [00:22:17]: Just have like S.Swyx [00:22:18]: Yeah. Is there a special, something special about S? No, just randomlyAli [00:22:21]: It just seems to be the one token involved.Swyx [00:22:23]: Yeah. And it'Philip [00:22:24]: Is thereSwyx [00:22:24]: And it's only temperature 0Ali [00:22:27]: NoSwyx [00:22:27]: Even at other temperaturesAli [00:22:27]: Even at like 0.9 or whatever, it will still, it will still collapse.Swyx [00:22:30]: That's weird, right?Ali [00:22:30]: It's, it is an inference problem to be honest, like a software problem. Like oftentimes, the image you run will-- like NVIDIA will release an image for instance, and if we will upstream the changes from their latest TensorRT-LLM image into our stack, we'll find that it fixes it. Or oftentimes this will only happen in an inference engine that you're using like SGLang. But if you were to switch to vLLM, that isn't the case. So it seems to be like an extremely like deterministic software issue and not really a model issue. It's not like a weights problem. Like I'- we'll say like, “Oh, it's a problem with the quant. We did PTQ wrong,” right? But that isn't, that doesn't make sense because the same weights used with a different inference engine does not repeat the problem. And sometimes it's, the kernels that are being used in the backend have like these very subtle sometimes race conditions, where if you were to use this model hosted on one cluster, you will never get this problem.Swyx [00:23:19]: Oh my God.Ali [00:23:19]: But if you host it on a different cluster, you will. And the reason is the KV cache transfer from a node to node in that one cluster is using a slower interconnect than the node to node in another cluster. So that exposes the race, whereas in another cluster it doesn't. So then you end up just like, okay, this model is not gonna be hosted on this cluster. We're gonna host it on, another cluster because that cluster exposed that problem. But then it ends up with like, okay, is it the software? Is it the model weights or is it the hardware?Swyx [00:23:42]: There is a thing about this with temperature 0 still not being deterministic, right?Ali [00:23:46]: Right.Swyx [00:23:46]: Mostly because of hardware. Even at temperature 0 same model, you won't always get the same output.Swyx [00:23:52]: Even-- But I'm surprised by the race condition one because, I thought PyTorch was a graph that like guarantees that you at least, execute things in the right order.Ali [00:24:02]: Well, yeah, true. Like I'm not, I'm not saying that there is. Like well, you have things like PTL optimizations where like you can start a kernel before the end of the previous kernel, and that's like ‘cause you want to do that because there'sSwyx [00:24:12]: It's like pipeliningAli [00:24:12]: Expense. Exactly.Swyx [00:24:13]: Yeah.Ali [00:24:13]: But it'- But you don't do it cleanly. Like you overlap a little bit of the execution. No, it is very possible that the kernel itself, like that one block that is supposed to be running in this instance of time, that kernel itself has a race condition. For instance, like a missing barrier. Like often if you're designing a kernel and you want it to make it to be very fast, if you don't test it extensively, you'll, you'll have certain threads access data points from registers before they've been written to by other threadsSwyx [00:24:36]: YeahAli [00:24:36]: For example, because like your barrier is wrong or your synchronization was wrong. But yeah, like the testing itself is very difficult in those like, andSwyx [00:24:42]: And there's no like borrow checkerAli [00:24:45]: What does that mean?Swyx [00:24:46]: Like Rust. Like the. If you're trying to have like memory safety It sounds like a comparable problem.Ali [00:24:52]: Well, yes, but you're working in CUDA, right, NVIDIA GPUs. Like- You just need a higher level language like modular Maybe that's what modular is supposed to do. I don't know.Quantization Quality and Vendor FidelityVibhu [00:25:00]: How do you see keeping quality of the model? So you talked about all these steps of, okay, you gotta do quantization, train your own speculative decoderAli [00:25:07]: RightVibhu [00:25:07]: Run on different hardware. Looking at other model providers, okay, you kicked off a inference speed race on the consumer end. What goes into keeping quality the same across them, right? Sure, you can run benchmarksAli [00:25:22]: YeahVibhu [00:25:22]: But, like, how do you determine how much quantization are there standards? What goes intoPhilip [00:25:27]: There's a few things on quality. Most inference optimizations are lossless. KV caching, for example. You are just recomputing or preventing recomputing the same values. Speculation, of course, if a draft token is wrong, it gets rejected. The main lossy optimization is quantization. And that really comes down to, number one, data format, number two, which parts of the model you choose to quantize, which layers, and number three, like doing a lot of calibration on the quantized weights, to ensure that you're preserving