Podcasts about Blackwell

Send us Fan MailIn Episode 267 of Book Talk Etc., Tina and Hannah took a week to do some mood reading, and for book talk, we answered questions from our listeners in an ask us anything style discussion.If you enjoy this commercial-free podcast, consider supporting us on Patreon! Your membership includes access to bonus episodes like What's in the Mailbag, Bookstore Browse: The Handsell, and Book Talk After Dark, invites to monthly community events like Mood Reader Happy Hour, and entry into our private Facebook group and Discord server- all for just $5 a month.Loving LatelyThe Obama Foundation & Museum (T)Sterling Point - Prime Video (H)Latest ReadThe Unknown | Riley Sager (T+H)Mood ReadsWonderland | Jennifer Hillier (T)The Dog Stars | Peter Heller (H)Mind of a Murderer | Michael Wood (T)Let's Kiss & Tell | Joss Richard (H)Current ReadThe Only Plane in the Sky (T)A Plagued Sea | Kim Bo-Young (H)If you prefer other shopping options, you can find today's books on Bookshop.org or Blackwell's. Purchasing through these links supports us with a small commission, at no extra cost to you.Support the showLet's Connect... Email us at booktalketc@gmailBTE on YoutubeTina's TikTok , IG @tbretc YT @tbretcHannah's TikTok , IG @hanpickedbooksJonathan IG @infiltrate_jayPodcast IG @booktalketcRenee's Substack Newsletter , IG@Itsbooktalk

This week, I first discuss the latest trending news in travel, including Google's AI Mode for Travel, Europe's water-level updates, and more. Later, I interview Phil Blackwell, VP, Business Development & Partnerships for Travel Insured International (TII). Blackwell shares why TII invests in travel advisors and how two Mastermind events this year are helping shape how they work with advisors. Additionally, Blackwell discusses the importance of the advisor-supplier relationships and business updates from TII. The interview with Blackwell begins after the 9-minute mark. Today's episode sponsor: Travel Insured International At Travel Insured International, we believe that power lies in partnership. And that means giving you the tools to streamline your workflow and help protect your reputation. Gain more control over your business (and its bottom line) with a custom Advisor Dashboard featuring a quote manager, commission and conversion tracking, and payment alerts. Plus, our Certified Specialist Program gives you the confidence to offer premium protection as a standard offering. Your clients aren’t alone. Now, neither are you. Visit TravelInsured.com. Where our people become your people. Have any feedback or questions? Want to sponsor the show? Contact us at Podcast@TravelPulse.com and follow us on social media @TravelPulse.See omnystudio.com/listener for privacy information.

Nvidia delivered another blockbuster quarter, reporting record revenue of $96.2 billion, up 106% from a year ago, as demand for AI infrastructure continues to accelerate. But the bigger story may be what comes next. Lance Roberts & Michael Lebowitz break down Nvidia's latest earnings, its extraordinary growth outlook, Blackwell and Vera Rubin demand, the massive buildout in AI infrastructure, and Nvidia's growing financial involvement across the AI ecosystem. We also examine the risks, including rising memory costs, margin pressure, enormous capital requirements, and questions about whether today's AI spending boom can deliver adequate returns. 0:00 INTRO 1:00 - Economic Recap & Multiplier Effect 6:31 - Market Push from NASDAQ 11:45 - Passwords, Needles & Rosso's toys 14:30 - NVIDIA Reports - Markets Respond 21:36 - How Ancillary Business Benefit from NVIDIA 24:03 - NVIDIA Stock Performance 25:32 - The Heat Map Game 27:25 - Portfolio Positioning & Nat Gas play 30:37 - The Economics of Gilligan's Island 35:33 - Productivity Factors of AI 38:04 - How Much Return on AI Investment will there be 39:57 - How Will AI be Transformative? 42:42 - We're All Investing in AI Hosted by RIA Advisors' Chief Investment Strategist, Lance Roberts, CIO, w Portfolio Manager, Michael Lebowitz, CFA Produced by Brent Clanton, Executive Producer ------- Do you enjoy our content? Rate us on Google: https://bit.ly/4b9JtEo ------- Watch Today's Full Video on our YouTube Channel: https://youtube.com/live/lGSWXLw9dPY ------- Articles mentioned in this report: "Fueling AI Data Centers: Behind The Meter Solutions- Part 1" https://realinvestmentadvice.com/resources/blog/fueling-ai-data-centers-behind-the-meter-solutions-part-1/ "Behind The Meter Solutions Investment Guide- Part 2" https://realinvestmentadvice.com/resources/blog/behind-the-meter-solutions-investment-guide-part-2/ "Productivity On Gilligan's Island: Episode 2" https://realinvestmentadvice.com/resources/blog/productivity-on-gilligans-island-episode-2/ "Bitcoin Up 22%: Has The Halving Cycle Begun?" https://realinvestmentadvice.com/resources/blog/bitcoin-up-22-has-the-halving-cycle-begun/ --- Watch today's "Before the Bell" report, "NASDAQ Technicals Are Turning Higher," https://youtu.be/33KMo6HoezY ------- Watch our previous show, "Are Today's Interest Rates Really That High?" https://youtube.com/live/lPCVe6O4LjM ------- Get more info & commentary: https://realinvestmentadvice.com/insights/real-investment-daily/ ------- * REGISTER for our next Dynamic Learning Series, "The Smart Way to Pay for College," Thursday, September 3, 2026: https://streamyard.com/watch/mcE7YgphgMns --- Visit our Site: https://www.realinvestmentadvice.com Contact Us: 1-855-RIA-PLAN --- Subscribe to SimpleVisor : https://www.simplevisor.com/register-new --- Connect with us on social: https://twitter.com/RealInvAdvice https://twitter.com/LanceRoberts https://www.facebook.com/RealInvestmentAdvice/ https://www.linkedin.com/in/realinvestmentadvice/ #NASDAQ #StockMarket #TechnologyStocks #MarketOutlook #Investing #Nvidia #NVDA #NvidiaEarnings #AIStocks

