Podcasts about flamingos

Matt and Abby open up about what it feels like to raise boys in a moment where conversations around masculinity feel louder and more polarized than ever. They unpack the difference between healthy and unhealthy masculinity, how criticism aimed at grown men can be internalized by young boys, and the role social media plays in amplifying extremes. They also share their biggest hopes as parents — raising sons with strength, humility, emotional awareness, and a sense of purpose in a rapidly changing world. This episode is sponsored by Zoc Doc, Better Help, Kindred Bravely, Olipop, and Flamingo. Zoc Doc: Stop putting off those doctors appointments and go to https://Zocdoc.com/UNPLANNED to find and instantly book a doctor you love today. Better Help:Sign up and get 10% off at https://BetterHelp/unplannedpodcast Kindred Bravely: Get 20% off your first order at https://KindredBravely.com/UNPLANNED with promo code UNPLANNED. Exclusions apply. OLIPOP: Get a free can of OLIPOP when you buy any 2 cans in store—just visit https://drinkolipop.com/unplanned Flamingo: Our listeners get the Flamingo Starter Set for just $7 at https://www.shopflamingo.com/unplanned Learn more about your ad choices. Visit podcastchoices.com/adchoices

The Flamingo Hotel and Casino is a Las Vegas icon, known for its lush outdoor wildlife habitat and its namesake pink residents. But at 5:00 a.m. on March 3, 2026, surveillance cameras captured a scene that was far from a tourist attraction. 33-year-old Mitchell Fairbarn allegedly breached the habitat, resulting in a terrifying ordeal for a flamingo named "Peachy." --For early, ad free episodes and monthly exclusive bonus content, join our Patreon! Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

Marzo de 1966 se merece una segunda entrega dedicada a recordar singles que alcanzaron su puesto más alto en el Billboard Hot 100 hace 60 años. Sonido Motown y otros sabores del soul, canciones sobre agentes secretos o superhéroes, los Beatles haciendo country o un original del cantante alicantino Jaime Morey adaptado al inglés aparecen entre las curiosidades del mes.Playlist;(sintonía) SLIM HARPO “Baby scratch my back” (top 16)WILSON PICKETT “634-5789 (Soulsville, U.S.A.)” (top 13)MARVIN GAYE “One more heartache” (top 29)EDWIN STARR “Stop Her On Sight (S.O.S.)” (top 48)TIM TAM and THE TURN ONS “Wait a minute” (top 76)THE FLAMINGOS “Boogaloo party” (top 93)JAMO THOMAS and HIS PARTY BROTHERS ORCHESTRA “I spy (for the F.B.I.)” (top 98)THE VENTURES “Secret agent man” (top 54)THE NEWBEATS “Shake hands (And come out crying)” (top 92)BOB KUBAN and THE IN-MEN “The cheater” (top 12)THE SUNRAYS “Andrea” (top 41)NEAL HEFTI “Batman theme” (top 35)JAN and DEAN “Batman” (top 66)DINO DESI and BILLY “Superman” (top 94)THE SEARCHERS “Take me for what I’m worth” (top 76)THE BEATLES “What goes on” (top 81)VERDELLE SMITH “In my room” (top 62)THE BARBARIANS “Moulty” (top 90)THE MAD LADS “I want someone” (top 74)Escuchar audio

In-sauna testical covers are being banned in Japan. Tourist arrested after kidnapping and torturing a flamingo at The Flamingo in Las Vegas. Police warn there are corcodiles everywhere after major flooding in Australia.Weird AF News is the only daily weird news podcast in the world. Weird news 5 days/week and on Friday it's only Floridaman. SUPPORT by joining the Weird AF News Patreon http://patreon.com/weirdafnews - OR buy Jonesy a coffee at http://buymeacoffee.com/funnyjones Buy MERCH: https://weirdafnews.merchmake.com/ - Check out the official website https://WeirdAFnews.com and FOLLOW host Jonesy at http://instagram.com/funnyjones - wants Jonesy to come perform standup comedy in your city? Fill out the form: https://docs.google.com/forms/d/e/1FAIpQLSfvYbm8Wgz3Oc2KSDg0-C6EtSlx369bvi7xdUpx_7UNGA_fIw/viewform

11am hour of The K&C Masterpiece! Which Mavs players need to be part of this team going forward? AM ON THE FM: Mario Day / A Canadian tourist and a flamingo... at The Flamingo. Gridiron Gravy: Catching you up on who went where and which top free agents remain unsigned

A Canadian tourist was arrested for stealing a flamingo while staying at the Flamingo hotel in Vegas! He's being charged with torturing the flamingo and his passport has been taken away until his hearing in May.

Jamie went to another workout yesterday and it's inspired BJ to look into the Stretch Lab! Remember, make your BJ and Jamie Bead Me Signs for Saturday! A Canadian tourist was arrested in Vegas for stealing a flamingo at the Flamingo hotel and his passport has been confiscated.

Komm mit auf die Schmusedecke, denn da ist immer etwas los! Der Flamingo ist ganz verdreht. Er würde gern wieder gerade und auf einem Bein stehen. Vielleicht kann ihm ja der Bär helfen. Das Sandmännchen hat dir aber nicht nur diese Geschichte mitgebracht, sondern auch noch das Kinderlied "Flamingos" von Robert Metcalf.

Learn more about your ad choices. Visit podcastchoices.com/adchoices

A shirtless tourist tries to kidnap a flamingo from a Las Vegas casino and somehow that's only the beginning. This week the gals dive into the absolute animal chaos of Vegas: the bizarre story of the Flamingo Hotel heist, the shocking truth behind a famous Vegas show featuring dancing orangutans, and Elvis Presley's pet chimp who developed a taste for whiskey and started attacking guests at Graceland. Meanwhile Cassie brings the darkness with the horrifying story of a French doctor who preyed on Jewish people trying to escape during WWII. Flamingos, primates, Vegas madness and wartime monsters... just another normal week on The Creep Dive.

A Las Vegas judge issues a strong warning for the man accused of kidnapping and torturing a Flamingo, Nevada offers its own proposal on the future of the Colorado river, an event highlights career opportunities for young women in construction and more on 7@7.

A dumbass tortured a Flamingo bird, David Copperfield is ending his show, Metallica is selling lots of tickets and Tim reviews EPiC The post FHBM #998: Countrifica Live at Sphere first appeared on Five Hundy By Midnight.