all the outliers. There's other tricks that you can do, though. A big one is long context, ‘cause one thing you asked at, right at the beginning is, “Oh, what's gonna happen if I send a 200,000 token request in?” So with a long input sequence, you need to, store a lot more information. You need to process a lot more tokens. And so even if a model has a context of a certain length, you might, as an inference provider, choose to build an API with a shorter context length, and of course a full length one as well. Because if someone doesn't need the full million token context, for example, you can get them better performance. I don't know if that's exactly like quality of the model. The way that I think about quality is to what degree are we faithfully serving the original model? If you think of a golden implementation of a model that performs exactly the way the model is designed to perform, I think of quality as how close are we getting to that, 100% fidelity of the model.Philip [00:27:13]: You can also, of course, think about quality from the training side and how do you push yourself past 100%. But when I think about purely inference optimizations, it's getting faster while staying as close to that 100% fidelity mark as possible. And certainly our standard internally is that, like you should not be able to tell the difference between our API and a, official API. I think Kimi in particular does a good job of vendor benchmarking hereAli [00:27:41]: YesPhilip [00:27:41]: Where they haveAli [00:27:42]: They released an actual vendor benchmark.Philip [00:27:43]: Exactly, yeah.Ali [00:27:44]: ‘Cause they accused, some people, Amazon? There was some provider that was not doing very well on Kimi's benchmark.Philip [00:27:50]: Yeah.Philip [00:27:51]: So, with Reflect we probablyVibhu [00:27:52]: This was a long time ago, right?Philip [00:27:54]: No.Ali [00:27:54]: Yeah, like threeVibhu [00:27:55]: They alsoAli [00:27:55]: Four, five months agoVibhu [00:27:57]: This also happened with, I don't remember which model, but they pulled out quite a few, and then they started a whole chart about this. It might have beenPhilip [00:28:03]: Kimi Vendor Verifier.Ali [00:28:04]: Yeah.Philip [00:28:05]: Yeah.Ali [00:28:05]: Yeah, ‘cause you, ‘cause you'd be pissed, right? Like if you'Philip [00:28:07]: Yeah.Ali [00:28:07]: If like if I'm a consumer and I'm using like Amazon's endpoint for instance, and I've used Kimi and I'm like, “Oh my God, like this is bad,” I'm not gonna say, “Oh, Amazon quantized the model in a bad way.” I'm gonna say, “Oh, Kimi sucks.” Right?Philip [00:28:17]: Yeah.Ali [00:28:17]: So it seems like that makes sense.Philip [00:28:19]: Yeah, they care. They care.Vibhu [00:28:21]: Justifiably.Ali [00:28:21]: Yeah, justifiably.Vibhu [00:28:22]: This is probably a stupid question, but just checking, has anything improved from main quantization?Philip [00:28:28]: Yeah.Vibhu [00:28:28]: Like, is quantization always strictly worse?Ali [00:28:30]: Well technicallyVibhu [00:28:32]: NoAli [00:28:32]: It's a lossy. QuantizationPhilip [00:28:33]: YeahAli [00:28:33]: Is a lossy, it's a lossy implementation.Philip [00:28:36]: Speed improvesVibhu [00:28:36]: Speed improves.Ali [00:28:37]: It the number, likeVibhu [00:28:38]: No, I' always look for inverse scaling laws.Philip [00:28:40]: Yeah.Ali [00:28:40]: Yeah.Vibhu [00:28:40]: This is something I learned from Noam Brown, where like things that normally act in one direction sometimes do.Philip [00:28:45]: Well, technically when you run a benchmark, because these models are deterministic, sometimes your,Ali [00:28:52]: YeahPhilip [00:28:52]: NVFP4 quant is like, two basis points higher than yourAli [00:28:56]: No, it's noise. It's noise.Philip [00:28:57]: Yeah, exactly. I'm like, yeah, it's, it's within. That's why I always say within margin of error.Philip [00:29:01]: And I stopped saying that because everyone assumes that what is, well, within some margin of error, we're barely inside of that to the worst, so we're saying. But yeah, sometimes it's just like, gives you a higher output score. But like Ali said, that's noise. To my knowledge, you're not necessarily making the results better. You're just trying to, again, like keep your fidelity as close to 100% to the original model.Layer Selection, KL Divergence, and Better QuantizationAli [00:29:27]: There is, to your point, research that we did on MP. I don't know if you are able to pullPhilip [00:29:31]: YeahAli [00:29:32]: A tweet we did. One of our research interns, Joshua, I think it's a tweet on how we have 20% better quantized GLM-5.2 than NVIDIA. Essentially what we found throughout like this month research is, okay, quantization is a lossy. It's. You're compressing the data from, occupying 16 bits to occupying, four bits, for instance. And so you're losing some information, and you're trying to minimize that. And so when I say that I'm gonna quantize the model, my job becomes how do I find the layers that I can quantize, and how to find the layers to not. For instance, with image models, I don't