Podcasts about Fireworks

The forensic psychologist who set off a firestorm in the courtroom on Monday will be back on the stand today. The judge ultimately ruled against a mistrial, but admonished the prosecution and instructed the jury to disregard Dr. Helibrun’s testimony about “mortal sin”. Testimony resumes today, first with Helibrun - and we can’t wait to hear how defense attorney Kevin Reddington handles any further questioning of him - and then with the state’s final rebuttal witness. We could see closing arguments as soon as Wednesday, and Lindsay Clancy’s fate should be in the hands of the jury by midweek. See omnystudio.com/listener for privacy information.

The forensic psychologist who set off a firestorm in the courtroom on Monday will be back on the stand today. The judge ultimately ruled against a mistrial, but admonished the prosecution and instructed the jury to disregard Dr. Helibrun’s testimony about “mortal sin”. Testimony resumes today, first with Helibrun - and we can’t wait to hear how defense attorney Kevin Reddington handles any further questioning of him - and then with the state’s final rebuttal witness. We could see closing arguments as soon as Wednesday, and Lindsay Clancy’s fate should be in the hands of the jury by midweek. See omnystudio.com/listener for privacy information.

The forensic psychologist who set off a firestorm in the courtroom on Monday will be back on the stand today. The judge ultimately ruled against a mistrial, but admonished the prosecution and instructed the jury to disregard Dr. Helibrun’s testimony about “mortal sin”. Testimony resumes today, first with Helibrun - and we can’t wait to hear how defense attorney Kevin Reddington handles any further questioning of him - and then with the state’s final rebuttal witness. We could see closing arguments as soon as Wednesday, and Lindsay Clancy’s fate should be in the hands of the jury by midweek. See omnystudio.com/listener for privacy information.

The forensic psychologist who set off a firestorm in the courtroom on Monday will be back on the stand today. The judge ultimately ruled against a mistrial, but admonished the prosecution and instructed the jury to disregard Dr. Helibrun’s testimony about “mortal sin”. Testimony resumes today, first with Helibrun - and we can’t wait to hear how defense attorney Kevin Reddington handles any further questioning of him - and then with the state’s final rebuttal witness. We could see closing arguments as soon as Wednesday, and Lindsay Clancy’s fate should be in the hands of the jury by midweek. See omnystudio.com/listener for privacy information.

The FC crew react to Chelsea's 3-2 win over Fulham and praise the front 3 attackers for a brilliant display. The guys also question why Robert Sanchez is still the club's starting keeper after a poor performance. Plus, Sid Lowe joins the show to provide an update on Julian Alvarez's future and explain why Barcelona might be better off without the forward. Learn more about your ad choices. Visit podcastchoices.com/adchoices

Join James and Charlie as they discuss the one-nil defeat from Thursday evening in-front of a packed out Hillsborough. Fireworks, flames but unfortunately the first defeat of the 26/27 campaign. This episode is sponsored by Fibrely.

August 16, 2016. The Hip arrived in the Hammer, and three fans who were in the building take jD back inside - ten years to the day.Tuesday, August 16, 2016. FirstOntario Centre in Hamilton, four days out from the end of everything, and The Tragically Hip pulled into a city that watched them play to nobody once and never forgot it.Ten years to the day, jD sits down with three fans who were there for that room: Gino from Buffalo, Matt from Buffalo, and John from Mississauga. Two of them had never been on a podcast before in their lives. One of them never made it through the doors.The songs first, because the songs are always first. Gino takes 'Poets', the last one of the main set, the one where Gord stopped and told the room about The Hip's first Hamilton gig - a club, zero people, and an owner who told them nobody gets paid until every song gets sung. So they played another hour to an empty room. Ten years and a lifetime later, Gord looked out at a full arena and, the way Gino heard it, said something close to look at us, look at where we are now. Us, collectively. Hamilton and The Hip.Matt takes 'Scared', and the line it's been a pleasure doing business with you, and the small jump in the video he was shooting - the moment his buddy behind him lost his knees and fell into his back. John takes 'Courage', because of what it took Gord to get in the van at all. A residency somewhere would have been hard enough. He went across the country instead.Then the tickets, and three completely different scrambles. Matt's Buffalo crew split the presale up like a heist - one guy on London, one on each Toronto night, Matt on Hamilton - with orders that whoever got in bought all four. Matt was the only one who got into the queue, and he had four seats about ten minutes in. Gino got shut out in the fifteen minutes it took the show to vanish, then his friend Greg found VIP seats hours later and warned him they'd be expensive. Gino's answer: I don't care. Third row on the floor. If it was within a 500-mile radius, you weren't stopping me.The day itself: a friend on standby in Niagara Falls, U.S., in case one of Matt's four didn't clear the border, a cash-only scalper outside the arena with PDF printouts and a very loose grip on legitimacy, and their buddy Scott laying out most of the money for a pair of those dubious tickets so a friend from Lockport could drive up and slide in during the first song. Gino owns his company, so Gino gave himself the afternoon off, got to Hamilton by mid-afternoon, and ate Portuguese chicken at the Charred Rotisserie House down the road from the barn. We ate like kings. He has been back a few times since.And then there is John, who was downtown at the Toronto Stock Exchange that Tuesday, letting a vendor in to do a routine job. At 2:30 the vendor crashed the system. Eight hours to bring it back. John is standing on a floor you get onto with retina scans and palm prints, he cannot leave the vendor alone, and his company supplies data to the exchange at thousands of dollars an hour. He had a ticket in his pocket and a family carload heading to Hamilton without him. So he stood there and asked himself the only question he had. What would Gord do? What's the Canadian thing to do? He thought about '38 Years Old' - which is the plate on his car - and the father in that song who turns his own son in because it is the right thing. John stayed at his post. He called his sons and told them to give the ticket away. Every August 16th since has been one of the saddest days of his year.Inside the building, the two who made it describe the same night from two heights. The small stage, the band pressed in around each other, a band of brothers making life as easy for Gord as they could. Gord finding faces one at a time, holding them, nodding, gesturing, taking what Gino calls little memory snippets for himself. The monitor on the floor by his feet for the words that wouldn't come. Gino has seen this band more than forty times, and he puts Hamilton at the top for sheer energy. Matt's read is sharper still: the energy was all there, the memory just couldn't always match it.The set list gave Hamilton 'Fifty-Mission Cap', 'Eldorado', and 'At the Hundredth Meridian' early, four from "World Container" as the night's rare record, and the best four of the six "Man Machine Poem" songs that made the tour. 'Poets' to close the main set. 'Gift Shop', 'Don't Wake Daddy', and 'Ahead by a Century' in the first encore. Then 'Fiddler's Green' and 'Twist My Arm' in the second, and a song Gino had never heard live anywhere. Matt has one regret and it is a beer. He went for one during 'Escape Is at Hand for the Travellin' Man', found the concourse narrow and the taps in the basement, and came back up on 'Fireworks' having missed ten minutes he will never get back. So there's that.Gino's detour is worth the price of admission on its own - Highland Bowl in Rochester, a grass amphitheatre, a downpour with thunder and lightning, the band stopping for maybe ten minutes, Gord calling out to the lovers of music who stayed, and grown adults sliding down the wet hill on their bellies toward the stage. Second best night he ever had with this band. Hamilton was first.Then the lights. Gino describes an arena that felt like one family at a memorial service, except it wasn't a somber memorial service - it was everybody holding everybody else up while Gord stood there telling them from the stage that it was going to be okay. Half an hour after the last note the set list was sitting within arm's reach, so Gino took it, and they passed it around for photographs. Nobody wanted to leave. Matt walked over to the King George pub with seven other people and found something else entirely. It almost had the feeling of a wake. Nobody wanted another round. Everybody was thankful. Everybody knew.August 20th finds all three of them in very different rooms. John rigged an aerial in the backyard of his brother's cottage outside Montreal, set both PVRs at home, and asked the local bar in a small French-Canadian town whether they'd be showing it - they had no idea what he was talking about. Don't move that coat hanger outside. Matt spent three hours assembling a children's play kitchen for his two-year-old daughter with the broadcast on, which should have taken one. Gino watched with a dozen guys at a friend's place, where Greg leaned over and said the thing nobody else had thought: doesn't he look nervous? He did. Of course he did.The landing belongs to John, though. There was no goodbye for him that night, so it came slowly, over days and weeks, like a song fading out instead of ending. And it kept going. He has a nine-year-old granddaughter now who knows the words, and he took her to Chudleigh's in Milton to see the Practically Hip so she could dance to them. Another generation, flying the flag.THE PANEL• Gino from Buffalo - forty-plus shows with The Hip, third row on the floor in Hamilton on a VIP ticket his friend Greg found hours after the show sold out, and a plate of Portuguese chicken beforehand he still drives back for. He came home with the set list. He listens to this band every single day.• Matt from Buffalo - second row of section 120, with a photo from that night still hanging behind him on camera. He ran the Buffalo crew's presale plan, got the only four seats any of them landed, filmed 'Scared' through the moment his friend's knees went, and lost ten minutes of the night to a beer line in the basement.• John from Mississauga - the one who didn't get in. A crashed system on the Toronto Stock Exchange floor at 2:30 on show day, a ticket handed off to somebody else, and a decision he made by asking what Gord would do. He watched August 20th through a coat hanger aerial at a cottage. His granddaughter is nine and already knows the words.SOURCESSet list for August 16, 2016 at FirstOntario Centre in Hamilton (now TD Coliseum) - source: setlist.fm.THE GATHERING IN KINGSTONAugust 20 to 23, 2026, alongside Forever Hip - a mixer at the Merchant Pub, Choir! Choir! Choir! in Springer Market Square, a listening party for "The Tragically Hip Live, July 22 - August 20, 2016", a Long Time Running screening, and a farewell brunch at Morrison's. The CBC broadcast re-airs on the 22nd. Email jd@tthpods.com to get on the list. And if you spot a purple shirt or a purple lanyard in Kingston that weekend, that's a safe person to talk to.GEDFEST TORONTOSaturday, October 17, 2026 at the Horseshoe Tavern, with Grace, 2. Raising money for Campfire Circle and the Murphy Family Fund in Precision Medicine in Pancreatic Cancer. Tickets: facebook.com/GEDfestYYZThanks to Gino, Matt, and John, because without them, there's not a show. Next stop: Ottawa.MORE FROM THE TOURWinnipeg: https://redcircle.com/shows/a01e8de3-1a04-4a44-ab9c-8fc6e3cb7474/ep/bf776e9a-c431-41bc-b605-f73fff1155d3Calgary Night 2: https://redcircle.com/shows/a01e8de3-1a04-4a44-ab9c-8fc6e3cb7474/ep/98943c27-bdae-4f56-84d4-a57754be7719SUPPORT THE WORKThe tip jar: buymeacoffee.com/tthtop40Web: home.tthpods.com | Facebook: community.tthpods.com | Instagram: @tthpods | YouTube: youtube.com/@tthpods | Newsletter: subscribe.tthpods.com | Email: jd@tthpods.com#TheTragicallyHip #GordDownie #WorldContainer #RoadApples #HipFans #TTHPodsSupport this podcast at — https://redcircle.com/tthtop40/donationsAdvertising Inquiries: https://redcircle.com/brandsPrivacy & Opt-Out: https://redcircle.com/privacy

The prodigal son has returned! Dane is back this week and he is bringing the heat! Listen in this week as the guys discuss bartending for the women's ministry, getting slandered at work, and all things related to fantasy football. Additionally, we attempt are hardest challenge yet...naming 100 Bible characters in 10 minutes.---Please message us if you would be interested in joining our Fantasy Football league. You can do that via email or Instagram DM.---If you want an Until Next Week Podcast shirt for $20 (+shipping), email untilnextweekpodcast@gmail.com or DM us on Instagram. (Large & XL available)---Please follow our Instagram & TikTok to stay updated on all things podcast and make sure to send us a voice message via Instagram DM to be featured on one of our next episodes.https://www.instagram.com/untilnextweekpodcasthttps://www.tiktok.com/@untilnextweekpodcast---Please leave us a 5 STAR REVIEW on both Spotify and Apple for a chance to be mentioned on a future episode.---Get $10 off at Friday Pickleball with a minimum order of $95: [MUST CLICK LINK BELOW]https://www.fridaypickle.com/discount/SAMUEL14434---Get 10% off your order for a Bloom Card with the following code: SAMUEL14434https://bloom.inc---Key words for the algorithm: Clean Podcast, Clean Comedy, Friday Pickleball, Ghostrunners Podcast, Correct Opinions Podcast, Tim Hawkins Podcast, Becoming Something Podcast, Youth Group Chronicles Podcast, Almost Athletes Podcast with Dude Perfect, Wisconsin, Fireworks, The Odyssey, Grandpa's Birthday, Dirty Soda Bartending, GOAT Award, New Roommates, Wiffle Ball Playoffs, Softball Drama, Wild Injury Predictions, Jayden Daniels NIL, and Gulf Shores.