0:00-1:00 – Show Open1:00-10:00 – Cookie Day and Dave tried new sugar-free Oreos10:00-13:00 – Middle name pride day13:00-16:00 – Why Gen Z isn't taking part in bar culture as much16:00-30:00 – AT-home happy hours are becoming more popular30:00-35:00 – Marketplace meetup turned into hit-and-run35:00-39:00 – Update on the guy who was attacked after PS5 Marketplace meet-up39:00-42:00 – SW Airlines testing cabin cleaner fees42:00-48:00 – United Airlines may ban you for watching things on your phone without headphones48:00-51:00 – Amber scared Dave this morning51:00-54:00 – Guy riding electric scooter on freeway54:00-58:00 – TGIFridays waitress helped kid who was having a meltdown at restaurant58:00-1:04:00 – University basketball team assistant coach accused of being a pimp1:04:00-1:08:00 – Snow storm trapped people at Alaska basketball gym1:08:00-1:18:00 – Britney Spears DUI update1:18:00-1:22:00 – Savannah Guthrie will return to Today Show1:22:00-1:24:00 – Kansas salt mine1:24:00-1:27:00 – Married couple from The Amazing Race suing for making them look horrible1:27:00-1:31:00 – Afro Man says lawsuit violates his freedom1:31:00-1:39:00 – Guy breaks into old couple's home and old lady threatens him with gun named ‘Sweet Jane'1:39:00-1:46:00 – Jail kitchen worker busted having sex with inmates1:46:00-1:49:00 – Guy admits to creepily touching women's hair on Metro trains1:49:00-1:58:00 – Teacher filmed himself peeing in classroom1:58:00 – 2:04:00 – Another rub and tug called Jenny Spa busted2:04:00-2:11:00 – Digital billboard rejected guy's ad that featured him shirtless2:11:00-2:15:00 – Camels booted from beauty contest for using injectables2:15:00-2:19:00 – Man shoved 2 kids off bikes at concert2:19:00-2:25:00 – Ask Dave & Chuck: Fiancée flipped out because he opened her mail2:25:00-2:31:00 – Ask Dave & Chuck: friend wants to take him to Scotland2:31:00-2:35:00 – Ask Dave & Chuck: friend can't keep a lady because he's a jerk2:35:00-2:45:00 – Ask Dave & Chuck: Should he dump his GF for telling him to leave his job?2:45:00-2:48:00 – Guy's TikTok videos helped bust him for reckless motorcycle driving2:48:00-2:52:00 – Bunk bed collapse caught on camera2:52:00-2:55:00 – Another kid gets stuck in same claw game as another kid recently2:55:00-2:56:00 – Daylight Saving Time2:56:00-3:00:00 – Woman used Find My iPhone to save husband from avalanche3:00:00-3:02:00 – Wendy's hiring Chief Tasting Officer3:02:00-3:05:00 – Girl Scouts selling cookies outside of weed dispensaries3:05:00-End – Canadian guy arrested after stealing flamingo from Flamingo hotel in Las VegasSee Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

This hour starts with the Chump Line, followed by Police Blotter Fax Friday and a look into the new epsiode of Meet the Experts with Bill Brusard of Autoshop Answers!  Visit the Howie Carr Radio Network website to access columns, podcasts, and other exclusive content.

Link Up w/The Morning Sickness Digitally All Over:Instagram: @hms_98_official, @bosskupd, @bretvesely, @dickToledoX/Twitter: @HMSon98, @DickToledo, @bretveselyFacebook: @HMSKUPDYouTube: @hmspodcast9320, @98kupdRequest/Call in/Wakeup Song line:(IN AZ) 585.9800More HMS: holmbergpodcast.com, 98kupd.comEmail: dtoledo@98kupd.com, bvesely@98kupd.com, bbogen@98kupd.comSee Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Douchebag.See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Tim Conway Jr. Hour 2 (3.5) Tim Conway Jr. covers a wild mix of SoCal chaos and headline-grabbing stories, starting with major freeway closures and work-zone headaches as Angel helps break down where traffic is getting tied up. The crew also talks about the Brady Bunch house in Studio City being named a historical monument, sparking a little Hollywood nostalgia, before getting into Apple’s latest affordable laptop. Then things get even stranger with the report of Britney Spears being arrested for DUI, a big interview with legendary horse trainer Doug O’Neill ahead of the Santa Anita Classic Meet, and one of the most bizarre stories of the night—a tourist accused of bird-napping and torturing a flamingo at a famous Las Vegas resort. It’s a packed hour with traffic, pop culture, racing, and pure insanity. See omnystudio.com/listener for privacy information.

The Break Room (FRIDAY 3/6/26) 8am Hour 1) It's the safe way to travel if you've been drinking even though it might be illegal 2) Travel agents still exist? 3) Flamingo thief

A tourist in Nevada faces felony animal cruelty charges after police say he broke into a flamingo habitat on the Las Vegas Strip, grabbed one of the birds, and carried it back to his hotel room while allegedly torturing it. An eighteen-year-old Wisconsin man will spend the rest of his life in prison after admitting he murdered his parents in a plot to fund an assassination attempt against President Trump. Drew Nelson reports.See omnystudio.com/listener for privacy information.

Link Up w/The Morning Sickness Digitally All Over:Instagram: @hms_98_official, @bosskupd, @bretvesely, @dickToledoX/Twitter: @HMSon98, @DickToledo, @bretveselyFacebook: @HMSKUPDYouTube: @hmspodcast9320, @98kupdRequest/Call in/Wakeup Song line:(IN AZ) 585.9800More HMS: holmbergpodcast.com, 98kupd.comEmail: dtoledo@98kupd.com, bvesely@98kupd.com, bbogen@98kupd.comSee Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

See omnystudio.com/listener for privacy information.

Shut up, shrimp. Flamingo assault!!! Selling your kidneys for Metallica tix. See omnystudio.com/listener for privacy information.

A man faces animal cruelty charges after allegedly kidnapping a Flamingo at a strip hotel-casino, California's Governor takes a trip to Las Vegas, a new sphere residency kicks off tonight and more on 7@7.

*5:00am: What Is Something That Unfairly Gets A Bad Rap? *6:00am: How Many Times Do You Re-Wear Gym Clothes Before Washing?, Movie Day With Son *7:00am: Do I Sign Up For Neighbor Mediation? *8:00am: What Is The One Thing Tourists Do In Las Vegas That Drives You Crazy?

Josh has Quick Takes as we honor Lou Holtz, Flamingos taken from the Flamingo in Las Vegas, Going solo for Tourney Time and more!

The man who broke into the flamingo enclosure at 5am and stole a flamingo to take to his room at the Flamingo casino in Las Vegas!

An 80-million-dollar yacht crashed into a bridge...somehow a woman who was handcuffed in back of a cop car was still able to consume cocaine that was hidden in her underwear...and a man stole a flamingo from the Flamingo casino in Vegas!?

www.buymeacoffee.com/reallifewithjennie www.christconnection.cc/jennie https://youtu.be/dhyMfvCq1Mo  - Youtube version

Welcome back to the GGtMC!!!This week we kick off No Murder March and we talk about When the Cat Comes (1963) directed by Vojtech Jasny and The Flamingo Kid (1984) directed by Gary Marshall!!!Emails to midnitecinema@gmail.comAdios!!!

GrabOne is getting a second life under brand new owners. Wellington's Paradigm Group has bought the GrabOne brand and assets. It went into liquidation last October, leaving many consumers with vouchers they couldn't use. The new boss Paul Raeburn says they're bringing back half a dozen key GrabOne employees. He says they're energised to get the platform back to what it was. "I stopped looking at GrabOne probably five or six years ago, because there wasn't anything I wanted to do there. We've got some real heat today - Cordis Hotels, Flamingo scooters, Holy Moly, all of those staples." LISTEN ABOVESee omnystudio.com/listener for privacy information.