Things are getting weird – looks like someone is panicking. The Dumber side of finance – let’s dig in. A big week ahead – major earnings could move markets. – Gold, Bitcoin moving on a lower USD. PLUS we are now on Spotify and Amazon Music/Podcasts! Click HERE for Show Notes and Links DHUnplugged is now streaming live - with listener chat. Click on link on the right sidebar. Love the Show? Then how about a Donation? PayPal.Donation.Button({ env:'production', hosted_button_id:'JJJHP2GDEJC7J', image: { src:'https://www.paypalobjects.com/en_US/i/btn/btn_donateCC_LG.gif', alt:'Donate with PayPal button', title:'PayPal - The safer, easier way to pay online!', } }).render('#donate-button'); Follow John S. Dvorak on X Follow Andrew Horowitz on X Warm-Up - Things are getting weird - looks like someone is panicking - The Dumber side of finance - let's dig in (5 top items) - Cold War and Hot War - Canada, China and Iran - Economic "D" day - no one cares - Saying goodbye to Dolly Parton Markets - Gold, Bitcoin moving on a lower USD - Oil starting to settle - Big Week - NVDA earnings on tap - all eyes! STUPID IS WHAT STUPID DOES - A DHU One-Time Series On The Dumber Side of Finance 1) LOWER BEEF PRICES - Trump plans temporary tariff relief for certain ground-beef imports as U.S. beef prices remain near record highs. - The plan would allow up to 300,000 metric tons of discounted imported ground beef over roughly 90 days. - Supposedly this will reduce the prices for consumers - Ranchers and several farm-state Republicans are pushing back, arguing cheaper imports could hurt domestic producers. ---- Question: Tariff reduction is going to reduce prices for consumers? Didn't someone say that tariffs are paid by the companies and exporting countries??? 2) IRAN SANCTIONS - BIG THREAT, LIMITED IMMEDIATE ACTION - Treasury warned countries doing business with Iran that they could face secondary sanctions, calling it an "economic D-Day." - New sanctions hit about 60 individuals, companies and vessels, but major Chinese financial institutions were not targeted. - China remains the biggest buyer of Iranian oil, making Chinese banks the potential pressure point. - Treasury says sanctions could expand into gold, digital assets, aviation, shipping and other sectors. 3) TREASURY TRIES TO TAME LONG-TERM YIELDS - Treasury will at least double the maximum size of certain long-term bond buybacks to $4 billion per operation. - The announcement briefly knocked nearly 10 basis points off the 30-year Treasury yield. - Critics say this treats the symptom rather than the cause - huge deficits, heavy issuance and persistent inflation risk. ------Bessent had to come out on Friday to say "at least" $4B - as bind yields moved back up -------Maybe there is a concern over the $40TRILLION debt we have amassed as a country? 4) SOMEONE GOT LONG BONDS AT THE RIGHT TIME - TLT took in roughly $529 million of net inflows on August 18, one day before Treasury's surprise bond-buyback announcement. - The next day Bessent doubled planned buybacks of 10- to 30-year Treasurys, immediately pushing long yields lower. - TLT jumped about 1.7% on the announcement. - No public evidence identifies a single buyer - but the timing of the unusually large inflow is certainly worth noting. 5) DOLLAR GETS HIT BY THE BOND RESCUE - The dollar weakened as investors questioned Treasury's expanded bond-buyback strategy. - The concern is that suppressing yields without fixing fiscal problems simply transfers pressure from bonds to the currency. - The 30-year yield has remained elevated despite Treasury intervention. - Bitcoin rallied as investors looked toward alternatives to traditional fiat assets. - PEOPLE: This is NOT Quantitative Easing - Just a move to buy longer dated and issue more short dated WALMART FLASHES A CONSUMER WARNING - Walmart U.S. comparable sales grew just 2.6%, its slowest pace in roughly six years and below expectations. - Shares dropped sharply after the report. - E-commerce remained strong, while store traffic and discretionary spending showed more weakness. - Walmart raised full-year guidance but gave a softer-than-expected third-quarter outlook. - As America's biggest retailer, Walmart's slowdown is an important read on the broader consumer. VENEZUELA HAS OIL - BUT CAN'T SHIP IT FAST ENOUGH - Tankers are waiting as long as 30 days to load Venezuelan crude because of aging ports, outages and equipment failures. - Venezuela has struggled to push exports much higher despite rising production and strong demand. - The Jose terminal handles about 70% of exports and has become a major bottleneck. - Venezuela's ports may impose a physical ceiling on production growth until infrastructure is upgraded. JANE STREET'S $15 BILLION AI HIT - Jane Street reportedly lost about $15 billion in July as AI and technology positions moved sharply against the firm. - A major source of the damage was exposure to concentrated AI trades that were forced to unwind. - Jane Street posted its first negative month of trading revenue since 2016. - As a reminder: this episode highlights how quickly crowded AI trades can overwhelm even sophisticated risk-management systems. LEAVITT LEAVES THE WHITE HOUSE - Karoline Leavitt resigned as White House press secretary after returning from maternity leave. - Trump said she would become a top outside adviser and prominent Republican voice heading into the midterms. - Looking fior a kinder and gentler replacement? HA! probably not OFF THE HIGHS NVDA - 9% off high AMD - 19% off high AVGO - 23% off high MU - 23% off high TSM - 13% off high MRVL - 28% off high INTC - 35% off high NVIDIA - THE MARKET'S NEXT STRESS TEST - Nvidia reports Wednesday - now basically a referendum on the entire AI trade. - Blackwell demand and hyperscaler spending are the key tells. - A strong guide could quickly reset tech sentiment. - A miss would raise the uncomfortable question: how much AI optimism is already priced in? - Revenue expected around $92.1 billion, up roughly 97% year over year. - Adjusted EPS expected around $2.09, nearly double last year. - Data Center revenue expected around $85.7 billion - still the main engine. - Biggest watch items: Blackwell demand, Vera Rubin timing, China sales and hyperscaler capex. - Options imply roughly a 5%-6% move after earnings. - Discussion .... NVIDIA - AI SERVERS GETTING EVEN MORE EXPENSIVE - Major Nvidia customers have reportedly been warned that AI server prices could rise more than 15% beginning early next year. - Higher memory costs are the main driver, affecting Grace Blackwell and upcoming Vera Rubin systems. - Nvidia is effectively passing higher component costs through despite gross margins around 75%. - ALWAYS INTERESTING THESE ANNOUNCEMENTS SO CLOSE TO EARNINGS - HMMMM CANADA-U.S. TRADE FIGHT GETS WORSE - Canada announced retaliatory tariffs on about $20 billion of U.S. goods after Washington imposed new 50% tariffs on Canadian imports. - Canada's tariffs range from 15% to 50% across roughly 700 products and begin September 8. - Canada also announced a C$7.5 billion support package for affected businesses and workers. - Another escalation that could hit autos, manufacturing costs and cross-border supply chains. HOUSING - BUYERS KEEP DISAPPEARING - July new-home sales plunged 10.5% to a 607,000 annual rate, the weakest since January. - Median new-home price fell to $393,800, the lowest in four years. - Mortgage rates are still around 6.8%, keeping affordability under pressure. - Only 5.2% of consumers said they expect to buy a home in the next six months - the sharpest drop in more than five years. JACKSON HOLE - WARSH GETS THE MICROPHONE - Kevin Warsh speaks next week at Jackson Hole. - Markets want to know whether sticky inflation, oil and tariffs are enough to keep the Fed in tightening mode. - Long yields are already doing some of the Fed's work. - One sentence could move bonds, the dollar and stocks. - Friday, August 28 at 10:00 a.m. ET. --- BUT - Didn't Warsh say less communication is more? They can't keep put of the spotlight. PCE - THE FED'S FAVORITE INFLATION READ - "The Fed's preferred gauge/measure of inflation" - July PCE lands next week. - Watch for tariff, energy and goods inflation creeping back into the numbers. - A hot print revives hike fears. - A soft print gives risk assets some breathing room - especially with long yields already elevated. THE CONSUMER VS. THE AI BOOM - MORE EARNINGS COMING - Salesforce, CrowdStrike and Marvell give another read on corporate AI spending. - Dollar General, Dollar Tree, Best Buy and Ulta test the other side of the economy. - The setup is getting interesting: companies are still spending aggressively on AI while consumers look increasingly selective. - If that gap widens, it could become one of the bigger market themes into the fall. AROUND THE WORLD EUROPE - QUIETLY GETTING INTERESTING - European stocks slipped this week, but money has started flowing back into the region. - Euro-zone business activity is growing at its fastest pace this year, while Q2 earnings growth for STOXX 600 companies is running surprisingly strong. - CONCEPT: Europe may be turning into the anti-U.S. trade - less AI concentration, cheaper valuations, but more sensitivity to energy. JAPAN - BETTER DATA, WORSE MARKET - Japan's Nikkei lost about 4% this week even as the economic data improved. - August manufacturing PMI jumped to 55.1, with new orders rising at the fastest pace since 2018. - Semiconductors and AI-related demand are helping drive the factory rebound. - Stronger growth also keeps the BOJ rate-hike discussion alive - good economy, potentially tougher market. BIG PHARMA - CHINA'S WEIGHT-LOSS GOLD RUSH - Lilly, Novo Nordisk and others are aggressively targeting China's obesity market, where obesity rates are projected to top 65% by 2030. - China bans direct prescription-drug advertising, so companies are using subway ads, gyms, influencers and "disease awareness" campaigns instead. - China's GLP-1 market could reach roughly 30 billion yuan over the next 5-7 years. - Interesting regulatory gray area: education campaigns that look a lot like drug advertising without actually naming the drug. Love the Show? Then how about a Donation? PayPal.Donation.Button({ env: 'production', hosted_button_id: 'JJJHP2GDEJC7J', image: { src: 'https://www.paypalobjects.com/en_US/i/btn/btn_donateCC_LG.gif', alt: 'Donate with PayPal button', title: 'PayPal - The safer, easier way to pay online!' } }).render('#donate-button-2'); THE CLOSEST TO THE PIN for SpaceX (SPCX) Winners will be getting great stuff like the new "OFFICIAL" DHUnplugged Shirt!   FED AND CRYPTO LIMERICKS   See this week's stock picks HERE Follow John C. Dvorak on Twitter Follow Andrew Horowitz on Twitter

In this episode, Adrien Blackwell talks about how to Stop Predicting and Start Manifesting Your Future. Adrien Blackwell spent more than a decade as a celebrity psychic before realizing something that completely changed the direction of her work: knowing what was going to happen wasn’t nearly as powerful as helping someone change what happens next. In addition to psychic-healer, she is also the host of the Making Miracles Happen Masterclass Podcast, where healing gets curious, spirituality gets challenged, and conversations with doctors, researchers, healers, and extraordinary thinkers explore what’s actually possible when we stop accepting our past as a prediction of our future. Adrien’s philosophy? You weren’t put here just to survive your life, manage your patterns, or make peace with less than you want. You can have it all, or get pretty damn close. For More Information ★ To learn more about Adrien Blackwell visit her website: AdrienBlackwell.com★ If you enjoyed the show, please leave us a five star iTunes review. Visit Spiritual Rockstar Podcast at https://yoursacredpurpose.com/ for more information!★ I encourage you to join our Rock Your Sacred Purpose Community on Facebook: https://www.facebook.com/groups/246228169428755★ Do you want to Meditate and Make Money? Grab your Free meditation today: YourSacredPurpose.com Show Notes ★ 2:55 – When I was 6, my sister wanted to go over to her friend’s house and I was like ‘don’t let her go because something bad is going to happen’.★ 9:06 – Everything that I got was 100%, so to me the logic is that I’m reading the book, for him he goes ‘you know you are psychic, right?’.★ 25:41 – I feel like there are parts of the future that are fixed, and most of the future is not fixed.★ 32:57 – You could change, but you are so stuck in a belief system, in the hurt, and in old energy that you actually won’t allow yourself to experience anything outside of your belief system. And so, your belief system becomes a self-fulfilling prophecy.★ 41:15 – Sometimes you actually need to change the brain, so I have techniques that actually change beliefs systems in a matter of minutes.★ 48:04 – I want everybody to know one thing …★ 52:57 – Listen to the Making Miracles Happen Masterclass Podcast here: AdrienBlackwell.com★ 53:19 – FREE GIFT – Sign up for the Making Miracles Happen Masterclass LIVE – A free online spiritual masterclass featuring powerful teachers, healers and experts here: https://makingmiracleshappenseries.com/★ 1:02:04 – It has been like that ever since, where she just wants me to know she is here. So, thank you for the message because I know she is always around.★ 1:03:38 – Do you want to Meditate and Make Money? Grab your Free meditation today: https://www.YourSacredPurpose.com★ 1:04:45 – You already love yourself, continue to do so. Listen to the Show The post 548: Adrien Blackwell – Stop Predicting and Start Manifesting Your Future appeared first on Your Sacred Purpose.

Send us Fan MailIn Episode 265 of Book Talk Etc., Tina and Hannah discuss books that feel like Lughnasadh, a Celtish holiday that celebrates the start of the harvest season and falls halfway between the summer solstice and the autumn equinox.If you enjoy this commercial-free podcast, consider supporting us on Patreon! Your membership includes access to bonus episodes like What's in the Mailbag, Niche Novels, and Book Talk After Dark. You'll get to join monthly community events like Mood Reader Happy Hour, and entry into our private Facebook group and Discord server- all for just $5 a month!Loving Lately@readandwright Notes App System (T)Lifelines - Dot by Letter (H)Lifelines WebsiteLatest ReadFruit Fly | Josh Silver (T)Return to Sender | Stephanie Parente (H)Book Talk and Reading for LughnasadhTom Lake | Ann PatchettThe Wild Beneath | Kelly Anderson (T)Habits of the Sea | Shea Ernshaw (H)Where I End | Sophie White (T)Jane Eyre | Charlotte Bronte (H)Shelf AdditionTokyo Express | Seicho Matsumoto (T)A Quiet Place | Seicho Matsumoto (T)Suspicion | Seicho Matsumoto (T)Too Much of Life | Clarice Lispector (H)If you prefer other shopping options, you can find today's books on Bookshop.org or Blackwell's. Purchasing through these links supports us with a small commission, at no extra cost to you.Support the showLet's Connect... Email us at booktalketc@gmailBTE on YoutubeTina's TikTok , IG @tbretc YT @tbretcHannah's TikTok , IG @hanpickedbooksJonathan IG @infiltrate_jayPodcast IG @booktalketcRenee's Substack Newsletter , IG@Itsbooktalk

College campuses can feel like hostile territory for students who abide by their Christian faith and values. On today's edition of Family Talk, Dr. James Dobson continues his insightful conversation with Morton Blackwell, founder of the Leadership Institute. Blackwell reveals how to help young believers find like-minded community, stand firm in their convictions, and become confident leaders instead of feeling outnumbered. To support this ministry financially, visit: https://www.oneplace.com/donate/707/29?v=20251111