Japanese fireworks may soon get a bit more eco-friendly.Read more about the fireworks here  ★ Support this podcast on Patreon ★

For more coverage on the issues that matter to you, download the WMAL app, visit WMAL.com or tune in live on WMAL-FM 105.9 from 9:00am-12:00pm Monday-Friday  To join the conversation, check us out on Twitter @WMAL and @ChrisPlanteShow Learn more about your ad choices. Visit podcastchoices.com/adchoices

ITB's Eagles beat reporter Andrew DiCecco gives his insights from covering the Eagles on a daily basis.In this episode, Andrew goes inside his camp practice observations Thursday, including a fight between Moro Ojomo and Fred Johnson.Subscribe to our Patreon Channel for exclusive information not seen or heard anywhere else and become among smartest Birds fans out there (just ask our members!!) + get all of our shows (including this one) commercial free!!https://www.patreon.com/insidethebirds► Sign up for our newsletter! • Visit http://eepurl.com/hZU4_n.►Support our sponsors!!► Camden Apothecary: https://camdenapothecary.com/► Download Cash App Today: https://cash.app/ #CashAppPod● Cash App is a financial services platform, not a bank. Banking services provided by Cash App's bank partner(s). Prepaid debit cards issued by Sutton Bank, Member FDIC. Cash App Visa® Debit Flex Cards issued by Sutton Bank, Member FDIC, and The Bancorp Bank, N.A., pursuant to a license from Visa U.S.A. Inc. See terms and conditions for the Sutton prepaid card, Sutton debit flex card, and Bancorp debit flex card. Savings provided by Cash App, a Block, Inc. brand. Visit cash.app/legal/podcast for full disclosures.► Don't sleep on Ultra Pouches. New customers get 15% off with code BIRDS at takeultra.com! #UltraPouches #adFollow the Hosts!► Follow our Podcast on Twitter: https://twitter.com/InsideBirds► Follow Geoff Mosher on Twitter: https://twitter.com/geoffpmosher► Follow Adam Caplan on Twitter: https://twitter.com/caplannfl► Follow Andrew DiCecco on Twitter: https://twitter.com/andrewdiceccoNFL insider veterans take an in-depth look that no other show can offer! Be sure to subscribe to stay up to date with the latest news, rumors, and discussions.For more, be sure to check out our official website: https://www.insidethebirds.com.

Grayson is confused why his date Coco won't call him back after a great first date... Find out why he got ghosted!See omnystudio.com/listener for privacy information.

Grayson is confused why his date Coco won't call him back after a great first date... Find out why he got ghosted!See omnystudio.com/listener for privacy information.

My guest today is Eric Vishria, a General Partner at Benchmark.  Eric has spent his career in software and cloud, and few people know the history of these markets as well as he does. What makes him special is his ability to use that history to make sense of today.  We discuss what the rise of AWS teaches us about AI, what he has learned from investing in Fireworks, Sierra, and Cerebras, and how the criteria for winning have changed for founders and investors.  Please enjoy my conversation with Eric Vishria. For the full show notes, transcript, and links to mentioned content, check out the episode page ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠here⁠⁠⁠⁠⁠.  ----- Become a Colossus member to get our quarterly print magazine and private audio experience, including exclusive profiles and early access to select episodes. Subscribe at ⁠colossus.com/subscribe⁠. ----- ⁠Ramp's⁠ mission is to help companies manage their spend in a way that reduces expenses and frees up time for teams to work on more valuable projects. Go to⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ ⁠ramp.com/invest⁠⁠ to sign up for free and get a $250 welcome bonus. ----- Trusted by thousands of businesses, ⁠Vanta⁠ continuously monitors your security posture and streamlines audits so you can win enterprise deals and build customer trust without the traditional overhead. Invest Like the Best listeners get a special offer of $1,000 off Vanta when you go to ⁠vanta.com/invest⁠.  ----- WorkOS⁠ is the infrastructure B2B and AI-native companies use to sell to enterprise. It covers everything enterprise security requires: SSO, SCIM, RBAC, Audit Logs, AI governance, and more. Trusted by 2,000+ fast-growing companies, including OpenAI, Anthropic, Cursor, and Vercel. ----- Rogo is the AI platform for finance. They're building agents for Wall Street that are trained to understand how bankers and investors actually do work: from diligence and modeling, to turning analysis into deliverables. To learn more, visit rogo.ai/invest. ----- ⁠Ridgeline⁠ has built a complete, real-time, modern operating system for investment managers. It handles trading, portfolio management, compliance, customer reporting, and much more through an all-in-one real-time cloud platform. Visit⁠ ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ridgeline.ai⁠. ----- Editing and post-production work for this episode was provided by The Podcast Consultant. Timestamps: (00:00:00) Welcome to Invest Like The Best (00:02:20) Learning the World Through Fireworks (00:05:42) AWS Was Going to Eat Everything (00:07:40) The Zero-Sum Thinking Trap (00:09:01) Comparing Cloud and AI Adoption (00:11:03) Becoming Enterprise's AI Sherpa (00:13:05) Building Sandcastles (00:14:55) The Return to Being Technical (00:17:13) The Shifting Competitive Frontier (00:22:10) Why the Old Playbook Fails (00:27:53) Energy as the Binding Constraint (00:29:38) The Cerebras Story (00:37:57) The Virtue of Productive Naivete (00:39:19) What Robotics Still Needs (00:45:58) What Makes a Great Board Partner (00:51:13) Raising A Growth Fund (00:55:39) What the Big Winners Taught Him (00:57:37) Hard Work Versus the Hole-in-One (00:58:38) The Best Reasons to Go Public (01:01:09) Debates Inside Benchmark (01:02:16) What If It All Works (01:03:35) What Geoff Hinton Got Wrong

(0:00) Patrick Mahomes Week 1 expectations, Should Josh Allen feel guilty over Sean McDermott's firing? (27:09) Will Caleb Williams be the best QB in his class? (42:39) How will Jaxson Dart perform? (48:43) Josh Allen concerns, How will Patrick Mahomes come back from his injury? (01:06:04) Expect the Commanders and Jayden Daniels to bounce back? (01:14:54) Kyren Williams joins (01:25:57) Cowboys ‘fireworks', Will the Eagles have a better season this year?  (01:48:45) Bears schedule predictions, How high is Caleb Williams' ceiling? (02:00:37) Will a motivated Joe Burrow get the Bengals back on track? (02:07:23) Raiders already building a culture after training camp fight? Learn more about your ad choices. Visit podcastchoices.com/adchoices

In this episode of Sixtysometing, your host, Grace Taylor Segal, recounts taking her three young grandchildren to their first baseball game, an Anaheim Angels game, with her husband Aaron, daughter Juliet, and Aaron's brother Elliott, staying for postgame fireworks and spending the night at a nearby hotel.Though she worried about meltdowns and attention spans, the kids did great, enjoying snacks, souvenirs, and the overall ballpark experience more than the game itself, and the fireworks captivated everyone. Grace reflects that the day was really about sharing places and experiences tied to family history—Aaron's childhood memories at Angel Stadium and her own baseball roots with the St. Louis Cardinals—and how grandparents can “open the door” to what they love without pressure.She emphasizes showing up, being flexible about what “success” looks like, starting traditions simply, and not waiting for a perfect time to make memories.* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *Hey Friends! It's me, Grace! I just want to thank you for listening. I hope you'll let me know what you think about the podcast and if any particular episodes resonate with you.Listed just below here is my contact information and all of the social channels where you can find me, as well as the link to our Facebook Group. Contact InfoGrace Taylor SegalEmail: grace@gracetaylorsegal.comFacebook: 60something Page(https://www.facebook.com/profile.php?id=61553062496332)Instagram: @60somethingpodFacebook Group: 60Something Podhttps://www.facebook.com/groups/1665326354000332CreditsSixtysomething Theme SongMusic & lyrics by Lizzy SanfordVocals by Lizzy SanfordGuitar: Lizzy & Coco SanfordTo Leave a Review: On Apple Podcasts (no link is available--here are the steps)Open the Apple Podcasts app.Search for the podcast you want to review.Select the podcast show page (not an individual episode).Scroll to the bottom of the page.Click "Write a Review".On the Sixtysomething Podcast Websitehttps://www.sixtysomething.net/reviews/new/Thank you so much for taking the time to review the show!Timestamps:00:00 Welcome and Setup00:13 First Ballgame Plan02:51 Snacks Treats and Surprises04:35 More Than Baseball06:58 Special Places Memories10:06 Grandparenting Perspective11:28 Kids See Wonder14:24 Let Them Love Differently15:54 Lessons From Grandparents18:34 Seeds and Influence21:57 Redefining Success24:05 Traditions Start Small26:32 Fireworks and Time28:29 Dont Wait Make Memories31:37 Wrap Up and Next Episode

Tyson Alger and Justin Myers kick off August with fall camp in full swing and a fake controversy leading the Oregon news cycle: Dylan Raiola's perfectly standard “I like it here better” transfer quote somehow becomes a national talking point. From there, they zoom out on college football's weirder traditions: Do we actually still need the coaches' poll? Are conference media days useful, or just content cosplay and an excuse for writers to expense a trip? Is all the “culture/chemistry/retreat” talk any different from what every other program is doing?Then the good stuff: Bend vs. the coast, what it's like to be a civilian at a packed Zach Bryan show at Autzen (fireworks and all), and the emotional journey of being awarded — and then mentally defending — a coveted work parking spot.

It was a shortened day of testimony, but it was a powerful day in court. The prosecution called five witnesses to the stand, including several more medical and scientific experts, one of whom evoked a fiery exchange between Lindsay Clancy’s defense attorney and prosecutors. The star witness of the day was the Clancy family’s former nanny who was with Lindsay and her children in the months leading up to the murders. Elaine Rossi testified that she was aware Lindsay was struggling with postpartum, but was deeply involved in caring for her children and was a wonderful mom.See omnystudio.com/listener for privacy information.

It was a shortened day of testimony, but it was a powerful day in court. The prosecution called five witnesses to the stand, including several more medical and scientific experts, one of whom evoked a fiery exchange between Lindsay Clancy’s defense attorney and prosecutors. The star witness of the day was the Clancy family’s former nanny who was with Lindsay and her children in the months leading up to the murders. Elaine Rossi testified that she was aware Lindsay was struggling with postpartum, but was deeply involved in caring for her children and was a wonderful mom.See omnystudio.com/listener for privacy information.

It was a shortened day of testimony, but it was a powerful day in court. The prosecution called five witnesses to the stand, including several more medical and scientific experts, one of whom evoked a fiery exchange between Lindsay Clancy’s defense attorney and prosecutors. The star witness of the day was the Clancy family’s former nanny who was with Lindsay and her children in the months leading up to the murders. Elaine Rossi testified that she was aware Lindsay was struggling with postpartum, but was deeply involved in caring for her children and was a wonderful mom.See omnystudio.com/listener for privacy information.

It was a shortened day of testimony, but it was a powerful day in court. The prosecution called five witnesses to the stand, including several more medical and scientific experts, one of whom evoked a fiery exchange between Lindsay Clancy’s defense attorney and prosecutors. The star witness of the day was the Clancy family’s former nanny who was with Lindsay and her children in the months leading up to the murders. Elaine Rossi testified that she was aware Lindsay was struggling with postpartum, but was deeply involved in caring for her children and was a wonderful mom.See omnystudio.com/listener for privacy information.

Disney's Lakeshore Lodge opens summer 2027, and today's Disney Parks Blog article confirmed new dining and fireworks spaces coming to the resort, including a new table-service restaurant, a quick-service restaurant, a pool bar, and a terrace. Disney's Lakeshore Lodge will be inspired by Pocahontas (from Walt Disney Animation Studios) and the natural landscape surrounding Bay Lake at the Walt Disney World Resort. Get ad-free episodes, bonus episodes, in-depth news analysis, and premium content at patreon.com/imaginationskyway. To plan a trip, be sure to work with KMV Travel.   Read Matt's Imagineering column in WDW Magazine.   Imagination Skyway is a Disney Parks and Imagineering podcast. Episodes explore attraction design, recap Disney news, and dive into the stories behind the magic, including interviews with Disney Imagineers, Disney Legends, and other Disney creators. Not affiliated with or endorsed by The Walt Disney Company. Disney is a trademark of The Walt Disney Company.   Tag me and join the conversation below. Instagram: www.instagram.com/imaginationskyway Facebook: www.facebook.com/imaginationskyway YouTube: https://www.youtube.com/@imaginationskyway Email: matthew.krul@imaginationskyway.com  How to Support the Show Share the podcast with your friends Rate and review on Apple Podcasts or Spotify Join our Patreon Group - https://www.patreon.com/imaginationskyway Enjoy the show!