1. The Social Media Geniuses at Tampa International AirportShoutout to TPA's viral social strategy.Recurring character: Phoebe the Flamingo (giant airport flamingo mascot).Known for humorous, tongue-in-cheek posts.2. The Viral PostRecap of the joke announcement:Claiming they already banned Crocs.Now tackling the “even larger crisis” — pajamas at the airport.Dramatic language: “The madness stops today. The movement starts now.”Clearly satire… right?3. The Internet Takes It SeriouslyOutrage from people outside Tampa.Major outlets pick it up:New York PostABC NewsInStyleHeadlines frame it as a real controversy.Discussion: Clickbait vs. context.4. Why the Joke Worked So WellDeadpan delivery.Commitment to the bit.Absurdity (a flamingo endorsing policy changes).The power of social media virality.5. The Pajama DebateShould you wear pajamas to the airport?Comfort vs. presentation.Travel culture: 1950s formal vs. modern casual.Personal takes:“Comfy is fine.”“At least look presentable.”Dressing for your destination.See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Bundeskanzler Merz ist zurück von seiner China-Reise. Das Land spielt bei der Unterstützung des russischen Kriegs gegen die Ukraine eine wichtige Rolle. In Peking wollte der Bundeskanzler u.a. seine Gastgeber dazu bringen, sich stärker für eine Beendigung des Ukraine-Krieges einzusetzen. In einem gemeinsamen deutsch-chinesischen Statement wird immerhin die "Ukraine-Frage" erwähnt. Und: Die früher regelmäßigen Regierungskonsultationen zwischen Deutschland und China sollen wieder aufgenommen werden, nachdem sie in den letzten Jahren ins Stocken geraten waren - ein Erfolg. "Kanäle wieder aufzumachen" - das sei das Ziel des Bundeskanzlers, sagt Prof. Dr. May-Britt Stumbaum, die Direktorin des Spear-Instituts, einer sicherheitspolitischen Denkfabrik, und assoziierte Professorin an der Bundeswehr-Universität München. Allerdings: Auf Augenhöhe sehe sich China nur mit den USA, Russland sei für Peking "nützlich". Wichtig für Deutschland sei, "zu Hause" stark zu werden und den eigenen Mittelstand zu schützen - insbesondere vor chinesischer Technologie- und Wirtschaftsspionage, so Stumbaum, die auch Reserve-Offizierin der Luftwaffe ist. Sie berichtet im Gespräch mit Anna Engelke von Chinas kognitiver Kriegsführung. Dabei geht es darum, die Wahrnehmung des Gegners zu verändern. Dafür nutze China auch TikTok: Durch kurze Videos solle die Konzentrationsfähigkeit der User:innen untergraben werden: "Innerhalb Chinas sind diese Reels wesentlich länger. Außerhalb Chinas sind diese Reels wesentlich kürzer und das hat die Auswirkung, dass sie die Konzentrationsfähigkeit systematisch untergraben. Und natürlich ist das Zentrum einer Demokratie, dass ich mich konzentrieren kann, um mich mit den Informationen auseinanderzusetzen", so die Sicherheits-Expertin. China habe sehr viel in KI und Big Data investiert: "Dass man damit eben Inhalte für soziale Medien kreieren kann und die stimulieren eben diese Polarisierung in der Gesellschaft", sagt Stumbaum.Außerdem berichten Stefan Niemann und Kai Küstner von der aktuellen Lage in der Ukraine: Diese konnte den dritten erfolgreichen Treffer ihres Marschflugkörpers "Flamingo" gegen russische Ziele verzeichnen. Nach Erkenntnissen des US-Thinktanks Institute for the Study of War hat Russland nach zweijähriger Invasion die ostukrainische Stadt Pokrowsk eingenommen. Unterdessen hat der Haushaltausschuss des Bundestages grünes Licht für Kamikazedrohnen gegeben.Lob und Kritik, alles bitte per Mail an streitkraefte@ndr.de Aktueller Lagebericht des Institute for the Study of War (ISW), u.a. zur russischen Eroberung Pokrowks: https://understandingwar.org/research/russia-ukraine/russian-offensive-campaign-assessment-february-26-2026/ Interview mit Prof. Dr. May-Britt Stumbaum: https://www.ndr.de/nachrichten/info/audio-429720.html Alle Folgen von “Streitkräfte und Strategien”: https://www.ndr.de/nachrichten/info/podcast2998.html Podcast-Tipp: “Ready for Liftoff! Der Raumfahrtpodcast”: https://1.ard.de/ready-for-liftoff

** VIDEO EN NUESTRO CANAL DE YOUTUBE **** https://youtube.com/live/sv8o_N7453k +++++ Hazte con nuestras camisetas en https://www.bhmshop.app +++++ El frente sur ucraniano vuelve a moverse con intensidad. Las fuerzas ucranianas han logrado recuperar 400 kilómetros cuadrados en sectores clave del sur, obligando a unidades rusas a retroceder y reorganizar sus posiciones defensivas. Si la cifra se consolida, estaríamos ante uno de los avances territoriales más significativos en esa zona en los últimos meses. Pero en una guerra prolongada, el terreno ganado solo es decisivo si se consolida, se fortifica y se resiste el inevitable contraataque. La cuestión no es únicamente cuántos kilómetros se recuperan, sino qué significa este movimiento: ¿Repliegue táctico ruso o deterioro estructural en el sur? ¿Puede Moscú absorber la pérdida sin movilizar nuevas reservas? ¿Estamos ante un cambio de dinámica o ante una oscilación temporal del frente? LIBRO · UN MUNDO CONVULSO. Claves geopolíticas para entender el siglo XXI Las guerras modernas se deciden por desgaste, resiliencia y capacidad industrial. Si quieres comprender cómo encajan avances territoriales como este en un conflicto prolongado, este libro ofrece el marco histórico y estratégico necesario. Compra aquí: https://amzn.to/4qqd41e En este episodio de Frente de Batalla, Francisco García Campa y José María Rodríguez analizan el alcance real del avance ucraniano y sus implicaciones operativas y estratégicas. SUSCRÍBETE A @BELLUMARTISHISTORIAMILITAR Y @BELLUMARTISHISTORIAMILITAR para no perderte ningún programa y únete a nuestra comunidad de apasionados por la historia, la geopolítica y el análisis crítico. Apóyanos para seguir creando contenido riguroso e independiente: Patreon: https://www.patreon.com/bellumartis PayPal: https://www.paypal.me/bellumartis Bizum: 656 778 825 Síguenos también en redes: Instagram: https://www.instagram.com/bellumartis Twitter / X: https://twitter.com/Bellumartis Bellumartis Historia Militar — Porque entender el pasado es prepararse para el futuro. #Ucrania #Rusia #GuerraEnUcrania #4AñosDeGuerra #Putin #Geopolítica #ActualidadMilitar #OTAN #EuropaEnGuerra #ConflictoUcraniano #GuerraModerna #Drones #AnálisisMilitar #FrenteDelEste #Bellumartis