On today's edition of Family Talk, Dr. James Dobson welcomes Morton Blackwell, founder of the Leadership Institute, to expose the pressures Christian and conservative students face on today's college campuses. Blackwell shares how his organization fights back through Campus Reform, uncovering abuses that don't survive the light of day. To support this ministry financially, visit: https://www.oneplace.com/donate/707/29?v=20251111

デルがタワー型ワークステーション「Dell Pro Precision」3製品を発売　Xeon 600やRTX Pro Blackwellを搭載可能。　デル・テクノロジーズは8月12日、ワークステーション「Dell Pro Precision」シリーズのデスクトップタイプの新製品として「Dell Pro Precision 9 T2」「Dell Pro Precision 9 T4」「Dell Pro Precision 9 T6」の3製品の販売を開始した。

Send us Fan MailIn Episode 264 of Book Talk Etc., Tina and Hannah tackle the TBR shelf and read books from their personal home libraries in an effort to kick those books off their TBRs. Every month, Tina and Hannah spin the wheel, and randomly select a patron selected prompt. This month, we read books from our TBR shelf that we found intimidating!If you enjoy this commercial-free podcast, consider supporting us on Patreon! Your membership includes access to bonus episodes like What's in the Mailbag, Bookstore Browse: The Handsell, and Book Talk After Dark, invites to monthly community events like Mood Reader Happy Hour, and entry into our private Facebook group and Discord server- all for just $5 a month.Loving LatelyJuley.app (T)The Commons | Rebecca IhimWelly Bandages (H)Latest ReadThe Elementals | Michael McDowell (T)Nevermoor: The Trials of Morrigan Crow | Jessica Townsend (H)Intimidating BooksThe Lion Women of Tehran | Marjan Kamali (T)The Stationery Shop | Marjan KamaliThe Calamity Club | Kathryn Stockett (H)East of Eden | John Steinbeck (T)East of Eden - Penguin Orange Classic EditionGirl on Girl | Sophie Gilbert (H)Shelf AdditionThe Silent Appeal | Janice Hallett (T)Hive Mind | Allison Gunn (H)If you prefer other shopping options, you can find today's books on Bookshop.org or Blackwell's. Purchasing through these links supports us with a small commission, at no extra cost to you.Support the showLet's Connect... Email us at booktalketc@gmailBTE on YoutubeTina's TikTok , IG @tbretc YT @tbretcHannah's TikTok , IG @hanpickedbooksJonathan IG @infiltrate_jayPodcast IG @booktalketcRenee's Substack Newsletter , IG@Itsbooktalk

Send us Fan MailIn Episode 263 of Book Talk Etc., Tina and Hannah share tips on how they remember the books they read, as well as share some thoughts on new book releases that they've read recently!If you enjoy this commercial-free podcast, consider supporting us on Patreon! Your membership includes access to bonus episodes like What's in the Mailbag, Bookstore Browse: The Handsell, and Book Talk After Dark, invites to monthly community events like Mood Reader Happy Hour, and entry into our private Facebook group and Discord server- all for just $5 a month.Loving LatelyMellow Comforter (T)Paper Meadow Mail Club (H)Opal Bones Mail Club (H)Jimothy Collection (H)Latest ReadDie for Me | Shirlene Obuobi (T)Sookie Stackhouse BooksSeatmate | Cara Bastone (H)New ReleasesThe Lowe Job | Grace Alexander (T)Magnolia Parks | Jessa Hastings Teddy Bears Never Die | Yeeun Cho (H)The Open Era | Edward Schmit (T)The Children | Melissa Albert (H)Shelf AdditionIn Cold Blood | Truman Capote (T)Something Followed Us Home | Cynthia Pelayo (H)If you prefer other shopping options, you can find today's books on Bookshop.org or Blackwell's. Purchasing through these links supports us with a small commission, at no extra cost to you.Support the showLet's Connect... Email us at booktalketc@gmailBTE on YoutubeTina's TikTok , IG @tbretc YT @tbretcHannah's TikTok , IG @hanpickedbooksJonathan IG @infiltrate_jayPodcast IG @booktalketcRenee's Substack Newsletter , IG@Itsbooktalk

Merneptah vs the Libyans. In 1221 BCE, Egypt faced a crisis. Tribal coalitions from the west, and some "Peoples of the Sea" were marching on the Nile Valley. They came in their thousands, armed and ready to settle. Caught by surprise, King Merneptah and his warriors hastened to the defence... Support the show and get early releases, ad-free episodes, bonus video content, and pictures/videos of animals I meet in Egypt and travelling www.patreon.com/egyptpodcast Music by Bettina Joy de Guzman; Luke Chaos; and Keith Zizza. Bibliography Cline, E. H. (2021). 1177 B.C. The Year Civilization Collapsed (Revised edn). Princeton University Press. D'Amato, R., & Salimbeti, Andrea. (2015). Sea Peoples of the Bronze Age Mediterranean c.1400 BC–1000 BC. Osprey. Darnell, J. C., & Manassa, C. M. (2007). Tutankhamun's armies: Battle and conquest during ancient Egypt's late eighteenth dynasty. John Wiley & Sons. Davies, B. G. (1997). Egyptian Historical Inscriptions of the Nineteenth Dynasty. Paul Aströms förlag. Dean, R. (2017). Warfare & Weaponry in Dynastic Egypt. Pen & Sword. Finné, M., Woodbridge, J., Labuhn, I., & Roberts, N. (2019). Holocene hydro-climatic variability in the Mediterranean: A synthetic multi-proxy reconstruction. The Holocene, 29, 847--863. https://journals.sagepub.com/doi/abs/10.1177/0959683619826634 Heagren, B. (2010). The Art of War in Pharaonic Egypt [Unpublished PhD. Thesis, The University of Auckland]. https://www.researchgate.net/publication/49463150_The_art_of_war_in_pharaonic_Egypt_an_analysis_of_the_tactical_logistic_and_operational_capabilities_of_the_Egyptian_army Healy, M. (1992). New Kingdom Egypt. Osprey Publishing. Kaniewski, D., Campo, E. V., Guiot, J., Burel, S. L., Otto, T., & Baeteman, C. (2013). Environmental Roots of the Late Bronze Age Crisis. PLOS ONE, 8(8), e71004. https://doi.org/10.1371/journal.pone.0071004 Kaniewski, D., Marriner, N., Bretschneider, J., Jans, G., Morhange, C., Cheddadi, R., Otto, T., Luce, F., & Van Campo, E. (2019). 300-year drought frames Late Bronze Age to Early Iron Age transition in the Near East: New palaeoecological data from Cyprus and Syria. Regional Environmental Change, 19, 2287--2297. https://doi.org/10.1007/s10113-018-01460-w Kitchen, K. A. (1975). Ramesside Inscriptions Historical and Biographical (Vols. 1–8). Blackwell. https://archive.org/details/KennethA.KitchenRamessideInscriptionsVol1 Kitchen, K. A. (1993a). Ramesside Inscriptions Translated and Annotated: Notes and Comments (Vols. 1–2). Blackwell. Kitchen, K. A. (1993b). Ramesside Inscriptions Translated and Annotated: Translations (Vols. 1–7). Blackwell. Kopanias, K. (2017). Mercenaries or refugees? The evidence from the inscriptions of Merenptah on the “Sea Peoples.” Journal of Greek Archaeology, 2, 119--133. https://www.ancientportsantiques.com/wp-content/uploads/Documents/AUTHORS/SeaPeoples/SeaPeoples-Kopanias2017.pdf Manassa, C. (2003). The Great Karnak Inscription of Merneptah: Grand Strategy in the 13th Century BC. The Yale Egyptological Seminar. Pollastrini, A. M. (2024). Helmets and Body Armour in New Kingdom Egypt. Bloomsbury Academic. https://doi.org/10.5040/9781350323520 Sabbahy, L. (2013). Depictional Study of Chariot Use in New Kingdom Egypt. In A. Veldmeijer & S. Ikram (Eds.), Chasing Chariots: Proceedings of the First International Chariot Conference (Cairo 2012) (pp. 191--202). Sidestone Press. Shaw, G. J. (2017). War and Trade with the Pharaohs: An Archaeological Study of Ancient Egypt's Foreign Relations. Pen & Sword. Shaw, I. (2019). Ancient Egyptian Warfare. Casemate. Snape, S. (2003a). The Emergence of Libya on the Horizon of Egypt. In D. O'Connor & S. Quirke (Eds.), Mysterious Lands (pp. 93--106). UCL Press. Snape, S. (2003b). Zawiyet Umm el-Rakham and Egyptian Foreign Trade in the Thirteenth Century BC. Sea Routes: Interconnections in the Mediterranean 16th-6th c. BC. https://www.academia.edu/107251529/Snape_ZUR_and_Egyptian_Foreign_Trade_Zawiyet_Umm_el_Rakham_and_Egyptian_foreign_trade_in_the_thirteenth_century_BC_ Snape, S. (2023). What was Zawiyet Umm el-Rakham for? Early Ramesside Strategy in the Libyan West. In F. Hoffmann & M. R. Abbas (Eds.), Perspectives on the Ramesside Military System (pp. 143--155). Zaphon. Spalinger, A. J. (2004). Review: Manassa, Colleen 2003. The Great Karnak inscription of Merneptah: Grand strategy in the 13th century BC. Journal of Egyptian Archaeology, 90(reviews supplement), 45--47. https://www.jstor.org/stable/3822286 Zutterman, C. (2003). The bow in the Ancient Near East, a re-evaluation of archery from the late 2nd millennium to the end of the Achaemenid empire. IrAnt, 38, 119--165. Learn more about your ad choices. Visit megaphone.fm/adchoices

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

Bumper to Bumper Radio, the car guys on KTAR, 92.3 FM in Phoenix, AZ, broadcast every Saturday from 11:00 am ...