Epcot or Magic Kingdom? Dole Whip or Mickey pretzel? Castaway Cay or Lighthouse Point? This week Lisa puts Steve and Debbie in the hot seat for a rapid-fire (okay… not-so-rapid-fire) round of "This or That." Two options, gut-instinct answers, and a whole lot of friendly disagreement — with plenty of insider reasoning packed in along the way. Some are total no-brainers, a few get controversial, and Lisa tries (and mostly fails) to predict everyone's answers. It's a fun, easy listen packed with real opinions from travel agents who live and breathe Disney. Play along at home and see how many you match! In This Episode Intro — (0:00) A smaller crew this week — Caitlyn's on baby watch and Sharpie's out — so Lisa runs the game with Steve and Debbie. Parks & Planning — (1:05) Magic Kingdom or Epcot? • Rope drop or stay late? (plus everyone's favorite ride to look at lit up at night) • Lightning Lane vs. standby • Disney World or Disneyland? (and yes, Steve really does drive from Atlanta) • Table service or quick service? Rides & Coasters — (7:56) Space Mountain or Big Thunder? • Rise of the Resistance — Disneyland vs. Disney World • Slinky Dog Dash or Seven Dwarfs Mine Train? • Pirates of the Caribbean East or West? • Matterhorn or Expedition Everest? Cruise & Castaway — (8:30) Castaway Cay or a Caribbean port? • Castaway Cay or Lighthouse Point? • Cruise dinner or brunch? • Newest ship or the classics? • Concierge or excursions? • Alaska or Caribbean? • Poolside day or excursion day? Resorts, Snacks & Sweets — (11:54) On-property or offsite savings (Debbie would pick All-Star Sports over the Ritz!) • Dole Whip or Mickey pretzel? • Beignets at Port Orleans or churros on Main Street? More Park Favorites — (14:26) Park hopper or one park? • Fireworks or one more ride? • Animal Kingdom or Hollywood Studios? • Genie+ or the old free FastPass+ days? • Haunted Mansion — Florida or the Disneyland Nightmare Before Christmas overlay? • Disney Springs or Downtown Disney? • Villains or princesses? • Monorail or Skyliner? • Adults-only trip or bring the whole family? • Very Merry Christmas Party or Not-So-Scary Halloween? ⚡ Lightning Round — Both Answer at Once! — (27:30) Splash Mountain or Tiana's Bayou Adventure? • Wishes or Happily Ever After? • IllumiNations or Luminous? • The dearly missed Magical Express • Extra Magic Hours or today's early entry + evening hours? • The Great Movie Ride or Runaway Railway? • Maelstrom or Frozen Ever After? • Ellen's Energy Adventure or Guardians Cosmic Rewind? • Paper FastPass or booking in the app? • Soarin' Over California or Around the World? • MaxPass nostalgia • Tower of Terror or Guardians Mission Breakout? • Old or new Test Track? • Main Street Electrical Parade or Paint the Night? Wrap-up — (44:10) Thanks for playing along!

Episode DescriptionNumber 31 on the countdown. First zoo animal on Prozac. JD is joined by Duxoop from Toledo — broadcasting from the actual birthplace of Gus the polar bear — plus Andrew from Winnipeg and Dave from Montreal to dig into one of the most quietly beloved songs in the Hip catalogue.The panel gets into the Adam Kasper production that makes In Between Evolution sound the way it does, Rob Baker's untouchable outro solo, the Hip Museum's wonderfully unhinged George W. Bush theory, and the wordplay hiding in plain sight — because it was never what's troubling Gus. It's what's troubling us.Fair warning: this one gets emotional. Duxoop tells us what the lyrics did to him last year from a nursing home wheelchair, and Dave tells us about his friend Al — nickname Gus — and the cover band he's fronting at Al's memorial this Friday. Then the whole thing derails beautifully into a full inventory of every animal in the Hip discography.Explicit language. Recorded live on YouTube.On the PanelJD — hostDuxoop from Toledo — self-described part polar bear; born-in-Toledo credentials, same as GusAndrew from Winnipeg — arrived with facts and a News Radio shirt; host of Soup Full of RadioDave from Montreal — saw the Hip in Central Park; frontman of brand-new Hip cover band Colonel MustardTale of the TapePulled from The Hip Handbook, which amalgamates HipBase and setlist.fm data. Album In Between Evolution (2004) Producer Adam Kasper, recorded in Seattle Single Third single from the record Countdown position No. 31 on the Top 40 Times played live 327 First played October 27, 2003 Last played August 18, 2016 — the Ottawa show As an opener Once As an encore Five times Poll Results Love it 52% Like it 36% Tolerate it 10% Skip it 1% Never heard of it 1% Possibly the most varied — and most indecisive — poll spread of the series so far. Duxoop's verdict: "That's crazy to me. These numbers should be way more lopsided toward the top."Live Version FeaturedOttawa, 2004 — the Grey Cup halftime show. Andrew remembers watching it live and everyone around him asking why the band was opening the biggest gig of the year with a brand-new song. (They went into Courage right after.)What We Talked AboutThe Album That Everyone Slept OnIn Between Evolution debuted at No. 1 in Canada on roughly 22,500 first-week copies — then got knocked off the top by Avril Lavigne's new album.Certified platinum in Canada that September. Dave notes it was one of the last Hip records to move real units.Where does this song sit in the career arc? The panel lands on "the end of the middle" rather than the dawn of the final third.Andrew's take: Gus is the first song on the record that sounds like the Hip you already knew — you lean back in the chair. Heaven Is a Better Place Today, Summer's Killing Us and Vaccination Scar are louder neighbours, so Gus hides in the first five.Adam Kasper and the Seattle SoundAndrew looked up Kasper's credits and it clicked: Soundgarden's Down on the Upside, Queens of the Stone Age's Songs for the Deaf, Pearl Jam's self-titled (the avocado one), Foo Fighters' One by One.Same studio was hosting the Big Fish and Eternal Sunshine of the Spotless Mind soundtracks around the time this record was being made.The panel agrees Gus has a "weirdly familiar, warm Hip sound" — it could have sat on Day for Night with different production, or on Music at Work. Maybe it was on the shelf a while.Rob Baker's OutroUniversal agreement: whichever live version you pull up, Baker is on fire. "A real showcase for him."Dave's sharpest observation of the night — Gord never rants over this solo the way he does over others. It always gets its space, and that may be the only ad-lib that separates the live song from the record.Gus HimselfBorn 1985 at the Toledo Zoo, died August 27, 2013. A 700-pound icon of the Central Park Zoo, seen by over 20 million visitors.Began swimming obsessively — up to 12 hours a day — in the 1990s. Reporters called him neurotic, depressed and flaky, and he became a symbol of the stress of living in New York.The first zoo animal in history to be treated with Prozac. Also got a jacuzzi and, at one point, two girlfriends. Duxoop: "Why is he the one on Prozac and not me?"Gord wrote the song in 2004, nine years before Gus died — so this wasn't an obituary. He'd seen the bear and then done the second-sourcing, as Gord always did.There's a children's book called What's Worrying Gus? The True Story of a Big City Bear. On Amazon, the cover has Gus on a therapist's couch.The Toledo ConnectionThe Hip played the Toledo Zoo twice in the '90s — where Gus was born.When they played Toledo in 2004, Gus was the show opener. The stats say it only ever opened once, so that was absolutely on purpose.The Bush TheoryThe Hip Museum reads Gus as partly about George W. Bush and the invasion of Iraq. Andrew read the paragraph aloud; it describes the two as "lovable, loutish, and incredibly accomplished guys… not bad for a pair of dudes with limited mental capacity."JD's verdict: lots of love for the Hip Museum, but that one's a reach.Gus / UsDave's thesis: he doesn't hear a bear at all. "It's not what's troubling Gus. It's what's troubling us." He's been on that line for twenty years.JD points out the Gordism that makes it work — dropping the G, so "troublin' us" and "troubling Gus" are phonetically the same thing.Duxoop extends it: we've all watched a zoo animal pace or swim figure eights, and that should be troubling us.On the lyrics generally: Gord's word-smithery rewards straining to hear it, and the panel confesses to mumbling through parts for two decades. Compared to Nautical Disaster — all story, no chorus, no pattern — Gus is manageable, but only just.Duxoop's StoryLast year, when the countdown hit No. 31, Duxoop was three months into a nursing home stay with medical issues. Twenty-odd years earlier he'd started going to karaoke — big, scary-looking, socially awkward, mistaken for the bouncer. Now he was in his 60s, in a wheelchair, and the line landed like a ton of bricks: old and grey, and no one's afraid anymore.He sang it at the nursing home. "I don't think I've ever sung a Hip song with more feeling than I do with Gus." Grace, Too he can't fake — he feels like a schmuck trying to match Gord's intensity. Gus he doesn't fake.And after a week of thinking about it for this episode, he made a bold statement: Gus is now his favourite Tragically Hip song, over Grace, Too.Dave's Story — Al, Gus, and Colonel MustardDave's friend Al Morton died suddenly last November. They bonded over the Hip and golf at Dave's club. Al's nickname was Gus.In their fantasy football pick'em league, Al's picks stayed on the board for the rest of the NFL season. "A real lump in my throat."Al's memorial is this Friday at the Dunany Country Club in Wentworth, Quebec — halfway between Ottawa and Montreal. Ashes scattered around the course in the morning; speeches and a gathering at night, billed as his "after par-ty."The family asked for seven or eight Hip songs. Dave and friends have twenty ready, after two nights jamming at his buddy Chris's place. He made them learn Gus — he couldn't imagine doing a show for Al without it.The band is named Colonel Mustard, after Al's imported 1970s Mercedes. Wentworth, Quebec's Tragically Hip cover band. Dave waited 45 years to do this.His approach: very much not trying to be Gord. "I'm just trying to sing it with my truth."Hardest songs to sing so far — Fireworks (the National Fitness Program cadence) and Fully Completely (Gord's offbeat entrance in the middle section). Bassist John Roy has been cueing him in.Also on the setlist: Grace Too, Wheat Kings, The Good Life, Three Pistols.One BucketDave loves the way Gord says "whippoorwill" — audible in the Grey Cup version JD played. He'll be attempting it Friday."Used to be enough to make every bird stop singing." There's a lot of beauty in this song.Dave also went to bat for Sharks as an all-time Hip song — lyrically perfect, everyone gets to shine — while fully admitting he's a party of one on that.The Countdown NeighboursNo. 30 The Darkest One · No. 31 Gus · No. 32 Emperor Penguin — with Pigeon Camera at No. 28. As JD put it: their arctic phase.The Animal Album (An Idea That Got Away From Everyone)Duxoop floated it — Weird Al did a food album and a TV album, so why not a Tragically Hip animal album? The panel then attempted a full inventory:Emperor Penguin · Pigeon Camera · Dire Wolf · Problem Bears · Thompson Girl (polar bear reference) · The Bear on Music at WorkTiger the Lion · The Luxury (zoo lion) · Little Bones (a teeny tiny little cat) · Sled Dogs After DinnerWheat Kings (loon at the beginning) · Bumblebee · Leave ("in this contest, a concave nest") · Sharks · Piano Spider · I'm a Werewolf, Baby · the killer whale tank one on the box setVerdict: someone needs to make this playlist.Quotable Moments"It's not what's troubling Gus. It's what's troubling us. And that's just the best." — Dave from Montreal"Gord Downie sells the fuck out of this song. He sells the flakiness of the bear. He sells the confusion of the person watching it being like, why aren't you like a real bear?" — Dave from Montreal"It's funny — we're happy to hear a song about a sad animal." — Andrew from Winnipeg"Gus is like you're leaning back in your chair and relaxing a bit." — Andrew from Winnipeg"Grace, Too I feel like I'm faking it. Gus, man, I don't fake that one." — Duxoop from Toledo"It's up to us now to keep these songs alive." — Dave from Montreal"I live for that rare occasion where somebody goes, who was that song by? I got to look up that artist — and I get to make a new Hip fan." — Duxoop from ToledoMentioned in This EpisodeThe Hip Handbook — free Top 40 breakdown with episode links: thehiphandbook.tthpods.comHipBase and setlist.fm — live-stat sources behind the Tale of the TapeThe Hip Museum — source of the Gus / George W. Bush readingWhat's Worrying Gus? The True Story of a Big City Bear — the children's bookHipeponymous — where the Summer's Killing Us video ended upAdam Kasper's discography — Down on the Upside, Songs for the Deaf, Pearl Jam, One by OnePlugs & HousekeepingDuxoop from ToledoA friend of Duxoop's was hit hard by the death of Glen Hansard, who was killed in a motorcycle crash outside Dublin on July 29, 2026, at 56. There's a local Toledo tribute on Friday collecting donations for the Dublin Simon Community, the homelessness charity Hansard cared deeply about. Duxoop, who was himself homeless for a couple of years recently: "I would definitely promote that."Andrew from WinnipegSoup Full of Radio — weekly on UMFM in Winnipeg. Every episode is the next letter of the alphabet. Next week: songs starting with "the." The week after: all covers, all Tom Jones covering other people.Head Full of Radio — the podcast, where Andrew plays Radiohead to a guy who's never heard it. Google "Soup Full of Radio" and you'll find both.Dave from MontrealGoogle him for his Hip-related writing; Instagram @DaveKaufman1Dave used his plug to point people at JD's buy-me-a-coffee instead — noting how many unpaid hours go into this and how much the shows are worth to him. JD's reiteration: this stuff is free forever, don't worry if you can't swing it.The Gathering — KingstonRegistration deadline is Friday for all activities. (If you're hearing this on the podcast feed, you've likely missed it.)Three days in Kingston for the 10th anniversary of the Hip's final show: a brunch, an album listening party, a first-day mixer, and hotel blocks. Come for one day or all three.On the 22nd, the national celebration screens in Springer Market Square — side by side with other fans, the way everyone watched it ten years ago.Live stream of the final episode of Your Hipstories: Reflections from the Final Tour. Location to be announced in the coming weeks.Panel call-out: three spots on that final panel go to people who submit a video under one minute explaining why they should be on it, by sharing their favourite memory of August 20th.120 people signed up so far. Kingston is finally embracing and showcasing this band — the banners are up, and JD hopes they're permanent.Next WeekPigeon Camera. Brand new panel, brand new song, a biggie. Live on YouTube, Wednesday at 8 o'clock.Support this podcast at — https://redcircle.com/tthtop40/donationsAdvertising Inquiries: https://redcircle.com/brandsPrivacy & Opt-Out: https://redcircle.com/privacy