quantize modulation layers, and I don't quantize out projections because those two are. Like out projection is what you see as the user. Modulation is what the model sees or understands. Right, exactly. And so to his paper, do you have the. It doesn't have the. Yeah. It's a long paper. I don't know if I can findVibhu [00:30:25]: If there's a part to search or it's probably in the thread.Ali [00:30:28]: It's probably in the thread.Vibhu [00:30:29]: Yeah.Ali [00:30:29]: But the long and the short is it is very possible that quantizing more of the model makes the results. Like if I have a model that I quantize layers one, five, and 10, and another model where I only quantize layers one and It is possible that the model in which I quantized more information is going to perform better because the quantization errors have canceled out. And so what Joshua showed in his mathematical proof where he had like a verifier in, is that you can predict which layers are going to have quantization errors that will cancel out with each other, and you choose to quantize those layers. And so the result of doing this mathematical quantization is you end up with a model that's 20% more quantized than another provider, so you get 20% more throughput of it because there's more layers than running an NVFP4, and your quality is better than that other quant because the layers that you chose to quantize have their errors cancel out, like one layer skewed to the right one layer skewed to the left, one layer skewed to the right. Your final logits distribution is more similar to the original distribution of the model, so you have better fidelity. And so the way we proved this was with KL divergence. So instead of just scoring on the benchmarks, we scored the KL divergence between the logit distribution of the quantized model and the logit distribution of the original full precision model, and we showed that with this technique we get. If your probability distribution on the logits which token it wants to select is more of the same as the original model, you're probably gonna end up staying true to the original model. So yeah, so it seems like previously before this, it seemed like the industry was, well, the more you quantize, the worse it's gonna be, ‘cause the more loss you introduce. That's not exactly, not necessarily true. So yeah, doesn't improve it, but can cancel out.Philip [00:31:57]: I think it might be this, but reminds me a good bit about pruning where you can prune off certain layers.Philip [00:32:03]: But very interesting. Didn't know this was a whole paper you guys put out.Ali [00:32:06]: It's. Fun fact, it was originally 72 pages, this paper, and then we decidedPhilip [00:32:11]: WowAli [00:32:11]: We can't tell. We couldn't release it. So it's now 45.Swyx [00:32:15]: Still 39 pages, so very substantive. We talked about evals and all these things and, like what's possible in terms of speedup? Like it's like probably like the numberInference Speedups and BenchmarkingSwyx [00:32:25]: Thing that people do wanna care about, and it's something that you wrote about in your post. Like official API is 70 tokens per second, and you push it up to 90. Is that like a normal thing?Philip [00:32:36]: So what's cool about working in inference, the reason that I think inference is going to be a useful place to do engineering for a long time, is that if you look at highly optimized domains like, say, finance, if you're in finance, you measure how much better you got in basis points. It's like, “Oh, I got five basis points better, like twentieth of 1% better,” that's huge news because everything is so optimized. When we publish optimizations, it's 20%, it's 100% it's 200%. So there's still probably like a lot further to go, honestly. Like you'll, you'll know that inference is pretty much solved when researchers start publishing about how they got 1% faster at something.Swyx [00:33:19]: Which by the way, because I am from the finance background, in the ‘70s, that was the margin at the time. When you did quantitative finance research, you would findAli [00:33:27]: And like 20%, tens of percent.Swyx [00:33:29]: That's. Yes.Philip [00:33:29]: Yeah.Swyx [00:33:30]: And now it'Philip [00:33:31]: Tiny fractionsSwyx [00:33:32]: For those people interested, look up Andrew Lo's paper. He had a really interesting illustration of quant, stat arb, distribution, narrowing down from like those kinds of 20% differences in the ‘70s, down to nothing today, which is very cool.Philip [00:33:48]: Exactly, and we're at the beginning of the same type of thing. Now benchmarking is hard. I think anyone will tell you that, and benchmarking provider speeds is hard because there's so many variables that go into it. What hardware are you using? How much load do you have on the system? What's the exact nature of the prompts and input and output sequence lengths? All that stuff. But overall, when you start stacking these improvements, you're looking at multiples. You can look at it. The most common form, of course, is TPS, tokens per second, which