The Fellowship assembles one last time before the final curtain.Not for a quiet conversation.For war.This week on Regarding Music from The Elder, Chaz, Wolfy, Scott D. Monroe, Corey Morrissette, Laura Morrissette, Sean McGinity, Debbie and Michael Pastore, Dan and Pang Papathopoulos—and even virtual appearances from Heath McCoy and Kevin Brown—return for the largest panel in the series as Scott unveils the penultimate chapter of his screenplay.What follows is the most ambitious chapter yet.Corey awakens blind after destroying the mysterious orbs, only to discover he has begun seeing the world in an entirely new way. The Cat discovers that loyalty has consequences. Terminus Village reveals its tragic past through Sypha's memories. Allies fall. Hidden powers awaken. Philosophies collide. And Mr. Blackwell returns more terrifying—and more unstable—than ever before.Meanwhile, the panel compares cover versions, discusses the band's late-career what-ifs, and marvels at how an album once dismissed as an infamous misfire has somehow become the foundation for an unexpectedly compelling fantasy epic.Along the way:The largest cast ever assembled for a Regarding season 5 episodeCorey loses his sight... and gains a different visionThe Cat's shocking fateSypha's heartbreaking return to her childhood villageThe Fox's final stand—and the ideas that die with himMr. Blackwell arrives like a force of natureA screenplay racing toward its final confrontationThis Week's Song:I — KISSFINAL VERDICT:Every great adventure reaches the point where escape is no longer the objective.The only way out......is through.The ShowIn this season of Regarding…, the panel tackles KISS's Music From The Elder one song at a time—testing whether its epic ambition holds up under scrutiny. Alongside the analysis, Scott D. Monroe's original screenplay tries to turn the album's abstract mythology into an actual story.Ambition meets accountability.GO BONELESSCertified boneless in the state of Ohio by the Boneless Podcasting Network. Go Boneless. Boneless Makes a Better Podcast. Hosted on Acast. See acast.com/privacy for more information.

Send us Fan MailIn Episode 262 of Book Talk Etc., Tina and Hannah share some August book releases that they are excited about! If you enjoy this commercial-free podcast, consider supporting us on Patreon! Your membership includes access to bonus episodes like What's in the Mailbag, Bookstore Browse: The Handsell, and Book Talk After Dark, invites to monthly community events like Mood Reader Happy Hour, and entry into our private Facebook group and Discord server- all for just $5 a month.Loving LatelyTony's Chocolate (T)Pipsticks (H)Latest ReadThe Shampoo Effect | Jenny Jackson (T+H)BOTM BooksFruit Fly | Josh Silver (T)Sunlight Finds You | Laura Moriarty (H)These Walls Remember | They Say a Girl Died Here | Sarah PinboroughThe Unheld | Luke Larkin (T)Deadly AnimalsThe Hill in the Dark Grove | Liam Higginson (H)Portrait of a Witch Undone | KS Shay (T)Crocodilopolis | John Manuel Arias (H)Current ReadsDie for Me | Shirlene Obuobi (T)The Children | Melissa Albert (H)If you prefer other shopping options, you can find today's books on Bookshop.org or Blackwell's. Purchasing through these links supports us with a small commission, at no extra cost to you.Support the showLet's Connect... Email us at booktalketc@gmailBTE on YoutubeTina's TikTok , IG @tbretc YT @tbretcHannah's TikTok , IG @hanpickedbooksJonathan IG @infiltrate_jayPodcast IG @booktalketcRenee's Substack Newsletter , IG@Itsbooktalk

Send us Fan MailIn Episode 261 of Book Talk Etc., Tina and Hannah celebrate the power of short books–from  short stories and novellas to essay collections and anthologies! Join us as we explore why these bite-sized reads can leave just as lasting an impact as a full-length novel.In this episode, we also spin the wheel to randomly select our next topic for our August episode of Tackling the TBR Shelf! These TBR tackle prompts are chosen by our patrons as a way for them to take part in our reading lives.If you enjoy this commercial-free podcast, consider supporting us on Patreon! Your membership includes access to bonus episodes like What's in the Mailbag, Bookstore Browse: The Handsell, and Book Talk After Dark, invites to monthly community events like Mood Reader Happy Hour, and entry into our private Facebook group and Discord server- all for just $5 a month.Loving LatelySummer Reading Campaign - Melanated Reader (T)Soola Bag (H)Soola Pouch (H)Latest ReadKeep Them Close | David Ellis (T)Look Closer | David EllisThe Best Lies | David EllisEverything to the Sea | Alicia Upano (H)Novellas and Short StoriesAt Home With the Horrors | Sammy ScottHuman Sacrifices | María Fernanda AmpueroI Know A Place | Nat Cassidy (T)Metamorphosis | Franz Kafka (H)Obstetrix | Naomi Kritzer (T)Automatic Noodle | Annalee Newitz (H)Shelf AdditionSeekers of Deer Creek | Thao Thai (T)People in Love | Claire Daverley (H)If you prefer other shopping options, you can find today's books on Bookshop.org or Blackwell's. Purchasing through these links supports us with a small commission, at no extra cost to you.Support the showLet's Connect... Email us at booktalketc@gmailBTE on YoutubeTina's TikTok , IG @tbretc YT @tbretcHannah's TikTok , IG @hanpickedbooksJonathan IG @infiltrate_jayPodcast IG @booktalketcRenee's Substack Newsletter , IG@Itsbooktalk

When an active shooter turned a military installation into a combat zone, Navy Airman Ryan Blackwell found himself staring down the barrel of a weapon. Drawing on base-level readiness and pure instinct, Blackwell's split-second decisions during the assault highlight the realities of survival under extreme duress. This episode breaks down the raw mechanics of an insider threat encounter and the warrior mindset required to survive it.

Send us Fan MailIn Episode 260 of Book Talk Etc., Tina and Hannah If you enjoy this commercial-free podcast, consider supporting us on Patreon! Your membership includes access to bonus episodes like What's in the Mailbag, Bookstore Browse: The Handsell, and Book Talk After Dark, invites to monthly community events like Mood Reader Happy Hour, and entry into our private Facebook group and Discord server- all for just $5 a month.Loving LatelyThe Cluci Wristlet (T)Yoto Player (H)Latest ReadWhere the Wildflowers Grow | Terah Shelton Harris (T)The Missed Connection | Tia Wliliams (T)SuperlativesThe One Day You Were My Husband | Rosie WalshAll the Blood We Share | Camilla BruceStrangers Behind Closed Doors | Catherine Adel WestVigil | George SaundersInto the Blue | Emma BrodieCleopatra | Saara El-ArifiYesteryear | Caro Claire BurkeTop 5 So Far!The Reformatory | Tanarive DueThe Caretaker | Marcus KliewerNightwatching | Tracy SierraKeeper of Lost Children | Sadeqa JohnsonNesting | Roisin O'DonnellJohn of John | Douglas StuartLost Lambs | Madeline CashLady Tremaine | Rachel HochhauserEverything to the Sea | Alicia UpanoIf you prefer other shopping options, you can find today's books on Bookshop.org or Blackwell's. Purchasing through these links supports us with a small commission, at no extra cost to you.Support the showLet's Connect... Email us at booktalketc@gmailBTE on YoutubeTina's TikTok , IG @tbretc YT @tbretcHannah's TikTok , IG @hanpickedbooksJonathan IG @infiltrate_jayPodcast IG @booktalketcRenee's Substack Newsletter , IG@Itsbooktalk

Chaz and the assembled Order of Mildly Concerned Scholars (Wolfy, Scott, Corey, and returning guests Heath McCoy, Laura Morrissette, Debbie Pastore, Michael Pastore, and Dan & Pang Pappathopoulos) arrive at Terminus Island expecting a rescue mission.Instead......they're invited to dinner.Because this week isn't really about "Escape from the Island."It's about what happens when the hero finally reaches the villain... only to discover the villain has already set the table.What unfolds is less another chapter of Scott's ever-expanding Elder screenplay and more a psychological chess match disguised as a fantasy adventure. Corey and his companions battle a crystal-powered war wagon, survive an underground labyrinth, and descend into a cavern so impossibly vast it feels like the album itself has finally become a location.Meanwhile, Mr. Blackwell isn't waiting on a throne.He's waiting with a goblet.And manners.There are flashbacks revealing the terrible truth behind Mrs. Blackwell's death.There are creatures pouring from fractures in reality.There is a warhorse with a change of heart.There is a fox serving bitter wine mixed with crow's blood because, at this point, subtlety has officially left the island.And hovering over everything is the growing realization that Blackwell no longer sees himself as the villain.He sees himself as the only adult left in the room.There is action.There is tragedy.There is a battle wagon that somehow feels perfectly reasonable by Episode 10.And somewhere beneath Terminus Island, Corey begins asking the most dangerous question of the entire season:What if saving Mr. Blackwell... is no longer possible?Featuring:Heath McCoy embracing his inner Macho Man for a rap performance nobody saw comingLaura Morrissette, Debbie Pastore, Michael Pastore, and Dan Papathopoulos returning as the Fellowship pushes toward the finaleA battle wagon armed with purple lasers, flaming wrecking balls, and one very angry catA dinner invitation that somehow feels more threatening than the battleMr. Blackwell proving that the most dangerous villains are often the most politeTHIS WEEK'S SONG:"Escape from the Island" — KISSFINAL VERDICT:Sometimes the hardest part isn't escaping the island...It's escaping the story waiting for you when you get there.The ShowIn this season of Regarding…, the panel tackles KISS's Music From The Elder one song at a time—testing whether its epic ambition holds up under scrutiny. Alongside the analysis, Scott D. Monroe's original screenplay tries to turn the album's abstract mythology into an actual story.Ambition meets accountability.GO BONELESSCertified boneless in the state of Ohio by the Boneless Podcasting Network. Go Boneless. Boneless Makes a Better Podcast. Hosted on Acast. See acast.com/privacy for more information.

Chicago Bears special teams star Josh Blackwell came through with two huge plays last season that led to two comeback wins. Here's why Blackwell will continue to be a top under-the-radar player next season.Become a supporter of this podcast: https://www.spreaker.com/podcast/shaw-local-s-bears-insider-podcast--3098936/support.

Philippa is joined by three brilliant crime writers — Abir Mukherjee, Jane Casey, and Sarah Hilary — to talk all about St Hilda's Crime Fiction Weekend, the unique Oxford crime fiction event running 4th–6th September. Each guest is allotted a school role (head girl, school council rep, and chair stacker) which determines the questions they're asked — and the chaos that follows is exactly as fun as it sounds.

Cardiovascular ambulatory surgery centers (ASCs) are no longer a future concept — they are quickly becoming a critical part of care delivery. In this episode of MedAxiom HeartTalk, host Melanie Lawson talks with Jerry Blackwell, MD, MBA, FACC, president and CEO of MedAxiom, and Aric Burke, founder and CEO of Atlas Healthcare Partners, about the momentum behind cardiovascular ASCs and the role MedAtlas CV is playing in advancing this transformation. Together, they explore how the right clinical expertise, operational model and physician alignment can help health systems deliver high-quality, efficient and lower-cost cardiovascular care in the outpatient setting.