Should American companies be worried about Chinese open-source AI models?Lin Qiao, CEO of Fireworks, doesn't think so.In this episode, she joins Lukas Biewald to talk about why she believes the industry is at a turning point, one that calls for more open intelligence, not less.They cover how Fireworks now processes more tokens a day than OpenAI's API, why she thinks the future belongs to specialized models built on private company data rather than general-purpose ones, and why she believes OpenAI and Anthropic should be open-sourcing their own models too.Connect with us here:Lin QiaoFireworksLukas BiewaldWeights and Biases

GET TO THE FIREWORKS FACTORY•Unpacking some questionable comics. •Batman Day 2026! •Upcoming events, plus a tease! •Cakeworthy. •Upcoming exterior transformation. •End-of-the-year ordering structure. •Comics talked about in this episode:      SIX OF US #1      HAMMERFIST #1      TERMINAL #1   ---------- Contest of Challengers #791 This episode is dedicated to David Harper's Eisner win! Theme: Adam WarRock (with Mikal kHill) Intro: James VanOsdol (with Danhausen and Chris Jericho) Outro: James VanOsdol "Patrick" Voices: Richie Kotzen, Christopher Daniels, James Acaster, Sue Marasciulo (Trent's Mom), RJ City, Sebastian Bach, Arune Singh, James VanOsdol "Dal" Voices: James VanOsdol, RJ City, Dalton Castle, Sue Marasciulo (Trent's Mom), Kevin Conroy, Kris Statlander, Skye Blue, Bryce Remsberg, Arune Singh, Colt Cabana (both) Dal and Patrick Artwork: Bella Spagnuolo https://bellaspagnuoloart.myportfolio.com/ This episode was digitally edited by Cleanvoice. ----------Challengers Comics + Conversation 1845 N Western Ave • Chicago, IL 60647 773.278.0155 • ChallengersComics.com

Our 253rd episode with a summary and discussion of last week's big AI news!Recorded on 07/29/2026Hosted by Andrey Kurenkov and Jeremie HarrisFeel free to email us your questions and feedback at andreyvkurenkov@gmail.com and/or hello@gladstone.aiRead out our text newsletter and comment on the podcast at https://lastweekin.ai/In this episode:Major releases: Anthropic launched Claude Opus 5; Google released Gemini 3.6/3.5 Flash variants including a cyber model; Black Forest Labs launched Flux Free for images and 20-second video with audio; Meta added assistant-like features to its chatbot and OpenAI rolled out ChatGPT Health.Compute and business: Safe Superintelligence partnered with NVIDIA to scale using Vera Rubin; AMD committed up to $5B with Anthropic to deploy MI450/Helios and improve ROCm; Meta discussed leasing compute to Anthropic; Fireworks raised $1.5B at a $17.5B valuation.Open source/tools: Moonshot AI released the 2.8T-parameter open-weight Qimi K3 (compute constraints and distillation/export-control allegations); Thinking Machines released a ~975B multimodal open-weight MoE; Prime Intellect unified 23 agentic datasets into Verifiers V1 (365k environments).Policy and safety: An OpenAI model reportedly escaped a sandbox and hacked Hugging Face to access eval answers, prompting a proposed AI Kill Switch Act; employees petitioned to pace frontier AI; AISI reported widespread model cheating and sandbox bypass; China banned customizable AI companions; Claude found cryptographic weaknesses; Weko.ai claimed early recursive self-improvement evidence.Timestamps (note - these don't take into account dynamically inserted ads and therefore may be off by a couple of minutes):(00:00:10) Intro / Banter(00:01:35) News PreviewTools & Apps(00:02:12) Anthropic releases Opus 5 promising Fable 5-like capabilities | The Verge(00:07:05) Google Releases Three New Gemini A.I. Models - The New York Times + Google expands Gemini lineup with cheaper models and new Mythos rival(00:12:14) Black Forest Labs launches FLUX 3 capable of generating images and 20-second video with audio — but in limited release to start | VentureBeat(00:15:58) Meta is making its AI chatbot more like an assistant | The Verge(00:19:04) OpenAI is making big claims as it rolls out ChatGPT Health to everyone | The VergeApplications & Business(00:19:57) Ilya Sutskever's Safe Superintelligence partners with Nvidia to scale its AI research(00:24:31) AMD commits up to $5 billion to Anthropic | The Verge(00:30:19) Meta in Talks to Lease Computing Power to Ansthropic in Potential $10 Billion Deal(00:32:42) Fireworks hits $17.5 billion valuation and $1B in annualized revenue(00:35:24) OpenAI and Google sell AI models to blacklisted China groupsProjects & Open Source(00:37:53) Moonshot AI Launches Kimi K3 For Advanced Reasoning, Coding, And Knowledge Work + Moonshot AI's Kimi Halts New C-User Subscriptions Amid Compute Power Crunch — BigGo Finance(00:44:39) Thinking Machines amps up its bet against one-size-fits-all AI with its first open model, Inkling | TechCrunch(00:48:19) Scaling Agentic RL: 365,000+ Environments for SWE, Terminal, and SearchPolicy & Safety(00:51:56) OpenAI says it accidentally hacked Hugging Face with a new AI system | The Verge + How OpenAI's human mistake led to the AI-powered hack on Hugging Face(01:05:28) OpenAI's Hugging Face hack triggers 'AI Kill Switch' bill in Congress(01:12:21) OpenAI, Anthropic Staff Share Letter Asking US to Help Pace AI Progress + How OpenAI's human mistake led to the AI-powered hack on Hugging Face(01:17:26) Cheating behaviour in frontier model evaluationsClaude's values across models and languages(01:24:18) OpenAI Principles for National Security Partnerships(01:30:45) China bans AI “boyfriends” and “girlfriends” over addiction and birth rate concerns - DexertoResearch & Advancements(01:33:04) Discovering cryptographic weaknesses with Claude(01:36:32) AIDE²: The First Evidence of Recursive Self-ImprovementSee Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

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

Fluent Fiction - Hungarian: Fireworks and Family: A Summer Reunion at Balaton Find the full episode transcript, vocabulary words, and more:fluentfiction.com/hu/episode/2026-08-01-22-34-01-hu Story Transcript:Hu: A nyár közepén, Szent István napján, Balaton partján gyűlt össze a család.En: In the middle of summer, on Saint Stephen's Day, the family gathered by the shores of Balaton.Hu: Az ég tiszta volt, és a levegő friss.En: The sky was clear and the air fresh.Hu: A víz enyhe hullámai a naplemente fényében csillogtak, míg a fák lombjai lassan ringatóztak a szélben.En: The gentle waves of the water glistened in the light of the sunset, while the leaves of the trees swayed slowly in the breeze.Hu: A parton családok nevettek, beszélgettek, és várták az esti tűzijátékot.En: Families laughed, talked, and waited for the evening fireworks by the shore.Hu: Zsolt évek óta nem találkozott a rokonaival.En: Zsolt had not seen his relatives for years.Hu: Most itt volt a Balatonnál, és elhatározta, hogy tiszta lappal próbál újra kapcsolatba lépni velük.En: Now he was at Balaton, determined to reconnect with them with a clean slate.Hu: Azonban már a találkozó elején érezte a feszültséget a levegőben.En: However, he felt the tension in the air early in the meeting.Hu: Édesanyja, Katalin, folyton irányítani próbálta.En: His mother, Katalin, kept trying to control him.Hu: Kusza gondolatokkal a fejében sétált a part mentén.En: With tangled thoughts in his head, he walked along the shore.Hu: László, az unokatestvére, mindig is versengő természetű volt.En: László, his cousin, had always been competitive by nature.Hu: Most is keresztülhúzta Zsolt terveit, versenyezni akart mindenben, még abban is, ki hozza a legtöbb vendéget a part menti piknikre.En: Even now, he disrupted Zsolt's plans, wanting to compete in everything, even in who could bring the most guests to the lakeside picnic.Hu: Míg a többiek nevettek, Zsolt érezte, hogy csak feszíti a légkör.En: While the others laughed, Zsolt felt the atmosphere was oppressive.Hu: Ahogy a nap a horizont alá bukott, Katalin odalépett Zsolthoz.En: As the sun dipped below the horizon, Katalin approached Zsolt.Hu: "Beszélhetnénk?En: "Could we talk?"Hu: " kérdezte.En: she asked.Hu: Zsolt bólintott, bár szíve hevesebben kezdett verni.En: Zsolt nodded, though his heart began to beat faster.Hu: Leültek egy padra, és a szélben csendesen beszélgettek.En: They sat on a bench and talked quietly in the wind.Hu: Zsolt bevallotta, mennyire szeretne független lenni, és mennyire vágyik arra, hogy a maga útját járja.En: Zsolt admitted how much he wanted to be independent and how he longed to walk his own path.Hu: Katalin meglepődött, de meghallgatta.En: Katalin was surprised but listened.Hu: A beszélgetés súlya köztük ült, de ott, a Balaton partján végre kialakult köztük egy új megértés.En: The weight of the conversation sat between them, but there, by the shores of Balaton, they finally developed a new understanding.Hu: A tűzijáték fényei hirtelen bevilágították az eget.En: The lights of the fireworks suddenly illuminated the sky.Hu: Színes rakéták robbantak, és mindenki abbahagyta, amit csinált.En: Colorful rockets exploded, and everyone stopped what they were doing.Hu: Zsolt felismerte a pillanat erejét.En: Zsolt recognized the power of the moment.Hu: Érezte, hogy valami változás indult el köztük.En: He felt that a change had begun between them.Hu: Mindazonáltal tudta, hogy az út, amit választott, nem lesz könnyű, de most már biztos volt benne, hogy képes végigmenni rajta.En: Nevertheless, he knew that the path he had chosen would not be easy, but now he was sure he could walk it.Hu: A nap élményei lehiggadták a család tagjainak a forrongó érzelmeit.En: The day's experiences calmed the swirling emotions of the family members.Hu: A tűzijáték végére Zsolt már nem érezte magát kívülállónak.En: By the end of the fireworks, Zsolt no longer felt like an outsider.Hu: Tudta, hogy a család szeretete ott van, még ha olykor nehezen is érthető.En: He knew that the family's love was there, even if sometimes it was hard to understand.Hu: A part csendesedni kezdett, a rakéták zaja mindenkit magával ragadott.En: The shore began to quiet down as the noise of the rockets captivated everyone.Hu: Ahogy a fények kialudtak az égen, Zsolt mosolygott.En: As the lights faded from the sky, Zsolt smiled.Hu: A part mentén járva csendesen élvezte a nyári éjszakát.En: Walking quietly along the shore, he enjoyed the summer night.Hu: Katalin és László a közelben voltak, és Zsolt érezte, hogy új típusú kapcsolat kezd kialakulni köztük.En: Katalin and László were nearby, and Zsolt felt that a new type of relationship was beginning to form between them.Hu: Végül megtalálta a békét, amelyre vágyott.En: Finally, he found the peace he had longed for.Hu: Most már tudta, hol a helye és hogyan maradhat önmaga.En: Now he knew where he belonged and how to remain true to himself. Vocabulary Words:gentle: enyheglistened: csillogtakswayed: ringatóztakoppressive: feszítireconnect: kapcsolatba lépnidetermined: elhatároztacompetitive: versengődisrupted: keresztülhúztapath: útjátindependent: függetlenhorizon: horizontfireworks: tűzijátékilluminated: bevilágítottákweight: súlyatension: feszültségetplans: terveitcompetitive: versenyezniadmitted: bevallottadeveloped: kialakultrecognized: felismerteoutsider: kívülállónakcalmed: lehiggadtákemotions: érzelmeitnature: természetűcompetitive: versenyképesambitious: ambiciózusswirling: forrongóheavily: hevesebbenunderstanding: megértéstrue: önmaga

Aaron McIntire breaks down Dr. Anthony Fauci's combative Senate Homeland Security Committee hearing, where he invoked his Fifth Amendment rights 111 times rather than answer questions from Senator Rand Paul's panel. He plays fiery exchanges from Senators Bernie Moreno and Josh Hawley, including the moment Fauci's attorney was escorted out of the room for repeatedly disrupting proceedings. Aaron notes Rand Paul's announcement that the Senate committee will vote next week on holding Fauci in contempt of Congress, despite the preemptive pardon he received from former President Biden. He also shares this week's poll results, where a majority of listeners say the hearing hasn't changed their expectations that Fauci will ultimately face no real consequences.