How much does a black bear eat each day? Do flamingos fly? And why does a giraffe have hooves? Curious questions from you listeners guide us into God's wild and wonderful world on this special kid-made episode!Episode Links:Explore Apologia's award-winning science courses and curriculum at: https://www.apologia.com/Episodes Mentioned:Lesson 23: Bears Don't Hibernate — 4 Cool Ways God Designed Creatures to Rest: https://player.captivate.fm/episode/3511d4f1-617f-4742-b3f9-ab6c55f1ea50/Lesson 9: Are All Black Bears Black?: https://player.captivate.fm/episode/78e0351c-b224-40f7-9041-7627d528eef2/Lesson 31: What Is The Difference Between A Turtle And Tortoise?: https://player.captivate.fm/episode/dccb26a1-9dde-498f-8d6d-573f1478e243/Why Do Giraffes Have Spots? Lesson 105: https://player.captivate.fm/episode/57dc20eb-2e33-4b58-9765-fa726c5ba736/Why And How Do Leaves Change Colors? Lesson 49: https://player.captivate.fm/episode/38cf2f29-1a36-4c54-81ba-becb473615a2/Can a Narwhal Get a Brain Freeze? Lesson 102: https://player.captivate.fm/episode/5beea601-57de-4eaf-a8d7-9f95a80f57cf/This podcast contains paid advertisements.This podcast uses the following third-party services for analysis: Podder - https://www.podderapp.com/privacy-policy

Eiropas līderi kara ceturtajā gadadienā pulcējas Kijivā un apliecina atbalstu Ukrainai. Karš turpinās kā novājināšanas karš ar lieliem dzīvā spēka zaudējumiem. Ukrainas ekonomika turas. Aktualitātes analizē Austrumeiropas pilitikas paētījumu centra vecākais pētnieks Armands Astukevičs un Zemessardzes komandieris brigādes ģenerālis Aivars Krjukovs. Sazināmies ar Latvijas Universitātes Ekonomikas un sociālo zinātņu fakultātes dekānu Jāni Priedi. Ar muti Kijivā, ar darbiem…? Vakar, 24. fbruārī, apritot ceturtajai gadskārtai kopš Krievijas agresijas kara eskalācijas Ukrainā, Kijivā ieradās vairāki Eiropas Savienības un tās dalībvalstu līderi. Klāt bija Eiropas Komisijas prezidente Urzula fon der Leiena, Eiropadomes prezidents Antoniu Košta, Somijas prezidents Aleksandrs Stubs, Norvēģijas, Zviedrijas, Dānijas, Horvātijas, Igaunijas, Latvijas, Islandes premjerministri, Lietuvas aizsardzības ministrs, arī NATO ģenerālsekretārs Marks Rite. Tas bija nepārprotams solidaritātes žests, kam jāapliecina savienības apņēmība turpināt atbalstīt Ukrainu visos iespējamos veidos. Līdz šim ir darīts daudz: savienības palīdzības kopapjoms tuvojas divsimt miljardu robežai. Tai skaitā, runājot par šobrīd īpaši aktuālo enerģētikas jautājumu, pārvietotas veselas elektrostacijas un piegādāti apmēram vienpadsmit tūkstoši ģeneratoru. Vairumam Eiropas valstu netrūkst vēlmes un gatavības, taču nupat palīdzības vezuma ceļā kā kupls cinis jau atkal aptupies Ungārijas premjers Viktors Orbans. Vispirms pirmdien notikušajā Eiropas Savienības ārlietu dienestu vadītāju sanāksmē Briselē Ungārijas ārlietu ministrs Peters Sijarto paziņoja, ka Ungārija neatbalstīs kārtējo Krievijai noteikto sankciju paketi, savukārt vakar, tieši pilna mēroga iebrukuma gadadienā, izpaudās pats Orbans, paziņojot, ka bloķēs arī jau nolemto Eiropas Savienības 90 miljardu atbalsta piešķīrumu Ukrainai. Par šo atbalstu agrāk tika panākta vienošanās, kas paredz, ka „negribošo koalīcija” – Ungārija, Slovākija un Čehija – tiek atbrīvotas no saistībām aizdevuma sakarā. Taču nu Budapeštas pusdiktators izdomājis, ka neatbalstīšot vispār nevienu Ukrainai labvēlīgu lēmumu, jo Kijiva, raugi, esot pārtraukusi krievu jēlnaftas piegādes Ungārijai un Slovākijai pa cauruļvadu „Draudzība”. Ukrainas valdība apgalvo, ka piegādes apstājušās, jo cauruļvads bojāts krievu lidrobotu triecienā. Duetā ar savu kaimiņu velk arī Slovākijas premjers Roberts Fico, kura dzimtenei arī Kremļa „melno zeltu” vajagot kā ēst. Viņš piedraudējis, ka ja piegādes neatsāksies, Slovākija pārtrauks elektroenerģijas piegādi Ukrainai. Bet kamēr Eiropas līderi neskopojas nīgriem izteicieniem par „Draudzības” trubai piezīsties kāro Viktoru, Vašingtonā Ukrainas jautājums šķiet nobīdīts otrajā plānā aiz iespējamās Irānas militārās pārmācīšanas, Epstīna failu blāķiem un, protams, prezidenta ķīviņa ar augstāko tiesu par tarifiem. Tiesa, pirms dažām dienām, kad bez nozīmīgiem rezultātiem bija noslēgušās trīspusējās ASV, Krievijas un Ukrainas sarunas Ženēvā, Baltā nama saimnieks pagarināja Krievijai noteikto sankciju termiņu. Baisi gausais karš Apritot ceturtajai gadskārtai kopš Krievijas pilna mēroga iebrukuma Ukrainā, karadarbība tiek raksturota kā novājināšanas karš ar lieliem dzīvā spēka zaudējumiem. Par to, cik dzīvību ziedots Kremļa diktatora iegribu un iedomu vārdā, ir visai aptuvens priekšstats, bet dažādi avoti lēš, ka Ukrainas pusē kritušo skaits varētu pārsniegt 60000, savukārt Krievijas pusē šīs aplēses svārstās no apmēram 180000 līdz 350000 un vairāk tūkstošiem. Kopējie