Send us Fan MailIn Episode 259 of Book Talk Etc., Tina and Hannah tackle the TBR shelf by reading through Book of the Month titles that have been on their shelves for more than 2 years.New for July- join our Patreon for access to our Summer Read-a-Thon! Participants read from a set of curated prompts, submit the titles they finish for points, and follow along on a game board as they go patreon.com/booktalketcIf you enjoy this commercial-free podcast, consider supporting us on Patreon! Your membership includes access to bonus episodes like What's in the Mailbag, Bookstore Browse: The Handsell, and Book Talk After Dark, invites to monthly community events like Mood Reader Happy Hour, and entry into our private Facebook group and Discord server- all for just $5 a month.Loving LatelySummer Bucket List (T)Water Table (T)The Crash (H)Maternal Instinct (H)Latest ReadThe Truth About Ruby Cooper | Liz Nugent (T)Sacculina | Phillip Fracassi (H)BOTM BooksBOTM Historical Data BaseNot That I Could Tell | Jessica Strawser (T)You're Invited | Amanda Jayatissa (H)The Joy Luck Club | Amy Tan (T)Once There Were Wolves | Charlotte McConaghy (H)Shelf AdditionDie for Me | Shirlene Obuobi (T)The Missed Connection | Tia Williams (H)If you prefer other shopping options, you can find today's books on Bookshop.org or Blackwell's. Purchasing through these links supports us with a small commission, at no extra cost to you.Support the showLet's Connect... Email us at booktalketc@gmailBTE on YoutubeTina's TikTok , IG @tbretc YT @tbretcHannah's TikTok , IG @hanpickedbooksJonathan IG @infiltrate_jayPodcast IG @booktalketcRenee's Substack Newsletter , IG@Itsbooktalk

Join us each Sunday for services at 8:30 or 11AM. We stream our 11AM service on Youtube and Facebook under West Mobile Baptist Church.

Elizabeth Blackwell was born in London in the early 18th century, and was known in her lifetime for her achievements as a botanical illustrator. Research: “A Genuine Copy of a Letter &c.” Stockholm, August 20. H. Carpenter in Fleet Street, 1747. https://books.google.com/books/about/A_Genuine_Copy_of_a_Letter_from_a_Mercha.html?id=EPRbAAAAQAAJ Alexander, Isabella and Cristina S. Martinez. “2. The First Copyright Case under the 1735 Engravings Act: 
The Germination of Visual Copyright?” From Circulation and Control: Artistic Culture and Intellectual Property in the Nineteenth Century. Marie-Stéphanie Delamaire and Will Slauter, editors. https://doi.org/10.11647/OBP.0247 Beharrel, Will. “Elizabeth Blackwell's Curious Herbal.” The Linnean Society. 7/28/2021. https://www.linnean.org/news/2021/07/28/elizabeth-blackwells-curious-herbal Blackwell, Elizabeth (1737). A Curious Herbal. Containing Five Hundred Cuts of the most useful Plants, which are now used in the Practice of Physick. Engraved on folio Copper Plates, after Drawings, taken from the Life. By Elizabeth Blackwell. To which is added a short Description of ye Plants; and their common Uses in Physick. London: Printed for Samuel Harding in St Martin’s Lane, MDCCXXXVII (1737) Rubenstein QK99.A1 B53 1737 folio v.1 c.1. Scan of preface. https://blogs.library.duke.edu/rubenstein/files/2022/10/blackwell-preface-scaled.jpg Bruce, James. “Lives of Eminent Men of Aberdeen.” Aberdeen. The University Press. 1841. https://archive.org/details/b33028722/ Chelsea Physic Garden. “Curious Herbal; Curious Tale.” Newsletter. Spring-Summer 2005. Child, Lydia Maria. “Biographies of Good Wives.” Boston: Munroe & Francis. 1850. https://archive.org/details/biographiesofgoo00chil_0 Elliott, Brent. “The World of the Renaissance Herbal.” Renaissance Studies. Vol. 25, No. 1. February 2011. Via JSTOR. https://www.jstor.org/stable/24420235 Evenden, Doreen A. "Blackwell [née Simpson], Elizabeth (1699–1758), botanical author and artist." Oxford Dictionary of National Biography. August 08, 2024. Oxford University Press. Date of access 18 Jun. 2026, https://www.oxforddnb.com/view/10.1093/ref:odnb/9780198614128.001.0001/odnb-9780198614128-e-2540 Grosjean, A. N. L. "Blackwell, Alexander (bap. 1709, d. 1747), agricultural improver and government agent in Sweden." Oxford Dictionary of National Biography. June 08, 2023. Oxford University Press. Date of access 18 Jun. 2026, https://www.oxforddnb.com/view/10.1093/ref:odnb/9780198614128.001.0001/odnb-9780198614128-e-2539 Huler, Scott. “A Beautiful Find.” Duke Mag. 9/5/2023. https://dukemag.duke.edu/stories/beautiful-find Madge, Bruce. “Elizabeth Blackwell—the forgotten herbalist?” Health Information & Libraries Journal, 18: 144-152. https://doi.org/10.1046/j.1471-1842.2001.00330.x Monroe, Nicky. “Elizabeth Blackwell’s Curious Herbal.” RHS Libraries and Collections. https://www.rhs.org.uk/education-learning/libraries-at-rhs/articles/elizabeth-blackwell Newman, Joyce. “Will The Real Elizabeth Blackwell Please Stand Up?” New York Botanical Garden. 7/1/2013. https://www.nybg.org/blogs/plant-talk/2013/07/exhibit-news/will-the-real-elizabeth-blackwell-please-stand-up/ O’Keeffe, Lynda. “Guest post by Lynda O’Keeffe – A Curious Herbal Elizabeth Blackwell’s Pioneering Masterpiece of Botanical Art.” All Things Georgan. 3/8/2024. https://georgianera.wordpress.com/2024/03/08/guest-post-by-lynda-okeeffe-a-curious-herbal-elizabeth-blackwells-pioneering-masterpiece-of-botanical-art/ Pardoe, Heather and Maureen Lazarus. “Images of Botany: Celebrating the Contribution of Women to the History of Botanical Illustration.” Collections: A Journal for Museum and Archives Professionals, Volume 14, Number 4, Fall 2018, pp. 545–566. RHS Digital Collections. “Elizabeth Blackwell's Curious Herbal.” https://collections.rhs.org.uk/collection/111276 Royal College of Physicians of Edinburgh. “Elizabeth Blackwell: Prison, Plotting and the Curious Herbal.” https://www.rcpe.ac.uk/heritage/heritage-blog/elizabeth-blackwell-prison-plotting-and-curious-herbal Shirk, Henrietta Nickels. “Contributions to Botany, the Female Science, by Two Eighteenth-century Women Technical Communicators.” Technical Communication Quarterly. Vol. 6, No. 3. Summer 1997. Tyson, Janet Stiles. “Introducing Elizabeth Blackwell to Hans Sloane.” British Library Untold Lives Blog. 5/18/2021. Via Archive.org. https://web.archive.org/web/20210619032948/https://blogs.bl.uk/untoldlives/2021/05/introducing-elizabeth-blackwell-to-hans-sloane.html Tyson, Janet Stiles. “The Rubenstein Library’s disruptive copy of A Curious Herbal.” 11/14/2022. https://blogs.library.duke.edu/rubenstein/2022/11/14/a-curious-herbal/ Tyson, Janet. “'A Curious Herbal' as Material Witness.” The Linnean Society. 1/10/2023. https://www.linnean.org/news/2023/01/10/a-curious-herbal-as-material-witness See omnystudio.com/listener for privacy information.

Send us Fan MailIn Episode 258 of Book Talk Etc., Tina and Hannah chat about Land by Maggie O'Farrell, which was the Book Talk Etc. Community Read for the month of June. In this episode we also share 10 July book releases to get excited about!Join our Patreon for access to our Summer Read-a-Thon! Participants read from a set of curated prompts, submit the titles they finish for points, and follow along on a game board as they go patreon.com/booktalketcIf you enjoy this commercial-free podcast, consider supporting us on Patreon! Your membership includes access to bonus episodes like What's in the Mailbag, Bookstore Browse: The Handsell, and Book Talk After Dark, invites to monthly community events like Mood Reader Happy Hour, and entry into our private Facebook group and Discord server- all for just $5 a month.Loving Latelyyesstyle.com (T)Beauty of Joseon Sunscreen (T)Lip Oil (T)Skin 1004 Cream (T)Alaffia Body Wash (H)Latest ReadLand | Maggie O'Farrell (T+H)July Books on the RadarCountry People | Daniel Mason (T)Every Version of You | Natalie Messier (H)Not with A Bang | Temi Oh (T)The Mortons | Justine Larbalestier (T)The Man | Laura SimsLooker | Laura SimsHow Can I Help You? | Laura Sims (H)The New People | Andrea Uptmor (T)The Parisian Heist | Jo Piazza (T)Make Nice | Ryan Effgan (H)Shelf AdditionNot That I Can Tell | Jessica Strawser (T)Sacculina | Philip Fracassi (H)If you prefer other shopping options, you can find today's books on Bookshop.org or Blackwell's. Purchasing through these links supports us with a small commission, at no extra cost to you.Support the showLet's Connect... Email us at booktalketc@gmailBTE on YoutubeTina's TikTok , IG @tbretc YT @tbretcHannah's TikTok , IG @hanpickedbooksJonathan IG @infiltrate_jayPodcast IG @booktalketcRenee's Substack Newsletter , IG@Itsbooktalk

What does death reveal about who we really are?This week I'm joined by Professor Douglas Davies, Director of the Centre for Death and Life Studies at Durham University and one of the world's leading scholars of death, ritual, and belief. His work spans decades and disciplines — from the anthropology of funerals to digital legacy, from woodland burial to the theology of grief — and his central conviction runs through all of it: the dead live within us, and recognising that can help us live better.We talk about the ways death strips away the myth of the self-made individual, revealing that we are fundamentally relational beings — shaped by the people, places, and memories we carry. Along the way, we cover the full arc of how societies and individuals make meaning in the face of mortality.We talk about:Résumé virtues versus eulogy virtues — and why the gap between them mattersThe concept of "dividual" personhood, and why the idea of a fixed, separate self breaks down when we actually look at how people live and grieveHow grief theory has shifted from letting go to continuing bonds — and what that means for how we mournWhy funerals work: their role as social containers for emotion and meaningThe rise of celebration of life services and direct cremation, and what those trends tell usWoodland burial, the scattering of ashes, and the pull of relational placesDying alone, shame, dignity, and what COVID forced us to confront about communityDigital death platforms, online memorials, and why offline ritual still does something differentNew body disposition options, including alkaline hydrolysisPets, suicide, and the ways love complicates every tidy theory of griefFor further reading, Douglas's book Death, Ritual and Belief: The Rhetoric of Funerary Rites (now in its third edition) is a rich and authoritative guide to everything this conversation touches — available from Blackwell's and other independent bookshops.If this episode resonated with you, the best thing you can do is share it with someone. Word of mouth is genuinely how the podcast finds new listeners. And if you haven't already, leaving a review is hugely appreciated.Support the show