is bad naming by us in the industry, ‘cause there's two tokens per second. There's tokens per second, the throughput number, and the latency number.Ali [00:34:31]: TTMT, yeah.Philip [00:34:32]: Like total tokens per second out of the, out of the GPU as a throughput number. Most people only care about tokens per second as the latency number, which we should call ITL, intertoken latency, but we don't.Philip [00:34:44]: Anyway, so you can imagine a standard API without many optimizations for a 1 trillion parameter model operating somewhere in the 30 to 50 tokens per second range for reasonable traffic profile. And we generally see the goal of, pushing to 10X that. But, not necessarily day zero, but by stacking enough optimizations, if you have, say like four optimizations, each of which doubles performance. Or sorry, three optimizations, each of which doubles performance, then you stack that up, that's an 8X gain. That's the order of magnitude that we're working with in this space. We're trying to make things substantially faster, not just go from like 70 to 90.Swyx [00:35:38]: Are you saying you've. You have done that?Philip [00:35:40]: So let's say you have as a reasonable baseline, 30 or 40 tokens per second. You can achieve 10X that. So like on GLM-5.2, if you run it unquantized, perhaps on H100s even, and you're just using an off-the-shelf inference engine with no particular optimizations, no speculator, nothing extra around like KV routing, no disaggregation, you're, you're probably, yeah, looking at that like 30 to 40. You think that's like a reasonable baseline?Swyx [00:36:12]: Right. Right.Philip [00:36:12]: To get to something like 10X, there's a lot of trade-offs that you're making. If we're running at more like a 300, 400 tokens per second range, you are using the best hardware possible. You have a optimized speculator. You have done all of your quantization work. You are Seeing a pretty high cache hit rate. You are running with a reasonably small batch size and a parallelism configuration that is tuned for latency versus throughput, but it is possible. So the spreads that you see if you, like, go on artificial analysis or you go on OpenRouter and you look at, the worst provider to the best provider, oftentimes can hit that range. 10X is of course very aggressive. It's oftentimes maybe more of a four to six times improvement. But that's the performance that makes us really excited, is when we can get these huge gains, not just go from 70 to 90 tokens.Stacking Optimizations: NVFP4, Speculation, and DisaggregationAli [00:37:19]: It's also, like, hardware dependent. Like, ifPhilip [00:37:20]: YeahAli [00:37:20]: If you have a thing where you're serving it on just, like, a node of H100s and then you throw, like, you shard the model across, like, four nodes of B200s. Like, you can definitely increase the speed with just throwing more hardware at it. Like, normalizing for the same exact hardware and the same number of GPUs.Philip [00:37:35]: Yeah. Then you're looking at, like, a two to 4X improvementAli [00:37:38]: Right. RightPhilip [00:37:38]: Depending on the inference optimizations. So yeah, it's. Some of it's, what's the call, and some of it's who's the driver.Vibhu [00:37:46]: If you break down the two to 4X, say the example is run GLM-5.2Ali [00:37:51]: YeahVibhu [00:37:51]: On B200sAli [00:37:53]: YeahVibhu [00:37:53]: Single node, right? What's, like, the cost trade-off for effort to get, like, the last bit of juice out versus what should people just think of, right?Ali [00:38:01]: Spectre quantization. Yeah.Vibhu [00:38:03]: Spectre quantization.Ali [00:38:04]: That's, that's, that's like 95%. LikeVibhu [00:38:06]: And how far does that get you? And how easy is that for the average person to do? So say right I wanna throw the weights of GLM-5.2 on a node of B200s, how easy is it to find speculative decoder- decoder model or already quantized model? How much work goes into it?Philip [00:38:23]: If you're doing it up front, it's quite a lot of work. If you're doing it today, there's going to be people who have published things that you can just, you can just grab some NVFP4 weights. You can grab a speculator. Yeah, if we're thinking about, like, what are the 2Xs we're stacking, going from, BF16 to NVFP4 is, it's not quite a 2X, right? It's like. I think it's about, like, 30 to 40%, from 16 to 8, and then another 30 to 40% multiplied from, 8 to 4. So that doesn't quite get you a 2X, but, like, roughly a 2X. Speculator, roughly a 2X. Disagg on top of that if you're able to get enough hardware and put enough traffic through it, another roughly a 2X. And then you add in some, double-digit percent increase from having just a better runtime with, the latest kernels and stuff behind it. And that's how it stacks up.Ali [00:39:21]: YeahPhilip [00:39:21]: So building each of those, like, building the, quantized weights is, for someone who really knows what they're doing, hours to days of work. Building the speculator, again, like, hours to days of work. And the, disagg setup, hours to