zaudējumi, saprotams, ir vairākas reizes lielāki, un, kā domā NATO analītiķi, pērnajā gadā vien Krievijai tie varētu būt apmēram 400000 kritušo, ievainoto un bez vēsts pazudušo. Kā zināms, šī cena maksāta par pieticīgiem ieguvumiem – pērngad agresoram izdevies papildus iegūt mazāk par procentu no Ukrainas teritorijas, un tagad tā okupēta ir apmēram piektā daļa no kaimiņvalsts. Tomēr daudzi eksperti spriež, ka Krievijai esot vēl diezgan resursu, lai šādi turpinātu vismaz kādu gadu. Izskan gan arī viedokļi, ka rekrutēšanas apjomi atpaliek no dzīvā spēka zaudējumiem un tuvojas brīdis, kad var nākties izšķirties par piespiedu mobilizāciju. Pie tam frontē pēdējā mēneša laikā Krievijas spēki piedzīvojuši nopietnas komunikācijas problēmas. Īlona Maska kompānija beidzot atslēgusi no „Starlink” tīkla nelegāli iegūtos termināļus, kurus krievi izmantoja okupētajā Ukrainas teritorijā, savukārt Maskavas pati bloķējusi „Telegram” tīklu, un daudzas krievu vienības tādējādi palikušas bez ierastajiem saziņas līdzekļiem. Daļēji ar to tiek skaidroti Ukrainas spēku nesenie panākumi, atgūstot ap 200 kvadrātkilometru teritorijas Zaporižjes un Dņipro apgabalos. Vēl pirms tam decembrī krievu vienības izdevās izspiest no Harkivas apgabala Kupjanskas, kuru Krievijas armijas vadība jau bija pasludinājusi par pilnībā ieņemtu. No vienas puses, tie ir nenozīmīgi taktiski ieguvumi, kas, cita starpā, nav mazinājuši Krievijas spēku spiedienu Doņeckas apgabalā, no otras, tas ir apliecinājums, ka Ukrainas armija saglabā uzbrukuma potenciālu. Tiek gan atzīmēts, ka arī Ukraina saskaras ar nopietnām militārā personāla problēmām – apmēram divsimt tūkstoši karavīru, nespējot izturēt frontes apstākļus, esot patvaļīgi pametuši savas vienības. Vēl viens ļoti nepatīkams pārsteigums Krievijai bija sestdien notikušais Ukrainas raķešu trieciens militāro raķešu rūpnīcai Votkisnkā, Udmurtijas autonomajā republikā, aptuveni 1400 kilometru attālumā no Ukrainas robežas. Šajā rūpnīcā top mazā rādiusa raķetes „Iskander”, kas tiek izmantotas triecieniem pa Ukrainas teritoriju, un starpkontinentālās raķetes „Topoļ-M”.  Kā apgalvo Kijiva, trieciens veikts ar ukraiņu ražojuma spārnoto raķeti „Flamingo”. Ukrainas ekonomika turas Krievijas agresijas eskalācija 2022. gadā saprotami traumēja arī Ukrainas ekonomiku. Vairākkārt palielinājās Krievijas okupētās teritorijas apmēri, agresorvalsts uzsāka mērķtiecīgu infrastruktūras graušanu, miljoniem iedzīvotāju pameta valsti. Tiek lēsts, ka pagājušajos kara gados agresors pret Ukrainas teritoriju raidījis apmēram 13000 raķešu un vairāk nekā 140000 lidrobotu. Lielu daļu no tiem notriekusi ukraiņu pretgaisa aizsardzība, taču daļa savu mērķi sasnieguši. Tomēr Ukrainas iekšzemes kopprodukts, kas 2022. gadā saruka par gandrīz trešdaļu, nākamajos gados piedzīvoja zināmu atlabšanu, 2023. gadā pieaugot par vairāk nekā pieciem procentiem, 2024. gadā – par vairāk nekā trīs ar pusi, 2025. gadā – par aptuveni diviem procentiem. Kāpums prognozēts arī šogad, tiesa, šais prognozēs nebija ņemti vērā Krievijas nežēlīgi mērķtiecīgie triecieni enerģētikas infrastruktūrai. Resursa „Project Syndicate” autori, ekonomisti Tatjana Derjugina, Anastasija Fedika un Jurijs Gorodņičenko piesauc trīs galvenos faktorus, kas ļāvuši Ukrainas ekonomikai līdz šim saglabāt kondīciju, kas šobrīd pārspēj cerības pilna mēroga kara sākumā. Pirmkārt, tās ir ukraiņu militārās spējas, saglabājot kontroli pār savu gaisa telpu un lielā mērā neitralizējot Krievijas Melnās jūras floti. Attiecīgi Krievijai nav izdevies pilnībā apturēt Ukrainas eksporta plūsmu pa jūras ceļiem. Otrs faktors ir apjomīgā starptautiskā palīdzība, kas aizvadītajos četros gados bijusi vidēji ap 40 miljardiem dolāru gadā. Tā palīdzējusi kompensēt budžeta deficītu, kas ir aptuveni 25% no iekšzemes kopprodukta, segt lielu daļu izdevumu ieroču un energoresursu importam. Savukārt grandiozais militāro izdevumu kāpums – no sešiem miljardiem dolāru 2021. gadā līdz septiņdesmit miljardiem pērngad – ir jaudīgs ekonomikas stimuls. Pēc amerikāņu domnīcas „Jamestown” sniegtajiem datiem lidrobotu ražošanas apjomi Ukrainā pagājušogad sasnieguši četrus miljonus vienību, bet šogad varētu sasniegt septiņus miljonus. Kā trešais faktors tiek minēta Ukrainas Nacionālās bankas sekmīgā darbība, nodrošinot likviditāti un novēršot banku sistēmas sabrukumu. Un, kā atzīmē trīs minētie „Project Syndicate” autori, salīdzinoši stabilais ekonomikas stāvoklis nebūtu iespējams bez ukraiņu uzņēmēju un visas sabiedrības gatavības pielāgoties un paciest grūtības, un radoši meklēt risinājumus. Protams, Krievijas agresijas karš rada Ukrainai milzu zaudējumus un arī milzīgas problēmas, no kurām akūtākās šobrīd ir teju trīs ceturtdaļu elektroģenerējošo jaudu iznīcināšana un jūtams kvalificēta darbaspēka trūkums. Sagatavoja Eduards Liniņš.