June 28, 2026 AM Service

I've covered three British herbalists so far on this podcast, and each has made a notable contribution to the development of botany and apothecary practice in Britain. This week, we're going to meet Maud Grieve, whose contribution took a different form. Yes, she wrote a lot about plants. She wrote A Modern Herbal, in fact. Yet she also contributed to the wartime efforts of the First World War in the realm of medicinal plants. She's perhaps not as well-known as the other three herbal writers I've featured: Elizabeth Blackwell, William Turner, and Nicholas Culpeper. She didn't illustrate her herbal, like Blackwell. Nor did she translate Latin texts into English, like Turner or Culpeper. But she did get people growing–and using–their own herbs. Let's go to meet her in this week's episode of Fabulous Folklore! Find the blog post with all the images and references here: https://www.icysedgwick.com/maud-grieve/ Become a member of Herbaria here: https://school.rowanandsage.com/courses/herbaria?affcode=437598_3qokpyep Magical Legends of the North East Talk: https://ko-fi.com/s/7f42dec282 Get your free guide to home protection the folklore way here: https://www.icysedgwick.com/fab-folklore/ Become a member of the Fabulous Folklore Family for bonus episodes and articles at https://patreon.com/bePatron?u=2380595 Get weekly articles and bonus content at Substack: https://fabulousfolklore.substack.com/ Buy Icy a coffee or sign up for bonus episodes at: https://ko-fi.com/icysedgwick Find the Fabulous Folklore Bookshop, Icy's social media links, and other useful bits at: http://icysedgwick.com/start-here

Wandering Works for Us PodcastDate: 27 June 2026Title: What we did on our trip to Oxford UK, and what we would do differently next timeSummary of EpisodeOxford was the first proper stop on our England road trip with Julie and Ann — and it nearly didn't get the time it deserved. In this episode, we talk about driving into the city in 5 o'clock traffic (on the wrong side of the road, in a rental car twice the size of ours), the 30-minute Bodleian Library tour that left us wishing we'd booked the longer one, and the completely unplanned moment that turned out to be the highlight of the whole stop: stumbling into a boys' choir rehearsal inside the chapel at New College. We also get into the Ashmolean, Oxford's literary pubs, an accidental trip to Blackwell's, and everything we'd do differently if we went back.TakeawaysTwo weeks is the ideal maximum for a trip like ours.Driving in the UK requires patience and understanding of local road rules.Oxford's architecture and history make it a must-visit destination.Traveling with friends adds joy but also logistical challenges.Planning for local experiences and downtime enhances travel enjoyment. Let's TalkHave you been to Oxford? What did we miss? Find us on Instagram and Threads and tell us — we'd love to know.Chapters00:00 Introduction to the Podcast Experiment01:01 Reflections on the Grand Adventure04:00 Traveling with Friends: The Dynamics of Group Travel05:00 Arrival in Oxford: The Journey BeginsKey TopicsDriving from London to Oxford — first time behind the wheel on the "wrong" side of the road, plus a lunch stop at a pub claiming to be over 900 years oldFirst impressions of Oxford: more spread out than we expected, and very much a lived-in city rather than a museum pieceThe Bodleian Library and Divinity School tour — the Harry Potter hospital wing, the centuries-old "no books leave the library" rule, and why we'd book the longer tour next timeNew College and the moment that became the whole episode: walking into a chapel mid-rehearsal and just standing thereThe Ashmolean Museum, and our case for why you don't need to see everything to enjoy a museumPubs, Inklings, and an accidental bookshop detour at Blackwell'sWhat we'd do differently — and what we'd tell a first-timer planning their own visit Mentioned in This EpisodeWhere we stayed: Newton HouseComing up next: we head into the Cotswolds — one-lane roads, sheep, and the best gin we've had from a village storeImportant Links To follow all of our antics and adventures, please visit our social media pages and our website at wwforus.com! You can send us a message at any of these places, and feel free to email us at wandering@wwforus.comLike what we are doing? Buy us a gin and tonic and help us keep going!InstagramFacebookTiktokYouTubeLooking for a tour guide in Portugal? I have a whole list!Blog post for this episode: Best things to do in Oxford (1-2day guide)RESOURCES & LINKSLooking to plan your next trip to Portugal? We can help! Check out our guides and Itineraries at wwforus.comFree Lisbon ItineraryPacking ListEssentials for every tripRenting a car in PortugalLooking for a tour guide in Portugal? I have a whole list!

Catch the second half of Leslie Blackwell's incredible story of redemption and healing this week on Speak Up! Virginia. Leslie and Candi confront the lies told to women and share resources for truth and healing.For information and to find support, please visit: https://supportafterabortion.com/For more information on healing post-abortion, please visit: https://www.rachelsvineyard.org/To read more post-abortion testimonials, please visit: https://abortiontestimonials.com/To help push back on the lies in our culture and defend life, visit: https://www.familyfoundation.org/life

Send us Fan MailIn Episode 257 of Book Talk Etc., Tina and Hannah talk through some of their favorite book to screen adaptations, share some books they'd love to see adapted into film, as well as discuss their thoughts on books they read that have been turned into movies or television series.If you enjoy this commercial-free podcast, consider supporting us on Patreon! Your membership includes access to bonus episodes like What's in the Mailbag, Bookstore Browse: The Handsell, and Book Talk After Dark, invites to monthly community events like Mood Reader Happy Hour, and entry into our private Facebook group and Discord server- all for just $5 a month.Loving LatelyLibib (T)Arc Library App (T)Bobbie Goods Coloring Books (H)Latest ReadJohn of John | Douglas Stuart (T)Recitatif | Toni Morrison (H)Book Talk + Book to Screen AdaptationsStrangers Behind Closed Doors | Catherine Adel West (T)The Ballad of Songbirds and Snakes | Suzanne Collins (H)Shutter Island | Dennis LeHane (T)The Bookshop | Penelope Fitzgerald (H)Shelf AdditionNine Lives | Catherine Steadman (T)Games: A Love Story | Anna Maria Volkova (H)If you prefer other shopping options, you can find today's books on Bookshop.org or Blackwell's. Purchasing through these links supports us with a small commission, at no extra cost to you.Support the showLet's Connect... Email us at booktalketc@gmailBTE on YoutubeTina's TikTok , IG @tbretc YT @tbretcHannah's TikTok , IG @hanpickedbooksJonathan IG @infiltrate_jayPodcast IG @booktalketcRenee's Substack Newsletter , IG@Itsbooktalk

Diversifying income streams, building community, and personal storytelling with Rebecca Blackwell. ----- Welcome to episode 576 of The Food Blogger Pro Podcast! This week on the podcast, Bjork interviews Rebecca Blackwell. Last week on the podcast, Bjork chatted with Jenna Arend. To go back and listen to that episode, click here. How Rebecca Blackwell Turned Food Blogging into a Six-Figure Digital Business from an RV Rebecca started out as a food blogger, but somewhere along the way, she traded a stable paycheck for a life on the road, running her digital business full-time from an RV! In this episode, she opens up about what that transition actually looked like: the risks she took by leaving a steady job, how the RV lifestyle reshaped her thinking about work and creative freedom, and why she made the strategic decision to manage multiple websites rather than rely on a single income stream. The conversation also gets into the community side of Rebecca's business. Specifically, how she's used Substack to build genuine connections with a growing audience of food writers. She shares practical advice on growing a newsletter, engaging readers, and landing sponsorships that feel like a natural fit. You'll also hear her talk about how she's navigating the challenges of AI and shifting search algorithms, why leaning into personal storytelling has been her biggest differentiator, and what she's learned from building community through mastermind groups and in-person retreats. Three episode takeaways: Diversify your income streams: Branching out to different avenues instead of relying on a single blog or platform gives you more creative freedom and financial stability, especially important in an era of unpredictable algorithm changes! Personal storytelling is your biggest differentiator: As AI and search engines continue to evolve, what sets your content apart isn't information. it's your unique voice, perspective, and lived experience. Rebecca's journey from food blogger to RV-dwelling digital creator is a perfect example of a story no algorithm can replicate. Community is a growth strategy: Whether it's building a newsletter on Substack, joining a mastermind group, or hosting retreats, investing in genuine relationships with your audience and peers pays dividends that go far beyond traffic and page views. Resources: A Little and a Lot Of Batter and Dough Rebecca's Substack: Let's Get Lost Tiny Shiny Home Designing Your Life Substack Mastermind for Food Writers The Food Writers Business Lab Kit Follow Rebecca on Instagram here and here Join the Food Blogger Pro Podcast Facebook Group Thank you to our sponsors! This episode is sponsored by Member Kitchens. Learn more about our sponsors at foodbloggerpro.com/sponsors. Interested in working with us too? Learn more about our sponsorship opportunities and how to get started here. If you have any comments, questions, or suggestions for interviews, be sure to email them to podcast@foodbloggerpro.com. Learn more about joining the Food Blogger Pro community at foodbloggerpro.com/membership.

In this episode of GuildSomm: Into the Glass, Advanced Sommelier Alisha Blackwell-Calvert joins GuildSomm's director of education, Chris Tanghe, to blind taste three high-acid white wines. They discuss the differences between lees and oak aging as well as the evolution of typicity in classic winemaking regions.   Alisha is based in St. Louis and is currently the wine director for Madrina. She previously worked at Cinder House in the Four Seasons Hotel and was a James Beard semi-finalist for Outstanding Professional in Beverage Service in 2025. She judges for the Decanter World Wine Awards and serves as a mentor for the Bâtonnage Mentorship Program.   Thanks for listening. If you enjoy this episode, please consider leaving us a review, as it helps us connect and grow the GuildSomm community. Cheers!