days. Well okay, but like once you haveAli [00:39:39]: Once set up. Once set up. YeahPhilip [00:39:40]: Yeah, getting disagg working for the first time, I'm saying, of course, is very difficult.Philip [00:39:44]: The marginal implementationAli [00:39:48]: Like, if you're just grabbing, like if you are a person, like just a normal consumer who has access to, like, a node of B200s and you're wondering, “How can I just host it myself?” You don't need to quantize the model yourself. There's always gonna be, like, an open source quantized checkpoint. NVIDIA's gonna push one out if no one else does. You. Usually, the providers will have their own spec dec that they've trained as well. You don't need to train your own spec dec. You can just use that as well.Philip [00:40:09]: Yeah. Like, GLM-5.2 has its own MTP.Ali [00:40:13]: Right. Right.Vibhu [00:40:14]: What's multi token prediction?Philip [00:40:15]: Yes.Ali [00:40:16]: I'm justVibhu [00:40:16]: Can you explain that?Ali [00:40:16]: I'm just an expert.Ali [00:40:18]: I can do it for you in case I get it wrong?Vibhu [00:40:20]: No.Vibhu [00:40:21]: Yeah, you should correct if we're wrong, but their multi-token prediction can be used for self-speculative decoding.Ali [00:40:27]: I'm not sure. I'm not gonna correct that.Vibhu [00:40:28]: Okay. I'm semi-confident in thatAli [00:40:30]: Okay. YeahVibhu [00:40:30]: But someone can check. But it's useful to paint the story of, okay, not just the average person, but say a company wants to switch from serverless inference I wanna throw this up on. I wanna rent some GPUs, throw it up. These are the steps you take to do significantly faster than just put it behind vLLM.Ali [00:40:48]: Right.Vibhu [00:40:49]: I was waiting for a mention of Dynamo.Vibhu [00:40:51]: I feel like, that's supposed to be the baseline that you measure against.Dynamo, KV Routing, and Disaggregation ToolkitsPhilip [00:40:55]: I would think of Dynamo as less of a box system and more of a toolkit for building with. So when we talk about doing aware routing, when we talk about doing KV offloading, when we talk about doing, PD disaggregation, Dynamo fundamentally is. By the way, Dynamo is an open source library from NVIDIA.Ali [00:41:17]: We've done a pod with KylePhilip [00:41:18]: OkayAli [00:41:19]: Kyle Cranin.Philip [00:41:19]: Cool. So then your listeners know then that it supports all the different inference frameworks. And it is multi hardware, which is interesting.Ali [00:41:28]: But it's just a router, it's not like an optimizer layer.Philip [00:41:30]: Yeah. All it does, like, what Dynamo is good at, it is a library for moving information around your cluster, around your hardware. So if you have, KV cache on one place and you need it to be somewhere else, Dynamo coordinates NIXL for you to move that around.Philip [00:41:49]: That doesn't mean that, like, out of the box, you just say, “Pip install Dynamo,” and then you get, like, a massive performance speed up. It's more of a developer toolkit.Ali [00:42:01]: Yeah. I would have said it would. It comes with a set of defaults that you can then swap out.Philip [00:42:06]: It does. If the industry at large, I think, was, like, rolling out all of these deployments, standard, then I think it would be, like, a credible baseline. But, we've got to, we've got to benchmark against, like, what we're seeing in the wild.Speculative Decoding Methods: Medusa, EAGLE, n-Gram, and Spec-SpecVibhu [00:42:23]: I did wanna talk a little bit more about PD disagg, because that is probably, like, number three after quantized and speculative decoding. In your book though, I was just gonna pull out the book.Philip [00:42:31]: Yeah.Vibhu [00:42:32]: Like section 522 on Medusa, 523 on EAGLEPhilip [00:42:35]: YeahVibhu [00:42:36]: 524 on gram.Philip [00:42:37]: It's 55, would be disaggregationAli [00:42:42]: Yeah. Well, no, I just wanted to dwell a little bitPhilip [00:42:44]: YeahAli [00:42:44]: The other. Like, so what do you choose to include? What do you choose to not to include? Because there was all these other techniques.Philip [00:42:51]: Yeah.Ali [00:42:51]: Are these still relevant? Because I think they came out, like, a year and a half ago maybe.Vibhu [00:42:55]: Medusa is quite old.Philip [00:42:56]: Yeah, Medusa's old.Ali [00:42:58]: It was old.Vibhu [00:42:58]: But is it in the book as a good, here'sPhilip [00:43:01]: BaselineVibhu [00:43:01]: Baseline vanilla understand it?Philip [00:43:02]: Like you should know this.Vibhu [00:43:03]: Like I read the paper, I'm like, “ it makes so much sense.”Philip [00:43:05]: Yeah.Philip [00:43:05]: So with the book, I had a couple goals. One was to give people just a working vocabulary for the space as a whole, and the other was to give them some intuition about how each of these techniques works. As I mentioned in my AI Engineer talk, which is the first public addendum to this, the speculation space has moved much faster than everything else. So yeah, even at the time that I wrote the book Medusa, I very much included as a way for people to understand