Jeffrey destaca una noticia y unos lanzamientos que les encantarán.¿Qué estoy usando hoy?:SchonDSGN Black Ultem (M) Pelikan Edelstein JadeTinteríasMadrid Pen Show (13 y 14 de noviembre)Barcelona Pen Show Gris TrencadísLamy AL-star Flamingo y PinePelikan Classic M200 Festivales mundiales Cherry BlossomPlatinum #3776 Century 2.0 Demonstrator GTPlatinum BISOTintería del capítulo: ARMONÍA

2026-02-22 | UPDATES #138 | Massive blow for Russia — the Votkinsk strike.A bold, deep-strike into the heart of Russia's missile industry — Ukraine has hit the Votkinsk Machine Building Plant. It's a long-range, high-value operation that — if confirmed — could blunt Moscow's ability to replenish ballistic and cruise missiles.On the night of 20–21 February 2026 — Ukrainian forces say they struck the JSC Votkinsk Machine Building Plant in the Udmurt Republic, a core site that manufactures engines and components for systems including Iskander, Oreshnik and some larger strategic missiles. Kyiv's General Staff publicly stated the strike and confirmed some details, saying it used domestically produced FP-5 “Flamingo” ground-launched cruise missiles. (ukrinform.net)----------SUPPORT THE CHANNEL:https://www.buymeacoffee.com/siliconcurtainhttps://www.patreon.com/siliconcurtainhttps://www.gofundme.com/f/scaling-up-campaign-to-fight-authoritarian-disinformation----------A REQUEST FOR HELP!I'm heading back to Kyiv this week, to film, do research and conduct interviews. The logistics and need for equipment and clothing are a little higher than for my previous trips. It will be cold, and may be dark also. If you can, please assist to ensure I can make this trip a success. My commitment to the audience of the channel, will be to bring back compelling interviews conducted in Ukraine, and to use the experience to improve the quality of the channel, it's insights and impact. Let Ukraine and democracy prevail! https://buymeacoffee.com/siliconcurtain/extrashttps://www.patreon.com/siliconcurtainhttps://www.gofundme.com/f/scaling-up-campaign-to-fight-authoritarian-disinformationNONE OF THIS CAN HAPPEN WITHOUT YOU!So what's next? We're going to Kyiv in January 2026 to film on the ground, and will record interviews with some huge guests. We'll be creating opportunities for new interviews, and to connect you with the reality of a European city under escalating winter attack, from an imperialist, genocidal power. PLEASE HELP ME ME TO GROW SILICON CURTAINWe are planning our events for 2026, and to do more and have a greater impact. After achieving more than 12 events in 2025, we will aim to double that! 24 events and interviews on the ground in Ukraine, to push back against weaponized information, toxic propaganda and corrosive disinformation. Please help us make it happen!----------SOURCES: Reuters: governor of Udmurtia says Ukrainian drones damaged site.Kyiv Independent: report on Flamingo strike, eyewitness videos and social posts. Defence-blog / Military analysis: satellite imagery and workshop damage assessment. UNN / local Ukrainian reporting: photos showing extensive damage to workshops. Pravda / Ukrainian defence reporting: FP-5 Flamingo missile claimed by Kyiv. UPI & LA Times: international wire coverage and context on reach and significance. Militantnyi / regional defence outlet: open-source satellite imagery analysis reporting specific workshop damage (No. 22, No. 36). ----------

Send a textAs it's been a while; a new Q&A!!This week I answer questions on protein-levels for postpartum recovery (for athletes and non-athletes) and I explain why I think most people DO NOT need a Personal Trainer that specialises/has experience with a particular condition (AFTER rehab is completed, of course, because I'm not a maniac)As always; HPNB only has 5 billing cycles. So this means that you not only get 3 months FREE access, no obligation! BUT, if you decide you want to do the rest of the program, after only 5 months of paying $10/£8 a month you now get FREE LIFE TIME ACCESS! That's $50 max spend, in case you were wondering. Though I'm not terribly active on  Instagram and Facebook you can follow us there. I am however active on Threads so find me there! And, of course, you can always find us on our YouTube channel if you like your podcast in video form :) Visit healthypostnatalbody.com and get 3 months completely FREE access. No sales, no commitment, no BS. Email peter@healthypostnatalbody.com if you have any questions, comments or want to suggest a guest/topic       Playing us out "Dresden the Flamingo".

Or: Spanish Gold. A treasure hunt intrigues Shannon and Saylor, even though King Moses suggests it's bogus. Regardless, it should be a fun job… right? Or is there danger lurking…

Your host Andy, Ant and Dun take a look at Saturday's Horror Show vs Blackburn Rovers, and also have the pleasure of speaking with the voice of the R's, Tyler Morris about his career in commentary to date, and does he need a bungee cord for his own safety?- Now Sam Field has gone on loan, who is the new scapegoat?- It's the team, it's the whole damn team - Dun goes full Jason Kelce minus the gravelly voice, and mummers outfit...- After three away games where they put their case forward as one of the best units in the league, The defense rests... - Disaster averted - Both Madsen and Kone come out relatively unscathed -The cavalry arrives. JCs on the bench, chair, Varane - Sam Field IS better off on loan...Not sure about Morrison though- The Nourry Q&A...Took a strange turn...by 90 degrees- New York warms up, as does the Winter Olympics- The Tyler Morris interview- The return of Blighty Bulletin - which is just 'fine'.- Ant delves deep into his kitbag to find Hull and Southampton kits. Does Yellow and Teal pair well with a tinfoil hat?- ...and are there Flamingos in Flamingo Land - Just how bold will our predictions be for the away games at Hull and Southampton? - Jacob goes as dark as Ant's Originals gorgeous blackout shirt- Lovely Stuff - Joy in repetition, Bones and Meat in a box for Valentines Day and the awarding of inaugural QPR NYC Peace Prize.- No meet up at the Factory this weekend, back for the SaintsRate, review, share, follow, listen, stream, download, and check out QPR NYC on Big Cartel to check out our merch including Ant's Originals...

Aujourd'hui, je vous propose de repartir explorer les zones humides. 

Today we've got a wide variety of topics covering Elon Musk's Boring Company being tied to Epic Universe, a MASSIVE manatee rescue, and your Q&A!

Florida's feathered fight continues, as the effort to change the state bird finally passes through the House. But will it take flight in the Senate? Ryan, Dana, and Chris Trenkmann talk about this latest effort and the person who should be getting credit for it.

Florida's feathered fight continues, as the effort to change the state bird finally passes through the House. But will it take flight in the Senate? Ryan, Dana, and Chris Trenkmann talk about this latest effort and the person who should be getting credit for it. See omnystudio.com/listener for privacy information.