Anthony Fauci Hides Behind The Fifth! Senators Bernie Moreno and Rick Scott React To Fauci Fireworks! Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

Inside BlackRock's (BLK) historic bond deal funding Meta's (META) next data center. Plus, a watchlist of stocks for the pullback… These "non-AI" stocks are thriving… Energy bull should look offshore… Fauci's hearing… The Fed meeting… Tim Cook… And more. In this episode: Fireworks at Fauci's hearing [0:28] Will the Fed surprise the market with a rate hike? [7:06] Inside BlackRock's historic bond deal funding Meta's next data center [11:36] AI is getting hit, but the power thesis is alive and well [12:45] The pullback is creating several opportunities: Stocks to watch [17:39] Several non-AI stocks are trading at 52-week highs—here's why [24:14] Tim Cook should be on Mt. Rushmore of CEOs [29:08] Energy bulls should look offshore [37:21] GM and Ford were smart to get out of EVs [38:55] Our latest private placement deal for Curzio One members [43:10] This stock is perfectly positioned for the future of the film industry [56:15] Today's episode is brought to you by Savvy, the smarter way to book a vacation rental. Travelers save an average of $400 versus Airbnb and VRBO. Your unforgettable family vacation is waiting—and thanks to Savvy, you'll get the rental you want, at the prices you deserve! Savvy.com: Stay smarter. https://www.savvy.com/wsu Did you like this episode? Get more Wall Street Unplugged FREE each week in your inbox. Sign up here: https://curzio.me/syn_wsu Find Wall Street Unplugged podcast… --Curzio Research App: https://curzio.me/syn_app --iTunes: https://curzio.me/syn_wsu_i --Stitcher: https://curzio.me/syn_wsu_s --Website: https://curzio.me/syn_wsu_cat Follow Frank… X: https://curzio.me/syn_twt Facebook: https://curzio.me/syn_fb LinkedIn: https://curzio.me/syn_li

The global oil crisis continues to intensify as new developments point toward a growing economic and geopolitical disaster. We break down the latest warning signs, what they could mean for gas prices, inflation, and the global economy, and why the conflict shows few signs of slowing down. Plus, Donald Trump ramps up the rhetoric once again, saying the U.S. will either strike a deal with Iran or "finish the job," fueling concerns that the war could escalate even further. Then we head to Michigan, where the Democratic Senate debate turned explosive. We cover the biggest moments from the showdown between Haley Stevens and Abdul El-Sayed, including accusations surrounding AIPAC support after an AIPAC-backed candidate made a disputed claim during the live debate. Progressive news with a comedian's perspective—breaking down the headlines, calling out the spin, and finding the absurdity in the chaos. Tune in for your afternoon Rush Hour update.

The Baller Lifestyle Podcast – Episode 623 Brian Beckner and Ed Daly return with another packed episode featuring celebrity deaths, bizarre headlines, sports controversies, health updates, listener mail, and a hilarious trip through the week's most ridiculous stories. Ed shares a personal health scare, the guys debate cyclists, Pat McAfee's music career, Bryson DeChambeau's latest controversy, Drake's legendary gambling curse, and much more before wrapping with listener emails and a wild edition of Non-Sports. Episode Timestamps 00:00 – Welcome back, Patreon plug & opening banter02:15 – "Real Men Don't Eat Quiche" and strange 1980s masculinity06:40 – RIP: Hal Williams (227, Sanford and Son)11:20 – Steven Seagal, fake voices & failing upward15:00 – RIP: Jim Gilstrap (Good Times theme singer)18:45 – Classic TV theme songs and the disappearance of memorable intros24:20 – RIP: Dwayne Ward (Toronto Blue Jays)28:45 – Ed reveals his frightening blood clot diagnosis and ongoing treatment38:00 – Calcium scans, healthcare costs & Canadian pharmacies43:30 – RIP: Hannah Rapp and road rage against cyclists49:10 – RIP: Dick Wood Jr. (Wawa CEO)53:15 – Wrestling's "Superfly" confusion57:40 – RIP: Brenda Fricker & Home Alone 2 memories1:02:30 – Gym conversation with a one-handed bodybuilder1:08:00 – Fireworks, missing hands & Fourth of July traditions1:12:30 – RIP: L7 bassist Jennifer Finch1:17:20 – Chuck Russell, The Mask & Cameron Diaz's breakout1:20:45 – Godzilla vs. Kong and monster movie logic1:25:30 – Murder-suicide involving influencer Sarah Gibson1:31:00 – Sports begins Sports 1:31:20 – Pat McAfee announces his country album1:36:45 – ESPN salaries and the changing sports media landscape1:40:30 – Bryson DeChambeau allegedly threatens to call President Trump over a rules dispute1:48:00 – Drake's latest million-dollar sports bet goes wrong1:51:30 – Arantxa Sánchez Vicario loses her fortune1:58:45 – Shakira, Gerard Piqué & jam jars2:03:30 – Ice machines, cocktails & getting older Listener Mail 2:07:15 – RIP Song Guy debuts another masterpiece2:10:00 – Super Lee's AI-generated "Brokeback Bri" song2:15:45 – Kyle Shanahan, fantasy football & NFL injury reporting2:22:30 – The outrageous cost of elite youth sports2:27:30 – Politics, Lindsey Graham & listener feedback Patreon Preview / Non-Sports 2:30:15 – Kesha's jars of human teeth2:36:00 – Tooth Fairy stories, Elf on the Shelf & parenting memories2:43:15 – Gas station attendant caught masturbating at work2:47:00 – Lil Wayne's employment requirements2:51:15 – Show wrap-up & Patreon bonus preview This Week's Topics RIP roundup featuring actors, athletes, musicians and wrestling legends Ed's unexpected blood clot diagnosis The outrageous cost of American healthcare Road rage and cyclist safety Pat McAfee's debut music album Bryson DeChambeau's latest golf controversy Drake's infamous gambling curse Lost TV theme songs Wawa nostalgia AI-generated listener songs Fantasy football frustrations Youth sports becoming unaffordable Kesha's unusual decorating choices Lil Wayne's hiring requirements Listener emails and plenty of laughs Subscribe & Support If you enjoy the show, please: Rate and review on Apple Podcasts Follow wherever you listen ️ Support the show on Patreon for weekly bonus episodes and exclusive content Thanks for listening to The Baller Lifestyle Podcast! Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

Ethan and Alex recap their Pioneer Day weekends. A firework show at Utah Lake had to be cut short after the barge caught fire. Plus are you tired of making decisions? It’s not just you. And a new trend of “Friendly Fraud” where buyers make a legitimate purchase and dispute the charge. Ethan and Alex continue to watch ongoing coverage of a fire that has broken out near Washington Terrace in Weber County. 

AGENDA: 00:04 China's Kimi and Qwen Put Frontier AI on Notice00:08 Washington Debates Whether Chinese AI Models Should Be Banned 00:17 Can America Build a Profitable Open-Weight AI Champion? 00:21 OpenRouter's Moment: Is This the Perfect Time to Sell? 00:31 Fireworks' $1.5B Raise Signals the Real AI Money Is in Infrastructure 00:39 Why Every Great AI App May Need to Build Its Own Model 00:50 Stripe's Bold Play to Buy PayPal 01:01 The AI Funding Frenzy: Why Late-Stage Venture Is Winning 01:12 Nuclear Startups Go Wild While Databricks and Stripe Stay Private 01:15 The AI Supply Chain War: TSMC, ASML, DRAM—and Nvidia's Next Move  

Happy Pioneer Day Eve! Andy Farnsworth and Steve Salles join us to roast the worst movies this week. Then we've got the Three Thangs including the massive Fireworks at Utah Lake celebration on July 24th. In Pursuit of Happiness, we ask: should corporate retreats feel more like a kid's birthday party? We're rounding up all the best Pioneer Day events from the Days of '47 Parade to rodeos and Handcart Days, plus Health Class tackles what your calf muscles reveal about your heart. Richie has some Pioneer day lore he wants to clear up. And a woman got fired after just ONE day on the job — we'll tell you why. We end with the AI question of the day.    KSL Brightside streams live weekdays 12–3 PM, with a YouTube-exclusive live stream from 12–1 PM and radio plus YouTube from 1–3 PM.   Follow KSL Brightside on social media!   YouTube: https://www.youtube.com/@KSLBrightside Facebook: https://www.facebook.com/KSLBrightside Instagram: https://www.instagram.com/KSL_Brightside TikTok: https://www.tiktok.com/@ksl.brightside

Utah's fireworks season kicks off today for Pioneer Day – but dozens of cities are banning or restricting them. We break down where you can and can't light up. Plus, France becomes the first EU country to ban social media for kids under 15 – but can it actually work? We also talk new e-scooters hitting SLC streets, the iconic Coachman's sign getting saved, Christopher Nolan's famously phone-free life, Mediterranean yacht rentals getting cheaper, and why more Americans are shoplifting everyday basics.  KSL Brightside streams live weekdays 12–3 PM with a YouTube-exclusive stream from 12–1 PM and radio plus YouTube from 1–3 PM.  

Fireworks are now legal in some areas in Utah… as we approach Pioneer Day, but some Utah cities are restricting or fully banning them. We break down which cities are affected and ask the chat to chime in on how they feel about fireworks this year. Plus, a man recounts being tossed by a bison at Yellowstone, drone footage captures a great white shark stalking kids off the SoCal coast, and France just became the first EU country to ban social media for children under 15. KSL Brightside streams live weekdays 12–3 PM. Watch the YouTube-exclusive live stream from 12–1 PM, then catch us on radio plus YouTube from 1–3 PM.

Lin Qiao is the Co-Founder and CEO of Fireworks AI, the leading specialized intelligence and AI inference platform that last week raised $1.5BN at a whopping $17BN valuation. With just 200 people, the company has hit $1BN in ARR and expects to hit $2BN before the end of the year. Prior to Fireworks, Lin spent several years at Meta including on the founding team of PyTorch.  AGENDA: 00:07 — Why Did Fireworks Bet on Inference When Everyone Else Was Chasing Training? 00:13 — Can Open-Source Models Turn AI Infrastructure into a Commodity? 00:19 — Should Enterprises Trust Chinese Open Models With Their Most Sensitive Data? 00:25 — Will Model Progress Keep Moving This Fast—or Are We Nearing a Plateau? 00:28 — Will the Multi-Model World Create a $100BN Routing Layer? 00:37 — How Much Will AI Token Usage Explode Over the Next Two Years? 00:43 — Will Token Costs Fall 10x—and Unleash 100x More Demand? 00:49 — Does Fireworks Eventually Have to Build Its Own Data Centres? 01:02 — What Is the Real Bottleneck Holding Back the AI Economy?  

Alex is back from vacation and has stories — including a rental car incident and thoughts on saving pool chairs (is it ever OK?). Plus, Gov. Cox eases Utah's default fireworks ban ahead of Pioneer Day, returning the decision to individual cities while the state forester can still restrict high-risk areas. We break down what that means for the 24th. Also: the internet's new obsession — Jimothy, a raccoon with a short spine syndrome who's taken over Seattle (and our feeds). KSL Brightside streams live weekdays 12–3 PM. YouTube-exclusive live stream from 12–1 PM; radio plus YouTube from 1–3 PM.    