how the space evolved rather than what the most modern technique is. And now of course, there's DFlash, dSpark. There's, there's newer techniques even than EAGLE, although EAGLE is still very commonly used.Ali [00:43:51]: SpecSpecta.Philip [00:43:52]: Yes. Speculative decoding.Vibhu [00:43:54]: What canAli [00:43:56]: Oh, it's a paper by Tri Dao and it's like, it's doing speculative decodingVibhu [00:44:00]: HuhAli [00:44:01]: For the speculative decoder.Philip [00:44:02]: Oh, in spec- oh my God.Ali [00:44:02]: It's literally just an another. It's like, yeah, that's the most simple way to explain it, and it seems like he got trivial speed ups there. But it seems that the complexity with training, it's almost like in our mind at least, it's almost as complex as training GANs. Like it's like a very delicate balance and oftentimes you, it's just but yeah, it's literally speculative decoding on speculative decoding.Vibhu [00:44:21]: Speculative.Ali [00:44:22]: Yeah. We saw this paper.Vibhu [00:44:24]: It's interesting, right?Ali [00:44:24]: Yeah.Vibhu [00:44:24]: I wouldn't even expect it to be very particular to train, I wouldAli [00:44:29]: Right.Vibhu [00:44:29]: The naive part of me is like, okay, train speculative decoder.Ali [00:44:32]: But like, and it makes sense, like the whole idea of speculative decoding is you. It's like, it's like almost like the iPhone auto predict version but for a normal model, right? Like you're just, you're just, generating three tokens and you're like, okay, I'll do prefill on them. And so you save those three turns for your original model. Now your speculative decoder is doing three turns of auto regression, so why not just have an even smaller model?Ali [00:44:53]: The other question there is what are the size of speculators? So say forPhilip [00:44:58]: Right. It's like a billion parameters.Ali [00:45:01]: Like for MiniMax, it's. Yeah. It's like one layer. It's like one 60th of the original model usually.Philip [00:45:06]: Yeah. I think we should do a paper when we get back to the office.Philip [00:45:10]: SpeculativeAli [00:45:11]: SpeculativePhilip [00:45:11]: Decoding.Ali [00:45:13]: No, it's, it does seem like how, when do you stop? But then it also seems like if you're able to train spec-spec decode for instance, right? Like if you're able to have a small model that is accurately predicts what the intermediate speculator is gonna predict, that is able to predict what the original target model's gonna predict, then why not just use that smallest model directly, right?Vibhu [00:45:34]: Yeah. This isAli [00:45:35]: Like it seems likeVibhu [00:45:35]: Adjacent to the routing problem.Ali [00:45:36]: Right.Vibhu [00:45:36]: Yeah.Ali [00:45:36]: Right.Philip [00:45:37]: The thing with speculators is one of the practical constraints on using them is that you do have to run a small model on the same hardware that you're running the big model on. There is a orchestration and resource competition problem inherent in that, and that is one of the constraints on speculation in general, is that draft tokens cost resources to create and cost software complexity to manage. And so if you have like infinitely recursive speculators, you add in quite a bit of that complexity on the actual implementation within the inference engine as well, not just in the training process.Vibhu [00:46:17]: I was gonna say, I would wonder if you could do similar, like distillation and pruning of, it's the same thing, it's just a model. Can we not just distill a lot of the weights, quantize the speculator, out of my domain? The question that also comes up is, this is all for big server workloads, right? How much of this applies to, say I have this MacBook, I wanna run Gemma really efficiently. Similar problems, not the same?Local AI vs. Data Center InferencePhilip [00:46:45]: Pretty different. I talked to Selo, about this on his podcast a couple weeks ago. The difference between inference engineering for the data center and for production workloads versus inference engineering for local AI, is that we start with fundamentally like different constraints and different goals. With local AI, it's how do I fit this model onto my hardware and then make it less dumb? And with data center influence, it's how do I load this model and then make it less slow? And we care about less dumb, and they care about less slow. But the local AI inference engineering ecosystem, I think has a lot for us to learn from in the data center space. They are experts in various forms of quantization, including dynamic quantization that we just don't touch, in the pruning, in the distillation, in the, layer removal. There'Ali [00:47:42]: Layer removal matters less.Philip [00:47:43]: Yeah. There'Ali [00:47:44]: No one loves pruning really.Philip [00:47:45]: Yeah. Well, but the, but they doVibhu [00:47:46]: Which is surprising, right? But that's, that's a whole different thingPhilip [00:47:48]: Just to fit something on the laptop.Ali [00:47:50]: Right.Philip [00:47:50]: So yeah, it's a, it's an interesting, it's an