From rewriting Google's search stack in the early 2000s to reviving sparse trillion-parameter models and co-designing TPUs with frontier ML research, Jeff Dean has quietly shaped nearly every layer of the modern AI stack. As Chief AI Scientist at Google and a driving force behind Gemini, Jeff has lived through multiple scaling revolutions from CPUs and sharded indices to multimodal models that reason across text, video, and code.Jeff joins us to unpack what it really means to “own the Pareto frontier,” why distillation is the engine behind every Flash model breakthrough, how energy (in picojoules) not FLOPs is becoming the true bottleneck, what it was like leading the charge to unify all of Google's AI teams, and why the next leap won't come from bigger context windows alone, but from systems that give the illusion of attending to trillions of tokens.We discuss:* Jeff's early neural net thesis in 1990: parallel training before it was cool, why he believed scaling would win decades early, and the “bigger model, more data, better results” mantra that held for 15 years* The evolution of Google Search: sharding, moving the entire index into memory in 2001, softening query semantics pre-LLMs, and why retrieval pipelines already resemble modern LLM systems* Pareto frontier strategy: why you need both frontier “Pro” models and low-latency “Flash” models, and how distillation lets smaller models surpass prior generations* Distillation deep dive: ensembles → compression → logits as soft supervision, and why you need the biggest model to make the smallest one good* Latency as a first-class objective: why 10–50x lower latency changes UX entirely, and how future reasoning workloads will demand 10,000 tokens/sec* Energy-based thinking: picojoules per bit, why moving data costs 1000x more than a multiply, batching through the lens of energy, and speculative decoding as amortization* TPU co-design: predicting ML workloads 2–6 years out, speculative hardware features, precision reduction, sparsity, and the constant feedback loop between model architecture and silicon* Sparse models and “outrageously large” networks: trillions of parameters with 1–5% activation, and why sparsity was always the right abstraction* Unified vs. specialized models: abandoning symbolic systems, why general multimodal models tend to dominate vertical silos, and when vertical fine-tuning still makes sense* Long context and the illusion of scale: beyond needle-in-a-haystack benchmarks toward systems that narrow trillions of tokens to 117 relevant documents* Personalized AI: attending to your emails, photos, and documents (with permission), and why retrieval + reasoning will unlock deeply personal assistants* Coding agents: 50 AI interns, crisp specifications as a new core skill, and how ultra-low latency will reshape human–agent collaboration* Why ideas still matter: transformers, sparsity, RL, hardware, systems — scaling wasn't blind; the pieces had to multiply togetherShow Notes:* Gemma 3 Paper* Gemma 3* Gemini 2.5 Report* Jeff Dean's “Software Engineering Advice fromBuilding Large-Scale Distributed Systems” Presentation (with Back of the Envelope Calculations)* Latency Numbers Every Programmer Should Know by Jeff Dean* The Jeff Dean Facts* Jeff Dean Google Bio* Jeff Dean on “Important AI Trends” @Stanford AI Club* Jeff Dean & Noam Shazeer — 25 years at Google (Dwarkesh)—Jeff Dean* LinkedIn: https://www.linkedin.com/in/jeff-dean-8b212555* X: https://x.com/jeffdeanGoogle* https://google.com* https://deepmind.googleFull Video EpisodeTimestamps00:00:04 — Introduction: Alessio & Swyx welcome Jeff Dean, chief AI scientist at Google, to the Latent Space podcast00:00:30 — Owning the Pareto Frontier & balancing frontier vs low-latency models00:01:31 — Frontier models vs Flash models + role of distillation00:03:52 — History of distillation and its original motivation00:05:09 — Distillation's role in modern model scaling00:07:02 — Model hierarchy (Flash, Pro, Ultra) and distillation sources00:07:46 — Flash model economics & wide deployment00:08:10 — Latency importance for complex tasks00:09:19 — Saturation of some tasks and future frontier tasks00:11:26 — On benchmarks, public vs internal00:12:53 — Example long-context benchmarks & limitations00:15:01 — Long-context goals: attending to trillions of tokens00:16:26 — Realistic use cases beyond pure language00:18:04 — Multimodal reasoning and non-text modalities00:19:05 — Importance of vision & motion modalities00:20:11 — Video understanding example (extracting structured info)00:20:47 — Search ranking analogy for LLM retrieval00:23:08 — LLM representations vs keyword search00:24:06 — Early Google search evolution & in-memory index00:26:47 — Design principles for scalable systems00:28:55 — Real-time index updates & recrawl strategies00:30:06 — Classic “Latency numbers every programmer should know”00:32:09 — Cost of memory vs compute and energy emphasis00:34:33 — TPUs & hardware trade-offs for serving models00:35:57 — TPU design decisions & co-design with ML00:38:06 — Adapting model architecture to hardware00:39:50 — Alternatives: energy-based models, speculative decoding00:42:21 — Open research directions: complex workflows, RL00:44:56 — Non-verifiable RL domains & model evaluation00:46:13 — Transition away from symbolic systems toward unified LLMs00:47:59 — Unified models vs specialized ones00:50:38 — Knowledge vs reasoning & retrieval + reasoning00:52:24 — Vertical model specialization & modules00:55:21 — Token count considerations for vertical domains00:56:09 — Low resource languages & contextual learning00:59:22 — Origins: Dean's early neural network work01:10:07 — AI for coding & human–model interaction styles01:15:52 — Importance of crisp specification for coding agents01:19:23 — Prediction: personalized models & state retrieval01:22:36 — Token-per-second targets (10k+) and reasoning throughput01:23:20 — Episode conclusion and thanksTranscriptAlessio Fanelli [00:00:04]: Hey everyone, welcome to the Latent Space podcast. This is Alessio, founder of Kernel Labs, and I'm joined by Swyx, editor of Latent Space. Shawn Wang [00:00:11]: Hello, hello. We're here in the studio with Jeff Dean, chief AI scientist at Google. Welcome. Thanks for having me. It's a bit surreal to have you in the studio. I've watched so many of your talks, and obviously your career has been super legendary. So, I mean, congrats. I think the first thing must be said, congrats on owning the Pareto Frontier.Jeff Dean [00:00:30]: Thank you, thank you. Pareto Frontiers are good. It's good to be out there.Shawn Wang [00:00:34]: Yeah, I mean, I think it's a combination of both. You have to own the Pareto Frontier. You have to have like frontier capability, but also efficiency, and then offer that range of models that people like to use. And, you know, some part of this was started because of your hardware work. Some part of that is your model work, and I'm sure there's lots of secret sauce that you guys have worked on cumulatively. But, like, it's really impressive to see it all come together in, like, this slittily advanced.Jeff Dean [00:01:04]: Yeah, yeah. I mean, I think, as you say, it's not just one thing. It's like a whole bunch of things up and down the stack. And, you know, all of those really combine to help make UNOS able to make highly capable large models, as well as, you know, software techniques to get those large model capabilities into much smaller, lighter weight models that are, you know, much more cost effective and lower latency, but still, you know, quite capable for their size. Yeah.Alessio Fanelli [00:01:31]: How much pressure do you have on, like, having the lower bound of the Pareto Frontier, too? I think, like, the new labs are always trying to push the top performance frontier because they need to raise more money and all of that. And you guys have billions of users. And I think initially when you worked on the CPU, you were thinking about, you know, if everybody that used Google, we use the voice model for, like, three minutes a day, they were like, you need to double your CPU number. Like, what's that discussion today at Google? Like, how do you prioritize frontier versus, like, we have to do this? How do we actually need to deploy it if we build it?Jeff Dean [00:02:03]: Yeah, I mean, I think we always want to have models that are at the frontier or pushing the frontier because I think that's where you see what capabilities now exist that didn't exist at the sort of slightly less capable last year's version or last six months ago version. At the same time, you know, we know those are going to be really useful for a bunch of use cases, but they're going to be a bit slower and a bit more expensive than people might like for a bunch of other broader models. So I think what we want to do is always have kind of a highly capable sort of affordable model that enables a whole bunch of, you know, lower latency use cases. People can use them for agentic coding much more readily and then have the high-end, you know, frontier model that is really useful for, you know, deep reasoning, you know, solving really complicated math problems, those kinds of things. And it's not that. One or the other is useful. They're both useful. So I think we'd like to do both. And also, you know, through distillation, which is a key technique for making the smaller models more capable, you know, you have to have the frontier model in order to then distill it into your smaller model. So it's not like an either or choice. You sort of need that in order to actually get a highly capable, more modest size model. Yeah.Alessio Fanelli [00:03:24]: I mean, you and Jeffrey came up with the solution in 2014.Jeff Dean [00:03:28]: Don't forget, L'Oreal Vinyls as well. Yeah, yeah.Alessio Fanelli [00:03:30]: A long time ago. But