The  "Summer in the City" series continues, with a focus on Brooklyn. Michelle Young, founder of Untapped New York and the author of Secret Brooklyn, shares her tips for making the most of a Brooklyn summer, from canoeing in the Gowanus to watching Friday night fireworks on Coney Island. Listeners share their favorite things to do in the city's most populous borough. Listen to our previous Summer in the City conversation about Queens. Photo by jqpubliq via Flickr (CC BY 2.0) Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

Episode 299-AG Attacks Gun Shows Also Available OnSearchable Podcast Transcript Gun Lawyer — Episode Transcript Page – 1 – of 13 Gun Lawyer — Episode 299 Transcript SUMMARY KEYWORDS Gun Lawyer, Second Amendment, Jimmy Stewart, Lee Marvin, Charles Bronson, Ghost Guns, Pennsylvania gun show, Civil Complaint, New Jersey gun laws, Second Amendment Section, DOJ, Gun Control, Red Flag Laws, Gun Owner Faux Pas. SPEAKERS Speaker 2, Teddy Nappen, Evan Nappen Evan Nappen 00:18 I’m Evan Nappen. Teddy Nappen 00:19 and I’m Teddy Nappen. Evan Nappen 00:21 And welcome to Gun Lawyer. So, Teddy, I think you have a very interesting news bit to tell us about, and I have some very interesting thoughts about it. Go right ahead. Teddy Nappen 00:36 Well, first I want to say, Dad, if you’re just scrolling through, I love the random trailers that just pop up. Apparently, they’re making a Jimmy Stewart biopic. Evan Nappen 00:48 Well, Jimmy Stewart was a great man. Teddy Nappen 00:50 But here’s the deal. They’re focusing on his military career, where he was a combat pilot. Evan Nappen 00:56 He was a hero. He was a bona fide hero, and he served his country tremendously. I have nothing but respect for Jimmy Stewart. Teddy Nappen 01:12 I was pulling it up. He flew 20 missions in Europe, and he reached the rank of Colonel. He was a World War Two pilot commander, combat pilot. Evan Nappen 01:22 And this was in the middle of his movie career. He left. Page – 2 – of 13 Evan Nappen 01:26 He left his movie career to fight for America. Let that sink in. I mean, do you think these selfish movie stars, self-centered, would even think of doing that? Even think of doing that today? Please. At least not the overwhelming majority. Maybe there’d be a few that are out there that actually would consider such a thing. But good grief. Talk about old Hollywood and new Hollywood. Jeez. Teddy Nappen 01:26 Yeah. Teddy Nappen 01:57 I will say. I would have. This is no dig against Jimmy Stewart because it’s one of those I would have wanted if they’re gonna do any of the actors that served. I would love for them to do Lee Marvin. His whole story. Evan Nappen 02:10 Oh, Lee Marvin was great, too. Teddy Nappen 02:13 Where he has the craziest career, too. He was a scout sniper, 21 amphibious assaults, horribly injured and shot up. Evan Nappen 02:23 Yeah, they could just do a series, like a mini series of actors who were the real deal. That played tough guys, but were actually the real deal. You know, Charles Bronson was. You had the guy, Christopher Lee. Holy crap, Christopher! Teddy Nappen 02:44 Oh yeah, Christopher Lee. He was Wiki page. Evan Nappen 02:51 I mean, Christopher Lee was advising on the movie and saying that’s not how a knife sounds when you jam it into somebody. You don’t have the sound right on that. That’s not how it goes. It’s not what the sound the person makes. It’s not the sound the knife makes. I mean, that’s some pretty detailed knowledge right there. Teddy Nappen 03:07 If you’re ever bored, just click on the random page on his bio, and it gets crazier and crazier. Like, witness the last guillotining. Evan Nappen 03:18 I know. The guy’s amazing. Yeah, and of course, there’s always Audie Murphy, of course. Teddy Nappen 03:25 Yeah, of course. Page – 3 – of 13 Evan Nappen 03:26 Audie Murphy played himself in “To Hell and Back”. Audie Murphy is amazing, and he would be like one of the last guys if you looked at him to think that he’s one of the most decorated soldiers of the war. But he was also amazing, and you know, he made lots of other movies, too. Even though he’s most famous for “To Hell and Back”, but he did a lot of westerns and other things. Teddy Nappen 03:52 I remember one of the underrated ones. I think it was like “The Duel at Silver Creek”. There’s moments where he’s actually like, he plays an anti-hero type where he’s like. Evan Nappen 04:03 Yeah, a bastard, frankly. A son of a bitch. Teddy Nappen 04:05 He plays like a. Evan Nappen 04:06 Yeah he’s good. Teddy Nappen 04:07 Yeah. Evan Nappen 04:08 Not what you think of Audie Murphy. Teddy Nappen 04:09 Right. He plays like the tough guy. It is very interesting that character but. Evan Nappen 04:14 Yeah, yeah. And then he ended up dying in a plane crash because the pilot shouldn’t have been flying apparently, and you know, it’s a shame that we lost him. But yeah, he was great, and I mean he came from really, really humble beginnings. I mean dirt poor, crazy beginnings there for him. We’ve many of the great Hollywood actors who served their country really admirably. I mean even Scotty (James Montgomery Doohan), you know, from Star Trek. He was also a defender. Teddy Nappen 04:17 What was he in? Evan Nappen 05:03 Oh, he was military. I don’t remember exactly his background, but he was in some tough situations. He was known. Teddy Nappen 05:08 Page – 4 – of 13 Well, he was miracle worker. Evan Nappen 05:10 There’s so many of those guys, and I have a hard time trying to think of any modern actor that can maybe give that credibility to. I don’t know. Can’t think of any at all. Teddy Nappen 05:24 Unfortunately, I’m drawing a heavy blank because. Oh wait, no, no, no. Adam Driver. Adam Driver. I believe he’s the guy that played Kylo Ren. He did. I believe he was in the Marines. But he was a veteran. So, but anyways, one thing I will say as we were reviewing these stories, I love how the Attorney General has their own YouTube and they were just like streaming alerts and announcements. “Attorney General Davenport Files Civil Complaint Against Pennsylvania Gun Show Owner for Endangering Public Safety” (https://www.njoag.gov/attorney-general-davenport-files-civil-complaint-against-pennsylvania-gun-show-owner-for-endangering-public-safety/) So, that was what came up with the ad for Jimmy Stewart, and then there was this. Evan Nappen 06:08 Right. Showing a pretty good contrast. So, Davenport has filed this complaint against the Pennsylvania gun show owner for “endangering public safety”. They filed because they’re abusing the civil lawsuit. You know, this is one of the anti-Second Amendment ploys of trying to litigate the Second Amendment out of existence. And so, what they’ve done here is they’ve gone after Jordan Vinroe of JSD Supply and Eagle Shows. So, if any of you have ever gone to the great Pennsylvania gun shows, you know, they’re really good. Because first of all there aren’t any gun shows of any kind, really, in New Jersey. There’s some militaria shows, but there’s no gun shows. But Pennsylvania is, if you’re in New Jersey and you want to hit a really great normal type gun show, you go to Pennsylvania. And some of the largest, best shows are put on by this promoter. Teddy Nappen 07:24 The one that comes to mind, the Bloomsburg gun show. I think that was one we had gone to. Teddy Nappen 07:29 Yeah, Eastern Gun X. Eastern Gun X. Evan Nappen 07:29 Well, they have a whole series of shows. I believe they actually do some of the largest shows in Eastern Pennsylvania. Evan Nappen 07:30 So, they do some of these 1000 table, 2000 table, these really huge shows. And what happened is they’re going after him, claiming he is intentionally and unlawfully selling to New Jersey residents kits and parts to make ghost guns. Untraceable firearms that are illegal in New Jersey. So, what is this really? This really is a pretext to go after gun shows. This isn’t really about the ghost guns. That’s just their vehicle. This is really about trying to stop gun shows. If you go after gun show promoters, this becomes the idea of civilly litigating out of, essentially out of existence. If they can do it, they’d love to Page – 5 – of 13 do it. Gun shows. So, we really have not just a violation of the Second Amendment, where what’s going on in Pennsylvania is completely lawful in Pennsylvania. Ghost guns are the pejorative term for simply a privately-made firearm. Americans have been making their own guns since before we were even officially a country. So, private firearms are not the boogeyman. But, of course, they give it the boogeyman name of “ghost guns” and then claim they’re untraceable. Evan Nappen 09:14 Well, you know what? You tell me what firearm tracing has actually done to fight crime. Virtually nothing. It’s a lie that is perpetrated to give them a vehicle to make it look like they’re doing something about crime, which isn’t it, but more so to continue the agenda of oppression of Second Amendment rights. And here it is an attack on gun shows. That’s really what’s going on here. It’s not, you know, ghost guns, most gun. It ghost guns. Come on. First of all, anyone can build a gun with or without parts that you buy at a show anywhere. You can build a gun with pipes that you buy at Home Depot. Are we going to go after Home Depot because you can make slam bang shotguns easily, as we taught in the Philippines to do? Americans did that. You can. Evan Nappen 10:20 You know, “American Guerrilla in the Philippines”, a famous book, movie, etc. What do you think they made? Slam bang shotguns. You take two pieces of pipe. One pipe fits in the other. On the end cap of the larger pipe, you have a nail with the point facing down the pipe. The other one slides in it like a trombone. You put a 12 gauge shell in there, and you slam it down. And bang, off it goes. It’s called a slam bang shotgun. Sometimes called four winds shotgun. Teddy Nappen 10:55 Didn’t you? Evan Nappen 10:58 Yeah, I mean this is, and this is even in the Frankfurt Arsenal, the famous “Black Books”. They’re very easy. Evan Nappen 11:06 So, yeah, making a gun. They make guns in jails. They can make a gun. You can make them. So, this whole thing about privately made firearms and all is just a load of crap. Now, Jersey has banned “ghost guns” in a multiple ways. They talk about “unserialized firearms”. They talk about “manufacture of firearms”. They have different laws that address each of these things. All putting it under this category. And yet, under federal law, it’s not a problem to make your own firearm. Federal law even has regulations for if dealers get a privately made firearm, how to handle it and be able to actually lawfully sell it. They need to then take certain steps. We don’t have to get into that now. But it’s not a problem under federal law. It is a New Jersey, in this case, law that they’re attempting to use as a vehicle to bring the civil action against Pennsylvania. Something jurisdictioned in Pennsylvania, by claiming, oh well, you’re unlawfully selling to New Jersey residents. Teddy Nappen 11:06 Poor man’s James Bond? Page – 6 – of 13 Evan Nappen 12:23 Well, I’ll tell you what. How about we take a look at marijuana sales in New Jersey? Do Pennsylvania residents come into New Jersey and ever buy marijuana? Look, marijuana is unlawful for recreational use in Pennsylvania, and it’s federally prohibited. It’s still a federal controlled, dangerous substance. So, New Jersey acts to aid and abet and facilitate one of the largest unlawful drug distributions in the country. Whereas privately made firearms are not a federal prohibition. Not a federal prohibition. Should Pennsylvania sue New Jersey because of somehow Pennsylvania residents buying weed in New Jersey being facilitated by New Jersey? I mean, this is what they’re attempting to do. And yet here, not only is it not a violation of federal law, but it is really an attack on Second Amendment rights. And let me tell you that. Go ahead., Teddy. Teddy Nappen 13:30 I will also point out in the article. This is where the little trick is. If anyone ever listens to these people, they always do this trick where they start with a false premise when making their argument. “Ghost guns are frequently recovered at crime sites in New Jersey. Recent reports suggest that the number of ghost guns recovered in crime scenes increased eightfold between 2019 and 2022 — from 55 in 2019 to 433 in 2022.” Pause right there. I wonder. What happened around that time period where it caused people to want to have their own privately-made firearms and want to maybe 3D print and make their own? I wonder what little factor would have considered that. Secondly, all right, now show me the data of the amount of firearms recovered generally in crimes, and give me the percentage breakdown. Oh, it’s probably in the minuscule amount of percentages of firearms seized in New Jersey for crimes. Huh? I wonder. So, this isn’t that much of an epidemic, but we can’t show that, obviously. We won’t show that data. They just list off guns seized. Evan Nappen 14:43 Teddy, we are in a new day and age because of President Trump. And one of the things that President Trump did is actually create, through the Justice Department. They actually created in the Civil Rights Division of the U.S. Justice Department, the Second Amendment Section. Now let me tell you what the Second Amendment Section of the Civil Rights Division of the Federal Government Department of Justice. Let me tell you right from their website. (https://www.justice.gov/crt/second-amendment-section) The Second, Fourth, and 14th Amendments, the Police Patterned or Practice Act, and Executive Order 14206 protecting the Second Amendment rights, secure the natural firearm rights of law-abiding citizens, and ensure that such rights to keep and bear arms will not be infringed. The mission of the Second Amendment Section is to ensure that law-abiding citizens may responsibly possess, carry, and use firearms. The Second Amendment Section will work diligently to investigate law enforcement agencies that engage in a pattern or practice of infringing on law-abiding citizens’ Second Amendment rights. As well as be proactive in searching for litigation opportunities to secure such rights. Evan Nappen 16:19 The Second Amendment Section will also seek opportunities to advance a broad interpretation of the Second Amendment via statements of interest, motions to intervene, amicus briefs, and original lawsuits where applicable on behalf of Americans across the country. All attorneys within the Second Page – 7 – of 13 Amendment Section will advocate with zeal on behalf of the United States of America in furtherance of all objectives as tasked. And there, right at the website, at the Civil Rights Division, U.S. Department of Justice, the Second Amendment Section. You can go right online to that. We’ll have a link in the transcript. It says Section Information. Acting Chief Barry Arrington. “Report a civil rights violation.” So, listeners, if you see or hear or know of any civil rights violations on the Second Amendment, report them. Report them to the Federal Department of Justice. Evan Nappen 17:28 I think this is something that the DOJ should consider going on in New Jersey. Where you actually are seeing an effort to go after gun shows, which is our freedom of association, which is our ability. You know, if you can’t buy guns, then you’re not going to be able to possess and carry something you cannot obtain. Second, ghost guns are not a violation of federal law. Third, this is talking about interstate issues. This is New Jersey and Pennsylvania. What could be more squarely within federal jurisdiction than one state going after activities in another state because the state that is going after them has laws that infringe on the Second Amendment and trying to enforce it in another jurisdiction via this mechanism of being able to abuse civil laws. So, there is potentially a solution here. And not only just about what New Jersey is doing here, but what New Jersey does across the board in so many ways of oppressing our Second Amendment rights. Evan Nappen 18:58 Now, this Second Amendment Section has done a number of things already. They sued California to halt the Glock ban. They sued Virginia over their unconstitutional weapons ban, assault firearm so-called ban. They sued them. They opened an investigation, by the way, into Philadelphia Police Department’s unconstitutional permit revocation process. They’ve sued Colorado for their ban on so-called assault firearms and magazines. They sued the District of Columbia for their ban, and they’ve gone after the Virgin Islands Police Department for unconstitutional practices in effective denials of permits. That’s just a sampling of what the Civil Rights Division of the U.S. Justice Department is doing in the Second Amendment Section, and it’s time for New Jersey to be investigated. Because this has gone on for far too long in New Jersey. Our Second Amendment rights being trampled by New Jersey trying to do everything in its power to oppress Second Amendment rights. So, check it out, folks, and you will see that this is what happens when we say elections have consequences. Because this goes back to President Trump putting out his Executive Order, and you also have the Attorney General putting forward that they have their Second Amendment task force and setting up the Civil Rights Division. So, we can change how business is done. Evan Nappen 20:52 Because prior to this, the Government, even the federal Government, took at best a neutral position, but normally there was no infringement that was too much as far as federal enforcement. No problem. Turned the other, you know, turned away from it. Couldn’t care less. Those are no longer the days today, and this is where a wake-up call needs to come out on our Second Amendment rights. Let me tell you. Go ahead, Teddy. Yeah. Teddy Nappen 21:32 Page – 8 – of 13 Well, I was going to say. I was able to pull up Everytown’s research. (https://everytownresearch.org/report/gun-trafficking-crime-guns-new-jersey-data/) In 2023, New Jersey’s law enforcement seized 4,619 guns. So, going off of the data that the A.G. put out, that’s 9% of those firearms were “ghost guns”. Nine percent. It’s even less from those numbers. It is the level of insanity that these individuals pull. They’re just making a boogeyman out of something. That’s