interesting space. Not necessarily that like their techniques make sense for us to do in the data center, because we have different resources and different goals, but more that the process as well as the openness of that field is something to, admire.Ali [00:48:12]: Yeah. Like to your point, like, certain optimizations that would. Like for instance, Turbo Quantum Sharper, like it made such huge hype on that and we did like a whole deep dive on Twitter and like said, what is it? How does it work? Why is it good or not? And it took off and it was implemented on local devices because your memory bandwidth is so slow on like a MacBook, for instance. But try putting the same thing on like an NVIDIA GPU on a B200 Turbo quant would not be. Like, it would not be used. Like, NVIDIA - Like, NVIDIA made it clear that this is not a good optimization, and we've seen it firsthand where the overhead of doing dequantization, quantization of, in the kernel itself with turbo quant kernel, each end is much slower than the time that you save from doing the bandwidth. ‘Cause on the B200s, you have like 3.5 terabytes per second. You don't need decrease the storage that much. You don't need to do, FP4 KV cache. You don't need to use a requant. There's, there's, there's better optimizations to be made. But on Edge devices, it's extremely important, it's extremely useful. So, seems to be, like, different optimizations there, but then they're all uniquely combined with like all you wanna quantize the model, you wanna do speculative decoding, like certain common prefixes with bothPhilip [00:49:18]: Principles.Ali [00:49:19]: Yeah, exactly. Exactly. Exactly.Philip [00:49:20]: They also do a lot of work on, model parallelism, especially over, heterogeneous topology, where you have, some sparks and they are wired together with, Ethernet, DGX sparks.Ali [00:49:35]: Yeah, this is the Exo Labs guys.Philip [00:49:36]: Yeah. You have, a nu
Guvna B is joined by comedian and footballer, Sapphire McIntosh, to discuss her experiences as a Black woman from Leeds as well as exploring the Black media ecosystem with hosts of the Rigour & Flow podcast, Aiwan Obinyan and Tamanda Walker. @1Xtra on socialProduced by Unedited for BBC Radio 1Xtra.
Hour 3 for 7/31/26 Drew and Peter Grandich cover the importance of faith during anxiety and depression (1:00). Calls: the need for Divine Providence (16:56), investment (21:16), and getting outside (24:23). Then, Mark J. Warshawsky from AEI covers the new bill aimed at preserving Social Security (27:45). Topics: allocating cuts (42:16), Ponzi scheme (45:29), fraud (46:41), and Medicare for all (48:21). Links: https://petergrandich.com/ https://x.com/petergrandich https://www.aei.org/profile/mark-j-warshawsky/
Mentorship in Motion: The Power of Partnership Sandra Pretari-Hickson joins host Laura Reeves to discuss her 25-year partnership with Betty-Anne Stenmark and how their shared devotion and collaborative mentorship have preserved the iconic King's Mountain Dandie Dinmont Terrier legacy. This episode of the Dandie Dinmont Terrier Diaries is part of Pure Dog Talk's Clubhouse. The series features members of the Club Partnership Program who are working to preserve the knowledge of their legacy breeders. Episode Highlights The Spark of a Historic Partnership: Sandra shares the story of meeting Betty-Anne at a handling class, a chance encounter that led to Betty-Anne gifting Sandra her very first Dandie, "Kings Mountain Mousetrap," and cementing a 25-year breeding partnership.A Shared Vision for a Distinct Type: Working hand-in-hand, Sandra and Betty-Anne established an internationally recognized breed type. Sandra explains their 50-year joint commitment to never compromising on correct front assemblies, classic heads, athletic movement, and the signature long outline.Navigating the "Gene Puddle" Together: Sandra discusses managing a rare breed's limited gene pool (coined a "gene puddle" by Betty-Anne). By combining their expertise, they utilized tight line breeding, imported frozen semen from across the globe, and coordinated artificial inseminations to safely preserve the line.Honest Evaluation and Ruthless Selection: Mentorship requires tough, honest decisions. Sandra details how she and Betty-Anne evaluate slow-maturing puppies at 10 weeks and why maintaining breed integrity means placing dogs with significant faults into happy companion homes.Preserving the "Stand-Up" Terrier Temperament: Sandra emphasizes the importance of protecting the athletic, working drive of the Dandie Dinmont. She shares her dedication to keeping the "terrier in the terrier" through active socialization, barn hunt sports, and earthdog trials.Passing the Torch to the Next Generation: True partnership extends beyond a single lifetime. Sandra highlights how she is actively paying her mentorship forward by guiding two to three promising young breeders to ensure the King's Mountain legacy continues to thrive. For more information about joining Pure Dog Talk's Clubhouse Series and becoming a Club Partner, stop by the website.