Chef George Blackwell Smith IVTake a walk with me down Fascination Street as I get to know Chef George Blackwell Smith IV. In this episode, we chat about what led him to fall in love with cooking, and the taste of Louisiana specifically. Chef Blackwell Smith explains how he went from growing up in California, moving to Chattanooga, then to Louisiana, and why he went back to Tennessee. Then we discuss his company The Lucky Cajun This is a premium spice and seasoning blend company that prides itself on the freshest blends. I get the chef to explain to me what makes his spices and seasonings better than what you grab off of the supermarket shelf. We do trade restaurant horror stories before we get into the specifics of each of his product line offerings. The Lucky Cajun offers a hot sauce called The Green Boss, as well as several unique blends. Blackbeard's Smoke, Jerk, Cajun, and the Original. Chef Blackwell Smith explains his philosophy of 'the anatomy of the bite', and an ever-expanding selection of easy-to-read cookbooks. He teases some of the upcoming new products, and we even touch on some of the skills that make spending time in the kitchen a ton of fun. Make sure you check out what the chef has to offer over at TheLuckyCajun.com and follow him on social media.

Ohio Auditor Candidates In November, when voters go to the polls, they'll be voting on every statewide executive office in Ohio, one of those being auditor of state. The two candidates for that position are the current Secretary of State, Republican, Frank LaRose, and the Mayor of Maple Heights, Democrat, Annette Blackwell. They recently spoke at forums organized by the Akron Press Club. We'll bring you excerpts from that on Monday's "Sound of Ideas." You can watch the entirety of LaRose's speech here. You can watch the entirety of Blackwell's speech here. Guests: - Frank LaRose, Secretary of State; Candidate, Ohio Auditor - Annette Blackwell, Mayor of Maple Heights; Candidate, Ohio Auditor Foreign Relations with India Then, we turn our attention to international affairs and look at U.S.-India relations. The most populous country is home to a quickly growing economy, and could be a major player in global policies, from immigration to trade, over the coming decades. Akriti Kalyankar, a woman who worked as a journalist in India, is now coming to speak in Northeast Ohio on Thursday, at the invitation of the Cleveland Council on World Affairs. You can find out more about Kalyankar's appearance here. Guest: - Akriti Kalyankar, Fellow, Stimson Center

This week, Candi is joined by Leslie Blackwell, a former Pro-Choice activist turned March for Life speaker and warrior for Life, helping women heal from their abortions. Hear the first part of this compelling interview where Leslie shares her extrordinary testimony.For more information on Leslie's work, visit: https://www.facebook.com/SilentnomoreRVA/

Send us Fan MailIn Episode 256 of Book Talk Etc., Tina and Hannah talk about the types of thriller and horror novels that are perfect to read in summer months! We chat all things thrills & chills and chat a bit about summerween traditions.If you enjoy this commercial-free podcast, consider supporting us on Patreon! Your membership includes access to bonus episodes like What's in the Mailbag, Bookstore Browse: The Handsell, and Book Talk After Dark, invites to monthly community events like Mood Reader Happy Hour, and entry into our private Facebook group and Discord server- all for just $5 a month.Loving LatelyJoJo Gemstone Choker Necklace (T)Byoma Phyto-Mucin Glow Serum (H)Latest ReadThe One Day You Were My Husband | Rosie Walsh (T)One True Loves | Taylor Jenkins ReidLaura DaveThis Story Might Save Your Life | Tiffany CrumThe Wishing Spell | Chris Colfer (H)Book Talk + Books We Missed Out OnJanelle BrownJulie ClarkThe God of the Woods | Liz MooreGabby's YouTube ChannelThe Troop | Nick CutterThe Compound | Aisling RawleMarion | Leah Rowan (T)Sundial | Catriona Ward (H)Why Catriona Ward Decided To Explore TheAbhorrent World of Animal Testing - Article on Crime readsThe Other | Annie Neugebauer (T)The Last Time I Lied | Riley SagerSurvive the Night | Riley SagerLock Every Door | Riley Sager Final Girls | Riley SagerShelf AdditionThe Calamities | Chuck Wendig (T)The Odyssey | Homer (H)If you prefer other shopping options, you can find today's books on Bookshop.org or Blackwell's. Purchasing through these links supports us with a small commission, at no extra cost to you.Support the showLet's Connect... Email us at booktalketc@gmailBTE on YoutubeTina's TikTok , IG @tbretc YT @tbretcHannah's TikTok , IG @hanpickedbooksJonathan IG @infiltrate_jayPodcast IG @booktalketcRenee's Substack Newsletter , IG@Itsbooktalk

Semiconductors have moved from the background of the technology stack to the center of the AI economy. What used to be a specialized industry discussed mostly by engineers and investors is now shaping the speed, cost, and strategic direction of modern computing.In this episode of TechSurge, host Michael Marks speaks with Stacy Rasgon, Managing Director and Senior Analyst covering U.S. semiconductors and semiconductor capital equipment at Bernstein Research. Stacy has spent years analyzing the chip industry across cycles, but argues that the current moment feels different in scale: AI demand has created an unprecedented scramble for compute, memory pricing has surged, and companies across the stack are being forced to rethink capacity, architecture, and capital allocation.The conversation explains the 4 different kinds of semiconductor cycles—supply, inventory, product, and demand — and why Stacy believes the industry is currently in a demand cycle of unusual magnitude. The discussion also unpacks the distinction between DRAM and NAND, why high-bandwidth memory is becoming strategically central to AI systems, and how the physical realities of wafer capacity and silicon area are constraining supply in ways the broader market often misses.Stacy and Michael also discuss the hardware economics behind the current boom, with Michael pressing Stacy on why compute remains so scarce and how companies are improving performance through packaging and system design. Michael then moves the conversation beyond market headlines to the core business questions: who is actually paying for this compute, which use cases are generating real revenue, and whether AI spending is creating durable economic value or simply shifting costs elsewhere. Together, these questions highlight two of the episode's clearest insights: coding may be one of the earliest AI applications with meaningful willingness to pay, and inference, not training, is the real test of whether the current buildout becomes a lasting business or just another expensive wave of infrastructure.Stacy explains the concentration of power among the major wafer fabrication equipment players, the rise of ASICs as a meaningful share of AI silicon, Broadcom's rapidly expanding AI opportunity, and the growing role of Chinese companies as new entrants, especially in memory and semiconductor equipment. Along the way, the conversation asks the defining question facing the sector: is this just another semiconductor upswing, or the first true supercycle the industry has seen? Stacy believes that this might be the biggest supercycle he has seen in his career.Sign up for our newsletter at techsurgepodcast.com for updates on upcoming TechSurge Live Summits and future episodes.Links:Stacy Rasgon on LinkedIn: https://www.linkedin.com/in/stacy-rasgon-6924963Bernstein: https://www.alliancebernstein.com/corporate/en/home.htmlReferences Mentioned During the DiscussionNVIDIA Blackwell Platform: https://www.nvidia.com/en-us/data-center/blackwell-platform/High Bandwidth Memory (HBM) overview from Micron: https://www.micron.com/products/memory/hbmDRAM overview from IBM: https://www.ibm.com/think/topics/dramNAND flash overview from IBM: https://www.ibm.com/think/topics/nand-flash-memoryFurther ReadingMcKinsey on the semiconductor industry outlook: https://www.mckinsey.com/industries/semiconductors/our-insights/the-semiconductor-industry-in-2025Semiconductor Industry Association: 2025 State of the U.S. Semiconductor Industry: https://www.semiconductors.orgNVIDIA on the Blackwell architecture and AI infrastructure roadmap: https://www.nvidia.com/en-us/data-center/blackwell-platform/Broadcom AI investor materials and infrastructure commentary: https://investors.broadcom.comASML on lithography and advanced chip manufacturing: https://www.asml.com/en/technologyMicron on HBM and AI memory demand: https://www.micron.com/products/memory/hbmChapters[00:00:00] — Highlights[00:00:26] — Welcome to  the Episode[00:01:29] — Meet Stacy Rasgon[00:02:01] — Is This the First Real Semiconductor Supercycle?[00:05:33] — Inside the Strongest Memory Cycle in History [00:09:14] — Can Innovation Keep Up With AI Demand?[00:11:33] — Chiplets, Blackwell, and the New Economics of Compute [00:12:37] — What Could Signal the Cycle Is Slowing[00:14:26] — Vertical Integration at the Hyperscales [00:16:36] — The Difference between Apple and Meta[00:17:15] — What is Vertical Integration Being Done For?[00:18:15] — Will other bottlenecks develop as This Progresses? [00:21:13] — Oligopoly Pricing in the Market[00:22:22] — Any New Entrants into Memory?[00:23:46] — Why the Industry Must Pivot From Training to Inference[00:25:10] — Agentic Coding and the First Real AI Revenues[00:26:57] — Groq, Low-Latency Inference, and What GPUs Cannot Do Alone[00:29:28] —-Could The Smaller Companies All be Bought Up ?[00:30:19] — Why Semiconductor Equipment Matters More Than Ever [00:31:00] — How Semiconductor Equipment is Affected by the Cycle[00:32:55] — A Long Upcycle for Semiconductor Equipment Guys?[00:33:13] — The Big Five and the Rise of Chinese Equipment Players[00:34:24] — The Effects of Geopolitics[00:35:02] — Broadcom's Quiet AI Breakout[00:40:46] — ASICs vs GPUs and the Next Wave of Custom Chips[00:41:06] — Intel, Foundry Strategy, and the Long Turnaround[00:46:46] —-The Risks the Market May Still Be Underestimating[00:49:32] — Where Startups Still Have Room to Win[00:50:39] — What the Semiconductor Industry Could Look Like Next Year

Miss Darcy in 1995? Amelia Blackwell on Time Travel, Mysteries and Jane Austen When Miss Georgianna Darcy sets out to explore the woods surrounding her home, she comes upon a peculiar object. Of course, Miss Darcy must examine this strange thing further, but in the blink of an eye, she finds herself transported from 1799 to 1995. Uncertain as to what just happened, Miss Darcy soon realizes that she may be the only witness to a murder and that she's landed in the middle of a film crew shooting a Jane Austen movie. As Miss Darcy learns to navigate 1995, she discovers love, friendship and helps solve the murder. A Crime Through Time is a fun, enchanting mystery with charming characters including a Border Collie named Watson. Join me for this entertaining discussion with Amelia Blackwell about her real life inspiration and experience with a Jane Austen film, her love for animals and much more. For more information on Amelia, visit: instagram.com/ameliablackwellauthor/ or instagram.com/lisaglassauthor/ For more about my K-9 books, visit: kathleendonnelly.com Sit. Stay. Read. is a proud part of the Authors on the Air Global Network.