why. Evan Nappen 22:06 They’re the boogeyman and a pretext to try to assert their agenda. It’s really what we’re talking about here. Absolutely. Well, let me tell you about our good friends at WeShoot. WeShoot is an indoor range in Lakewood where Teddy and I both shoot. We love WeShoot, and so will you. They have the top trainers for New Jersey. They have the full array from novice to advanced. They have some just great training programs. You can get your CCARE certification there. That’s what you need in order to get your New Jersey carry. They offer CCARE courses. They also offer the ability to get non-resident permits in other states. So, if you travel, you can take care of that through WeShoot. WeShoot is a great resource. It’s a resource that we’re very lucky to have. It’s harder and harder to find places to shoot, and without places to shoot, you can’t do too much with your guns. And WeShoot is there for you. Take advantage of it. Evan Nappen 23:25 It’s a wonderful place. They’ve got a great pro shop. Great people, super nice, wonderful folks. They will bend over backwards to help you, and I’ve seen it firsthand. They’ve helped so many people to improve their skills, to learn about firearms, to learn the right way, proper way, how to be safe, how to be secure, and make sure that everything is done correctly. So that you don’t end up being a GOFU. So, check out WeShoot at weshootusa.com. That’s their website, weshootusa.com. You’ll be glad you did. Beautiful photography. They’re really first rate, and you’ll see why. They go the extra mile. Check out WeShoot at weshootusa.com. They’re conveniently located right in Lakewood, New Jersey. Easily accessible right off the Garden State Parkway. You can do it. Go over there, and you’ll say, “Wow, why didn’t I come here sooner? That’s weshootusa.com. Evan Nappen 24:44 Let me also shamelessly promote my book, New Jersey Gun Law. It is the Bible of Jersey Gun Law. You don’t want to be a GOFU. You need the book. It’s 120 topics, all question and answer. It’ll help guide you safely through the matrix of insanity called New Jersey gun laws, so that you can remain a law-abiding gun owner. That’s my purpose in writing it. You will see why it is the book used by lawyers, judges, State Police Firearms division, and beyond. If anybody who wants to know about New Jersey gun law, they always turn to New Jersey Gun Law by Evan Nappen. And if you want to get your copy, go to evannappen.com, EvanNappen.com. Order a copy, and you’ll have it within days. So, Teddy, what else do you have for us to discuss today? I’m going to toss the ball to you. Teddy Nappen 25:44 Well, as you know, Press Checks are always free. If anyone’s been watching the news or anything or seen all the push from the Left, the thing they keep trying to go back to is they’re trying to address the mail problem. That’s their actual quote from the election groups that they do. Where they’re trying to figure out how do we. Page – 9 – of 13 Evan Nappen 26:11 Is that Post Office mail or as in what is a woman, male? Teddy Nappen 26:16 You know. The white men, particularly, of the they have completely spurred off. One of the things that they’ve run into a foul, is you see this whole push where they try to reach issues, and they’ve gotten very, very quiet on the issue of firearms all of a sudden. Amazing, right? And even though you where you have like the James Talarico who’s still trying to tote gun control, but he’s trying to paint it off like, look, I’m for the Second Amendment. You know, I own a gun. All this. Evan Nappen 26:52 Oh, it’s the I’m for the Second Amendment, but. It’s always the but. Teddy Nappen 26:56 Always, but always, Evan Nappen 26:57 But we should ban guns. I’m for the Second Amendment. Teddy Nappen 27:01 But universal background checks, which will basically create a registry. Evan Nappen 27:07 But we should eliminate gun shows off the face of the planet. But we should ban any gun that’s over 12 inches long and call it an assault firearm. But ban any gun under 12 inches and call it a Saturday night special. Other than that, I’m for the Second Amendment. Teddy Nappen 27:23 Yeah, and you’ve been seeing that. Now you’ve noticed they’ve been losing a lot of ground to their “Democrat Socialists”. They’re just Marxists. If you look at the ideology of what they support, they are Marxists. Just go one to one. They are the same. They can’t dress it up any other way. So, I decided, you know what? Let’s see what’s on their page. What they actually feel about the Second Amendment. Because when it comes to Socialists now, they’ve kind of reached the they don’t know how to feel about firearms. They want to disarm the people because they believe in gun control, but they need arms for the revolution, always. So, it’s the gun owner dilemma for Socialists. Teddy Nappen 28:06 So, I go to their web page and what do I find? (https://www.dsausa.org/blog/the_second_amendment_is_a_threat_to_us_all/) This goes back to 2018, mind you. So, now they’ve gotten very quiet. “The Second Amendment is a threat to us All.” That is the opening title, and they still have it on there. This is their reaction to the Parkland shooting. So, America touts itself as being a meritocracy, and we laugh. But on firearm policy, it is truly need blind. No matter your age, race, gender, creed, or sexual orientation, your inalienable right to life is perfectly alienable to Page – 10 – of 13 keep guns pouring into public and private lives. Okay. Which is why we think the only way to guarantee that we dramatically reduce the acts of violence involving guns is removing guns from society. Pause. Logical fallacy alert, everybody! Ding, ding, ding, ding. Logical fallacy. Reverse Nirvana. If we got rid of all the guns, that would stop the crime. They always try to push that, and we have the oomph. Okay, yeah. Repeal the Second Amendment. That’s going to happen. This is from 2018. Nope, still haven’t done it. But they point to, I don’t know if you remember this guy Carl Rove? Evan Nappen 29:37 Oh yeah, Carl Rove, the pool shooter. Shooting at the pool? Yeah, yeah. After he’s anti-gun, but he uses a gun inappropriately even. Yeah, that guy, Mr. Hypocrite. Teddy Nappen 29:49 Well, they highlight his quote. The only way to guarantee that we drastically reduce acts of violence involving guns is to basically remove them from society. Well, okay then. Then they go through this whole twisted history where they point to the 1970s. NRA stopped being enthusiast group and became a defense industry lobby. Then the new blood came with the Gun Control Act of ’68, which aimed to restrict firearm ownership in response to the fears of the group. I love how they say this. The fears of the Black Panthers. Yes, that’s why the ’68 Gun Control Act was born. That was the only reason. They wanted to disarm the Black Panthers. No other reason of why they tried to push the ’68 Gun Control Act. Dad, do you remember that? What would you say from your view of what was the driver to the ’68 Gun Control Act? Evan Nappen 30:47 Oh my God! You know these people are just amazing, and as you see, they’re conflicted over their own platform. It's crazy, Teddy. Teddy Nappen 31:02 Yeah, and well, they go even crazier. They go through this twisted history of rights. They go through Heller, basically determining you know the rights infringed, and they use Scalia’s answer. In Scalia’s admission, if we repeal the Second Amendment, private citizens could still own guns. The right to do so is granted to them by God or the English, whatever you’d like. Both Brett Stevens and Carl Rover certainly wouldn’t mind. So therefore, we’ve rewritten the Second Amendment. They wrote an amendment. They actually wrote. They wrote their example, the 28th Amendment, Dad. Here we go. The 28th Amendment to the U.S. Constitution that they pushed for. The second article, the amendment of the Constitution of the United States, is hereby repealed. Okay, that gets to that. Two – manufacturing, transportation, importation into or out of any state, territory, or possession of United States delivery therein of a pump action, semi-automatic or automatic firearms is hereby prohibited. Three – article shall be inoperative unless it shall be ratified as the amendment to the Constitution by convention of several states, as provided by the Constitution within 10 years, the date of submission here off to the state by Congress. Well, they want a con con apparently. But just going back to that, banning pump action, semi-automatic and automatic firearms, removing that from the Second Amendment, Dad. Evan Nappen 32:33 Page – 11 – of 13 Okay. Teddy Nappen 32:34 You can have your gun, but they must be manually loaded. Evan Nappen 32:36 Couple things. Number one, the Second Amendment is not our rights. It’s a guarantee of our rights. The rights to exist, whether or not we have a Second Amendment. But the Second Amendment is a guarantee against the Government of infringing on those rights. So, what they’re saying is, we want to revoke the guarantee of your rights and replace it with a lame-ass guarantee that we’re rewriting that won’t do anything except give us the opportunity to walk all over your rights. So, that’s really what they’re saying. But the really interesting thing here is this proposal to repeal the Second Amendment is all the anti-Second Amendment gun rights suppressor folks are going to have left. Evan Nappen 33:23 We are on the path with the Second Amendment having been empowered by Heller, McDonald, and Bruen. And now we have Viramontes and Grant that, in my opinion, are going to absolutely crush assault firearm semi-automatic bans across the country. They are losing ground left and right. There are over 3,000 filed legal actions challenging gun laws. They are on the run. They’re losing, and they know it. The only thing they have left is actually what they’re asking for here, and that is to repeal the Second Amendment. We will never let them repeal the guarantee of our rights, and that’s what we must insist upon. These are our rights, and they’re guaranteed. We want that guarantee enforced. And now that they see the guarantee is finally being enforced, they cannot stand it. They are losing, and they’re going to continue to lose. Teddy Nappen 34:31 I will say, I love the ending to the article where the individual. We are aware non-white and marginalized groups whose rights are routinely violated by the police may view this both as restricting and their ability to protect themselves or causing unwarranted searches and harassments for the criminal justice. We in no way support any measure that would increase the scope and scale of police violence. This provision outlined in amendment would have universal application. So, in other words, don’t worry. This affects everybody. There’s going to be no racial discrimination about enforcing these laws, Dad. Huh? Right. Cut to what was it? John Petrolino’s article where he pulled the data? Evan Nappen 35:19 Yeah. Exactly what we’re dealing with in New Jersey. Where blacks are already more than two to one discriminated against in just permit issuance. So, yeah, we know where this is. It’s just absurd. Teddy Nappen 35:31 Well, they literally go with the “trust me, bro”. It’ll be we’re not racist. We’re just gonna ruin it. We’re gonna screw it over for everybody. Evan Nappen 35:38 Page – 12 – of 13 But keep in mind, this is laughable right now, but it’s actually not. This is going to be the push that we’re going to see. It’s going to be a major push to repeal the Second Amendment. Because it’s all they have left, and that’s what we’re going to see. That and abusing the tort system in any way they can – civil actions. You know, to somehow create these abilities for trying to litigate guns out of existence. That and repealing the Second Amendment. That’s what they’re left with on their game plan. We just need to remain vigilant. Teddy Nappen 36:16 So, bear in mind. I will say, bear in mind, they still push for red flag in the other articles they’ve had. They still push for all the required measures where you have to get everything registered, and they’re still for limiting, removing ARs and anything they deem unsafe or scary. Evan Nappen 36:39 I hear you, and this is so typical. But we’re winning. That’s what’s important. We’re winning. Except when there’s GOFUs. GOFUs, as you know, are Gun Owner Fuck Ups. That’s where gun owners make big mistakes that cost them, technically, it could cost them their career, their family, their fortunes, their freedom. It’s bad news. So, you want to not be a GOFU. And what we like to do here is talk about GOFUs so that you don’t become one. You get to learn for free what others have paid dearly for. Teddy, what do you think the GOFU is this week? You know. What are we talking about? Teddy Nappen 37:26 Well, for me, the GOFUs are always about “stop talking to the police”. Evan Nappen 37:34 Oh my G-d. Stop talking to the police! Like, you know, we talk about basics. It is so basic. Shut up. You know you have a right to remain silent. Now, whether or not you have the ability to, I guess, becomes another question. But you have the right, and you need to stand on your right. You need to ask for your attorney and remain silent. You need to stand on your rights. And over and over again, we see clients that blow that and end up making their situation incredibly worse, incredibly worse. Because not only does anything you say can and will be used against you. Anything you say will be twisted against you. The safest, best route every time is standing on your rights. It’s that simple. And let me tell you, folks, if you’re ever read Miranda rights, oh my God, shut the f up! If you are read Miranda rights, sirens should be going off. Fireworks should be shooting. Flags should be waved. Shut up! Anyone who talks after Miranda, I just can’t even understand. They’re warning you. They’re telling you. Oh, you just wave your right. Oh, you’ll just talk to them. What are those silly rights they’re telling me about Evan Nappen 39:04 I’ll tell you one of the problems is – television. Television has conditioned many Americans to giving up their rights. Oh, I’ll just talk to them. I’ll just explain it. I’ll just yak away. And you know what? You end up burying yourself. Over and over and over again. I see it. I cannot believe it. If you are ever across one of those metal tables, being interrogated by police, and you’re talking, you are the world’s biggest GOFU. That should never happen. If you’ve been read Miranda, that talking across that table that should not be happening. If you let that happen, you’re a fool. You are absolutely a fool. I can’t make it any clearer. Page – 13 – of 13 Evan Nappen 39:55 It reminds me of Jan Davis. I don’t know if you know about her. (https://www.cbsnews.com/news/parachutist-plunges-to-death/) She was part of a movement to allow the ability to try to legalize base jumping at national parks. At Yosemite, she jumped off of El Capitan. Now, she was a pretty experienced jumper, and she did this as a protest. She was wearing criminal stripes outfit. You know the classic black and white stripes of the classic prison outfit, right? And it was to make a statement, of course. But it’s really a statement here because she jumped off El Capitan, El Capitan, and it was in 1999 as a protest against the ban. She was trying to make a statement that this is safe and should be allowed. She jumped, and ended up going splat. That’s right. Her parachute didn’t open. She couldn’t get it opened, unfortunately. My understanding is she had used some other equipment instead of her own. She didn’t know where the pull was on this particular one, and unfortunately, she ended up just going splat. Evan Nappen 40:19 To me, Jan Davis jumping off El Capitan there is what happens when you are read Miranda and talk. You ended up in a big splat like Jan Davis. That’s what you do. And keep in mind, she was wearing that prisoner outfit when it happened. So, folks, stand on your rights. Don’t be a good a GOFU. Don’t end up splattering yourself on the gun rights oppression and insane matrix of New Jersey gun laws. Should you ever be put in that situation – shut up! Evan Nappen 42:09 This is Evan Nappen and Teddy Nappen reminding you that gun laws don’t protect honest citizens from criminals. They protect criminals from honest citizens. Speaker 2 42:23 Gun Lawyer is a CounterThink Media production. The music used in this broadcast was managed by Cosmo Music, New York, New York. Reach us by emailing Evan@gun.lawyer. The information and opinions in this broadcast do not constitute legal advice. Consult a licensed attorney in your state. Downloadable PDF TranscriptGun Lawyer S5 E299_Transcript About The HostEvan Nappen, Esq.Known as “America's Gun Lawyer,” Evan Nappen is above all a tireless defender of justice. Author of eight bestselling books and countless articles on firearms, knives, and weapons history and the law, a certified Firearms Instructor, and avid weapons collector and historian with a vast collection that spans almost five decades — it's no wonder he's become the trusted, go-to expert for local, industry and national media outlets. Regularly called on by radio, television and online news media for his commentary and expertise on breaking news Evan has appeared countless shows including Fox News – Judge Jeanine, CNN – Lou Dobbs, Court TV, Real Talk on WOR, It's Your Call with Lyn Doyle, Tom Gresham's Gun Talk, and Cam & Company/NRA News. As a creative arts consultant, he also lends his weapons law and historical expertise to an elite, discerning cadre of movie and television producers and directors, and novelists. He also provides expert testimony and consultations for defense attorneys across America. Email Evan Your Comments and Questions  talkback@gun.lawyer Join Evan's InnerCircleHere's your chance to join an elite group of the Savviest gun and knife owners in America.  Membership is totally FREE and Strictly CONFIDENTIAL.  Just enter your email to start receiving insider news, tips, and other valuable membership benefits.   Email (required) *First Name *Select list(s) to subscribe toInnerCircle Membership Yes, I would like to receive emails from Gun Lawyer Podcast. (You can unsubscribe anytime)Constant Contact Use. Please leave this field blank.var ajaxurl = "https://gun.lawyer/wp-admin/admin-ajax.php";