Editor, director, and producer Whit Conway explains how the Hulu comedy Never Change! was assembled on an unusually compressed post-production schedule. He discusses cutting scenes while production was still underway, building an organized Adobe Premiere Pro workflow, using searchable transcripts to locate alternate performances, and collaborating with director and co-editor Marty Schousboe. Conway also shares how constant iteration shapes comedy, why editors cannot become precious about an early cut, and how aspiring comedy filmmakers can create the work they want to be hired to make. In this episode, No Film School's GG Hawkins and guest Whit Conway discuss... How Whit's previous collaborations with director Marty Schousboe led to Never Change! Why the film's budget and delivery schedule required editing to begin during production Completing an initial assembly within a week of production wrapping Applying lessons from the fast-paced editorial workflows of Saturday Night Live and Joe Pera Talks With You Organizing footage into scene sequences and line-by-line collections of every performance How front-loading editorial organization saves time when producers request alternate takes Treating the first cut as a strong starting point rather than a version that must be protected Adjusting individual jokes by only a few frames and testing different reaction shots Preserving previous versions of scenes without formally tracking every editorial change Recognizing when a joke needs more work versus when the filmmakers have simply watched it too many times Using Premiere Pro's searchable speech-to-text transcripts to find improvised lines and specific takes Collaborating with directors who edit and directors who communicate their ideas conceptually Why strong directors know what they want while remaining open to better ideas from collaborators Managing creative disagreements, competing producer notes, and compressed deadlines How walks and time away from the edit room can help collaborators reset Why budget limitations ultimately determine when the filmmakers must put their pencils down Building a comedy-editing career by making long-form work, studying classic comedies, and creating projects with friends Treating unpaid personal projects as “extra credit” that can lead to new professional opportunities Memorable Quotes: “It is just my job to put the first version of the moment together as fast as I can and also be tickled by it.” “I think great directors do a couple of things. One, I think they know what they want. They know what the moment should be. Two, they are open and willing to see other versions of that moment.” “I think you don't have to have all the answers all the time.” “People understand you and know you through the work that you have done, that they have seen you do.” Guests: Whit Conway Resources: Watch Never Change! on Hulu Whit Conway's Website Adobe Premiere Speech to Text How These SNL Editors Cut an Emmy-Nominated Sketch With 60+ VFX Shots in 2 Days Cutting Comedy: Inside the Edit of Saturday Night Live's 50th Anniversary Field Producing on Last Week Tonight & Tackling Vimeo's EU/UK Problem Find No Film School everywhere: On the Web: No Film School Facebook: No Film School on Facebook Twitter: No Film School on Twitter YouTube: No Film School on YouTube Instagram: No Film School on Instagram
Mission drift, the wrong measure, the trade-in trap—and a little fresh-squeezed freedom. Plus, Malcolm Guite on writing poetry one walk at a time, Daniel Suhr on election integrity, and the Thursday morning newsSupport The World and Everything in It today at wng.org/donateAdditional support comes from Ambassadors Impact Network. Christian entrepreneurs scaling their companies often struggle to find capital sources that respect, and even celebrate, their redemptive mission. Ambassadors Impact Network connects these founders with angel investors seeking both financial outcomes and spiritual fruit. If you're an investor wanting to steward your capital toward gospel-advancing companies, learn about membership at ambassadorsimpact.comFrom Cedarville University. Located in southwest Ohio, Cedarville University is committed to biblical faithfulness and academic excellence, preparing students to serve Christ with conviction in every profession. Every one of its 175+ undergraduate and graduate programs are grounded in a biblical worldview, equipping graduates with the knowledge, skill, and conviction to serve wherever God leads. New online undergraduate degrees through Cedarville Online offer flexible, affordable education rooted in biblical truth and designed for today's learners. Learn more at cedarville.edu, and explore online programs at cedarville.edu/online.And from Equip by Unbound. Launching high schoolers through coaching, skills training, and interest-led projects. More at BeUnbound.us/world
On this edition of The Federalist Radio Hour, Founder and CEO of American Majority Ned Ryun joins Federalist Senior Elections Correspondent Matt Kittle to discuss virtue as the organizing factor in early America, the founders' concept of self-government versus the modern administrative state, and the plan for Republicans to reclaim institutional power through the upcoming election cycles.You can find Ned Ryun's new documentary, The Thread of Liberty: Keeping Our Republic, here.The Federalist Foundation is a nonprofit, and we depend entirely on our listeners and readers — not corporations. If you value fearless, independent journalism, please consider a tax-deductible gift today at TheFederalist.com/donate. Your support keeps us going.
© 2020-2026 Ivy Podcast Discovery LLC