Send us Fan MailIn Episode 255 of Book Talk Etc., Tina and Hannah share some analog friendly loving latelies to continue to help you stay off your phone this summer and cure your brain rot! We also share some thoughts on new books that have released recently, as well as some thoughts on tropes we love, book tropes we hate, book tropes we anticipate being popular as well as the ones that are already buzzing around in the publishing world.If you enjoy this commercial-free podcast, consider supporting us on Patreon! Your membership includes access to bonus episodes like What's in the Mailbag, Bookstore Browse: The Handsell, and Book Talk After Dark, invites to monthly community events like Mood Reader Happy Hour, and entry into our private Facebook group and Discord server- all for just $5 a month.Loving Lately@Tabletopfamily (T)These Books Will Heal Your Brain Rot (H)Latest ReadOpen Wide | Jessica Gross (T)The Night Tiger | Yangsze Choo (H)Book Talk + Books We Missed Out OnBuffalo Hunter HunterCoffin MoonTrad Wife | Saratoga Schaefer (T)Nothing Tastes as Good | Luke DumasTropesick  | Lauren Okie (H)Dissection of a Murder | Jo Murray (T) The Fountain | Casey Scieszka (H)Shelf AdditionBack Stabbers | Eliza Jabore (T)They All Fall in Love at the End | Haili Blassingame (H)If you prefer other shopping options, you can find today's books on Bookshop.org or Blackwell's. Purchasing through these links supports us with a small commission, at no extra cost to you.Support the showLet's Connect... Email us at booktalketc@gmailBTE on YoutubeTina's TikTok , IG @tbretc YT @tbretcHannah's TikTok , IG @hanpickedbooksJonathan IG @infiltrate_jayPodcast IG @booktalketcRenee's Substack Newsletter , IG@Itsbooktalk

Josh and Raul are joined by Brian Geisinger of 247Sports and the Buzz Beat podcast to break down the expectations for Duke newcomers John Blackwell and Joaquim Boumtje-Boumtje. Topics include: Whether Blackwell will operate more on ball or as a movement shooterBlackwell as a screen setterHow Duke has made use of similar players to Blackwell under Jon ScheyerBlackwell's midrange gameHow to evaluate Joaquim Boumtje-Boumtje's overseas productionBoumtje-Boumtje's shooting ability at his sizeAreas where Boumtje-Boumtje still needs to developHow Boumtje-Boumtje and Cameron Williams might complement and push each otherWhat year two of Boumtje-Boumtje might look like as more opportunities open upBrian's work can be found at https://briangeisinger.substack.com/

Send us Fan MailIn Episode 254 of Book Talk Etc., Tina and Hannah share their thoughts on Japanese Gothic, their community read for May. They also share thoughts on books from the past they feel they missed out on, as well as chat through what they think about when they consider what it means to "miss out" on books.If you enjoy this commercial-free podcast, consider supporting us on Patreon! Your membership includes access to bonus episodes like What's in the Mailbag, Bookstore Browse: The Handsell, and Book Talk After Dark, invites to monthly community events like Mood Reader Happy Hour, and entry into our private Facebook group and Discord server- all for just $5 a month.Loving LatelyMonthly Meal Planning (T)Monitor Clip on Fan (H)Latest ReadJapanese Gothic | Kylie Lee Baker (T+H)Book Talk + Books We Missed Out OnBuckeyeAtmosphereDemon Copperhead | Barbara KingsolverLessons in Chemistry | Bonnie GarmusJane Eyre | Charlotte Bronte1984 | George OrwellThe Last Party | AR Torre (T)Buffalo Hunter Hunter | Stephen Graham Jones (H)Still Life | Louise Penny (T)The Deal | Elle Kennedy (H)Shelf AdditionNerve Damage | Annakeara Stinson (T)Liars Dice | Juliet Faithfull (H)If you prefer other shopping options, you can find today's books on Bookshop.org or Blackwell's. Purchasing through these links supports us with a small commission, at no extra cost to you.Support the showLet's Connect... Email us at booktalketc@gmailBTE on YoutubeTina's TikTok , IG @tbretc YT @tbretcHannah's TikTok , IG @hanpickedbooksJonathan IG @infiltrate_jayPodcast IG @booktalketcRenee's Substack Newsletter , IG@Itsbooktalk

This week, Zoe is joined by Made in Chelsea star Emily Blackwell for one of our most honest and moving conversations yet.Emily opens up about the birth of her daughter Eva who arrived 11½ weeks early with no warning, no hospital bag and no pain relief. She shares what the 53 days that followed in the NICU really looked like and talks about guilt, gratitude, the community she's built with other NICU parents and why the journey doesn't end at discharge.Plus: leaving reality TV on her own terms, being a self-confessed Type B bride ahead of her Mallorca wedding and the one thing about motherhood that genuinely surprised her.Emma Spring Bank Holiday Sale is live! Get up to 25% off plus extra 5% using the code MAYSLEEP at Emma Sleep. Hosted on Acast. See acast.com/privacy for more information.

The Tony Awards often celebrate the names on a show poster, but every Broadway production is carried by countless artists whose work rarely fits neatly into a category. So in anticipation of this year's awards ceremony, we're spotlighting a few of the performers, collaborators, and creative forces behind nominated productions. Allison Blackwell knows the power of ensemble storytelling firsthand. Currently appearing in Ragtime at Lincoln Center Theater as Sarah's Friend, Allison is part of the acclaimed company behind one of Broadway's most resonant productions this season. She has built a career defined by extraordinary versatility, emotional depth, and a deep commitment to collaboration onstage.  In this conversation, Allison reflects on the path that brought her back to theater after nearly pursuing law school; what it has meant to revisit Ragtime across multiple productions; and what recognition for ensemble artists means to her personally. ----- LINKS Allison Blackwell: https://www.allisonblackwell.com/ Ragtime at Lincoln Center Theater: https://www.lct.org/shows/ragtime/  The New York Public Library for the Performing Arts: https://www.nypl.org/locations/lpa 

The United States said on Sunday it conducted self-defense strikes on Iranian radar and drone control sites in response to aggressive actions from Tehran. The sites are located on Iran's Goruk and Qeshm Islands.The Department of Commerce has issued new guidance to prevent Chinese companies from obtaining advanced U.S. artificial intelligence chips, such as Nvidia's most sophisticated Blackwell processors, through overseas subsidiaries.

Anthropic vient d'annoncer son premier trimestre rentable — et les chiffres ne tiennent pas la route. Au même moment, des milliers d'investisseurs découvrent que leurs parts dans les startups IA n'ont peut-être jamais existé. D'un côté, des entreprises qui lèvent des milliards en affichant une rentabilité de façade — de l'autre, des investisseurs et des utilisateurs qui paieront le vrai prix quand la marée se retirera.===========================

Send us Fan MailIn Episode 252 of Book Talk, Etc. Tina and Hannah share their 20 most anticipated summer releases. They also share what they've been loving lately, their latest reads, and chat about how they built their summer reading lists.If you enjoy this commercial-free podcast, consider supporting us on Patreon! Your membership includes access to bonus episodes like What's in the Mailbag, Bookstore Browse: The Handsell, and Book Talk After Dark, invites to monthly community events like Mood Reader Happy Hour, and entry into our private Facebook group and Discord server- all for just $5 a month.Loving LatelyCountertop Ice Maker - Govee (T)Murdle Mystery Puzzle Activity Book (H)Latest ReadFive | Ilona Bannister (T)One Star Romance | Laura Hankin (H)SUMMER BONANZAPlant Lady | Minyoung KangHelpness | Jessica KnollLittle Wonder | Sophie Chen KellerTake What You Can | Naima CosterThe Spin | Faith GardnerEverything to the SeaThe Open Era | Edward SchmidtThe Great Wherever | Shannon SandersFamous Men | Julie BuntinKeep Them Close | David EllisNothing to My Name | Kangkang Li KovacsHeart of Glass | Jennifer HillierTropesick | Lauren OkieThe Secret Dinner | Raphael MontezWhistler | Ann PatchettThe Windsor Affair | Melanie BenjaminThe Children | Melissa AlbertThe Lowe Job | Grace AlexanderLand | Maggie O'FarrellStrangers Behind Closed Doors | Catherine Adel WestCurrent ReadsTrad Wife | Saratoga Schaefer (T)Nesting | Roisin O'Donnell (H)If you prefer other shopping options, you can find today's books on Bookshop.org or Blackwell's. Purchasing through these links supports us with a small commission, at no extra cost to you.Support the showLet's Connect... Email us at booktalketc@gmailBTE on YoutubeTina's TikTok , IG @tbretc YT @tbretcHannah's TikTok , IG @hanpickedbooksJonathan IG @infiltrate_jayPodcast IG @booktalketcRenee's Substack Newsletter , IG@Itsbooktalk

Send us Fan MailIn Episode 251 of Book Talk Etc., Tina and Hannah chat about some new book releases, as well as discuss the woes of starting a new book, and how they set themselves up for success when it comes to cracking open a new read.If you enjoy this commercial-free podcast, consider supporting us on Patreon! Your membership includes access to bonus episodes like What's in the Mailbag, Bookstore Browse: The Handsell, and Book Talk After Dark, invites to monthly community events like Mood Reader Happy Hour, and entry into our private Facebook group and Discord server- all for just $5 a month.Loving LatelyHobonichi Cousin A5 (T)Pilot Fixon Erasable PensJet PensAnalog Basket (H)Siece Campbell TikTokLatest ReadForget You Saw Her | Noelle Ihli (T)Ask for AndreaInto the Blue  | Emma Brodie (H)Book Talk and EDNRStarting a book wellA Good Person | Kirsten King (T)Homebound | Portia Elan (H)Nothing Tastes as Good | Luke Dumas (T) Ruins | Lily Brooks Dalton (H)Shelf Addition:Hollow Bones | Jodi Picoult (T)Love Felt Like This | Julie Olivia (H)If you prefer other shopping options, you can find today's books on Bookshop.org or Blackwell's. Purchasing through these links supports us with a small commission, at no extra cost to you.Support the showLet's Connect... Email us at booktalketc@gmailBTE on YoutubeTina's TikTok , IG @tbretc YT @tbretcHannah's TikTok , IG @hanpickedbooksJonathan IG @infiltrate_jayPodcast IG @booktalketcRenee's Substack Newsletter , IG@Itsbooktalk