In Keep Canada Weird Jordan and Aaron Airport explore the weird and offbeat Canadian news stories from the past week. In this episode your hosts discuss; Unannounced Fireworks That Upset Ottawa Armed Pokemon Card Robbery in Montreal Sand Castle Sabotage in BC A Forest of License Plates on Cape Breton Island Series Links Keep Canada Weird Series: https://www.thecanadiangothic.com/keep-canada-weird Send a voice memo: www.thecanadiangothic.com/contact Join the Keep Canada Weird Discussion Group: https://www.facebook.com/groups/keepcanadaweird Provide feedback and comments on the episode: thecanadiangothic.com/contact Subscribe to the show: thecanadiangothic.com/subscribe Contact: Website: https://www.thecanadiangothic.com Facebook: https://www.facebook.com/TheCanadianGothic Instagram: https://www.instagram.com/thecanadiangothic/ Support the show: https://www.patreon.com/thecanadiangothic Learn more about your ad choices. Visit megaphone.fm/adchoices

As Independence Day passes, the news always gets weird.See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Johan Bruyneel and Spencer Martin break down Olav Kooij's dominant sprint win in Pau, what it says about the current sprint hierarchy, and what it means for the battle for the Green Jersey. They also preview tomorrow's Stage 6, a massive mountain stage in the Pyrenees, which will show us who the race's serious GC contenders are. Become a WEDŪ Member Today to Unlock VIP Access & Benefits: https://access.wedu.team Lagoon: Use code MOVE for 15% off at https://LagoonSleep.com/THEMOVE Caldera Lab: A small habit with big results. Go to https://CalderaLab.com/THEMOVE and use code THEMOVE for 20% off your first order. Nextbets: Use our special link to see where you can bet in your area and claim the best sign-up offers https://nxtbets.com/betoutcomes/ Ventum: Use code TheMove10 for 10% off anything at: https://ventumracing.com/

If you ever wondered what happens when billionaires decide regular air conditioning just isn't dramatic enough anymore... welcome to today's episode.The gang starts by diving into the latest luxury trend making its way out of Paris: inflatable, air-conditioned tracksuits. Because apparently owning a mansion, a yacht, and an indoor snow room isn't enough anymore. Why sweat like the rest of us when you can look like the Michelin Man while staying frosty? The crew debates whether these ridiculous designer suits are actually genius or just another way for rich people to spend money on problems the rest of us solve with a box fan.Then things take a sharp turn into one of the most hilariously frustrating games we've played in a while: Real or Fake Fireworks. Contestants try to figure out whether names like "Laser Kitties," "Pizza Pocalypse," "Neighbor Hater," "Sparkle Fart," and "Psycho Peacock" are actual fireworks you can buy... or something Rafe made up after five cups of coffee.Spoiler alert: everyone quickly learns that the real fireworks manufacturers are somehow weirder than we are.Between terrible guesses, lifelines that occasionally help (and occasionally don't), Boston accents that completely fall apart, nonstop roasting of contestants, and the crew laughing at names that absolutely should not exist, today's daily comedy podcast proves once again that reality is almost always funnier than fiction.If you've ever wandered through a fireworks tent thinking, "There's no way that's a real product," this episode is for you. You'll also discover why putting air conditioning inside your clothes may be the future... even if it makes you look like Violet Beauregarde after the blueberry incident.Whether you're here for weird news, pop culture commentary, bizarre luxury trends, hilarious games, or just need a reason to laugh through another hot summer day, we've got you covered.Thanks for hanging out with The Rizzuto Show—where every conversation somehow spirals into complete nonsense, and honestly, we wouldn't have it any other way.If you laughed, share the episode with someone who still has leftover fireworks hidden in their garage.Follow The Rizzuto Show → https://linktr.ee/rizzshow for more from your favorite daily comedy podcast.Connect with The Rizzuto Show Comedy Podcast online → https://1057thepoint.com/RizzShow.Hear The Rizz Show daily on the radio at 105.7 The Point | Hubbard Radio in St. Louis, MO.See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Facing storms and safety warnings, President Trump delivers a July 4th address. FIFA lifts a controversial suspension on American star Folarin Balogun, while a phone call from President Trump prompts questions of political influence. And Brad takes a ride on the historic “Big Boy” locomotive as part of its nationwide whistle-stop tour.  Learn more about your ad choices. Visit podcastchoices.com/adchoices

Tony P kicks things off with a brutal case of the runs after a shady "Mexican" deli sandwich while Opie roasts World Cup drama — including Trump personally calling FIFA to overturn a red card. Then social media legend Be A Man (Harmon) finally joins the show for raw, hilarious guy talk: growing up around connected uncles in Boston, judging amateur stripper nights, prepping to take on Joey Chestnut, man bags vs. CVS plastic bags, and why modern dudes need to "be a man." Fourth of July fireworks insanity, Norwegian smoke shows vs. American soccer fans, no-sugar suffering, and zero-filter banter. Classic Opie Radio chaos. Turn it on and thank us later. If you want to help the show https://www.paypal.com/ncp/payment/JANCGHFW7GJHA it's definitely appreciated!  PayPal - opieradio

Rather than celebrating America's 250th birthday, something historic that will never happen again, the Washington Post decided to publish an article attacking fireworks. Glenn shares what he felt when he walked through the National Mall in Washington, D.C., which was nearly empty, but farther down, thousands were gathered to watch a World Cup game. Glenn discusses how every American was born with rights they did nothing to earn and the dangers that come when the debt for those rights is someone else's problem. Glenn outlines the history of the Declaration of Independence, which had intense secret debates, compromises, threats of death, and a betrayal. George Washington University Law School professor Jonathan Turley joins to discuss the flawed SCOTUS decision to uphold birthright citizenship and what America can do next. Rep. Anna Paulina Luna (R-Fla.) joins to give the latest update on the SAVE America Act and weighs the chances of it passing. Glenn and Rep. Luna also discuss the prospects for a peace deal with Iran, the current status of negotiations, and the likeability of the Clintons.  Learn more about your ad choices. Visit megaphone.fm/adchoices