like, I'm curious how you think about the cycle of these ideas, even like, you know, sparse models and, you know, how do you reevaluate them? How do you think about in the next generation of model, what is worth revisiting? Like, yeah, they're just kind of like, you know, you worked on so many ideas that end up being influential, but like in the moment, they might not feel that way necessarily. Yeah.Jeff Dean [00:03:52]: I mean, I think distillation was originally motivated because we were seeing that we had a very large image data set at the time, you know, 300 million images that we could train on. And we were seeing that if you create specialists for different subsets of those image categories, you know, this one's going to be really good at sort of mammals, and this one's going to be really good at sort of indoor room scenes or whatever, and you can cluster those categories and train on an enriched stream of data after you do pre-training on a much broader set of images. You get much better performance. If you then treat that whole set of maybe 50 models you've trained as a large ensemble, but that's not a very practical thing to serve, right? So distillation really came about from the idea of, okay, what if we want to actually serve that and train all these independent sort of expert models and then squish it into something that actually fits in a form factor that you can actually serve? And that's, you know, not that different from what we're doing today. You know, often today we're instead of having an ensemble of 50 models. We're having a much larger scale model that we then distill into a much smaller scale model.Shawn Wang [00:05:09]: Yeah. A part of me also wonders if distillation also has a story with the RL revolution. So let me maybe try to articulate what I mean by that, which is you can, RL basically spikes models in a certain part of the distribution. And then you have to sort of, well, you can spike models, but usually sometimes... It might be lossy in other areas and it's kind of like an uneven technique, but you can probably distill it back and you can, I think that the sort of general dream is to be able to advance capabilities without regressing on anything else. And I think like that, that whole capability merging without loss, I feel like it's like, you know, some part of that should be a distillation process, but I can't quite articulate it. I haven't seen much papers about it.Jeff Dean [00:06:01]: Yeah, I mean, I tend to think of one of the key advantages of distillation is that you can have a much smaller model and you can have a very large, you know, training data set and you can get utility out of making many passes over that data set because you're now getting the logits from the much larger model in order to sort of coax the right behavior out of the smaller model that you wouldn't otherwise get with just the hard labels. And so, you know, I think that's what we've observed. Is you can get, you know, very close to your largest model performance with distillation approaches. And that seems to be, you know, a nice sweet spot for a lot of people because it enables us to kind of, for multiple Gemini generations now, we've been able to make the sort of flash version of the next generation as good or even substantially better than the previous generations pro. And I think we're going to keep trying to do that because that seems like a good trend to follow.Shawn Wang [00:07:02]: So, Dara asked, so it was the original map was Flash Pro and Ultra. Are you just sitting on Ultra and distilling from that? Is that like the mother load?Jeff Dean [00:07:12]: I mean, we have a lot of different kinds of models. Some are internal ones that are not necessarily meant to be released or served. Some are, you know, our pro scale model and we can distill from that as well into our Flash scale model. So I think, you know, it's an important set of capabilities to have and also inference time scaling. It can also be a useful thing to improve the capabilities of the model.Shawn Wang [00:07:35]: And yeah, yeah, cool. Yeah. And obviously, I think the economy of Flash is what led to the total dominance. I think the latest number is like 50 trillion tokens. I don't know. I mean, obviously, it's changing every day.Jeff Dean [00:07:46]: Yeah, yeah. But, you know, by market share, hopefully up.Shawn Wang [00:07:50]: No, I mean, there's no I mean, there's just the economics wise, like because Flash is so economical, like you can use it for everything. Like it's in Gmail now. It's in YouTube. Like it's yeah. It's in everything.Jeff Dean [00:08:02]: We're using it more in our search products of various AI mode reviews.Shawn Wang [00:08:05]: Oh, my God. Flash past the AI mode. Oh, my God. Yeah, that's yeah, I didn't even think about that.Jeff Dean [00:08:10]: I mean, I think one of the things that is quite nice about the Flash model is not only is it more affordable, it's also a lower latency. And I think latency is actually a pretty important characteristic for these models because we're going to want models to do much more complicated things that are going to involve, you know, generating many more tokens from when you ask the model to do so. So, you know, if you're going to ask the model to do something until it actually finishes what you ask it to do, because you're going to ask now, not just write me a for loop, but like write me a whole software package to do X or Y or Z. And so having low latency systems that can do that seems really important. And Flash is one direction, one way of doing that. You know, obviously our hardware platforms enable a bunch of interesting aspects of our, you know, serving stack as well, like TPUs, the interconnect between. Chips on the TPUs is actually quite, quite high performance and quite amenable to, for example, long context kind of attention operations, you know, having sparse models with lots of experts. These kinds of things really, really matter a lot in terms of how do you make them servable at scale.Alessio Fanelli [00:09:19]: Yeah. Does it feel like there's some breaking point for like the proto Flash distillation, kind of like one generation delayed? I almost think about almost like the capability as a. In certain tasks, like the pro model today is a saturated, some sort of task. So next generation, that same task will be saturated at the Flash price point. And I think for most of the things that people use models for at some point, the Flash model in two generation will be able to do basically everything. And how do you make it economical to like keep pushing the pro frontier when a lot of the population will be okay with the Flash model? I'm curious how you think about that.Jeff Dean [00:09:59]: I mean, I think that's true. If your distribution of what people are asking people, the models to do is stationary, right? But I think what often happens is as the models become more capable, people ask them to do more, right? So, I mean, I think this happens in my own usage. Like I used to try our models a year ago for some sort of coding task, and it was okay at some simpler things, but wouldn't do work very well for more complicated things. And since then, we've improved dramatically on the more complicated coding tasks. And now I'll ask it to do much more complicated things. And I think that's true, not just of coding, but of, you know, now, you know, can you analyze all the, you know, renewable energy deployments in the world and give me a report on solar panel deployment or whatever. That's a very complicated, you know, more complicated task than people would have asked a year ago. And so you are going to want more capable models to push the frontier in the absence of what people ask the models to do. And that also then gives us. Insight into, okay, where does the, where do things break down? How can we improve the model in these, these particular areas, uh, in order to sort of, um, make the next generation even better.Alessio Fanelli [00:11:11]: Yeah. Are there any benchmarks or like test sets they use internally? Because it's almost like the same benchmarks get reported every time. And it's like, all right, it's like 99 instead of 97. Like, how do you have to keep pushing the team internally to it? Or like, this is what we're building towards. Yeah.Jeff Dean [00:11:26]: I mean, I think. Benchmarks, particularly external ones that are publicly available. Have their utility, but they often kind of have a lifespan of utility where they're introduced and maybe they're quite hard for current models. You know, I, I like to think of the best kinds of benchmarks are ones where the initial scores are like 10 to 20 or 30%, maybe, but not higher. And then you can sort of work on improving that capability for, uh, whatever it is, the benchmark is trying to assess and get it up to like 80, 90%, whatever. I, I think once it hits kind of 95% or something, you get very diminishing returns from really focusing on that benchmark, cuz it's sort of, it's either the case that you've now achieved that capability, or there's also the issue of leakage in public data or very related kind of data being, being in your training data. Um, so we have a bunch of held out internal benchmarks that we really look at where we know that wasn't represented in the training data at all. There are capabilities that we want the model to have. Um, yeah. Yeah. Um, that it doesn't have now, and then we can work on, you know, assessing, you know, how do we make the model better at these kinds of things? Is it, we need different kind of data to train on that's more specialized for this particular kind of task. Do we need, um, you know, a bunch of, uh, you know, architectural improvements or some sort of, uh, model capability improvements, you know, what would help make that better?Shawn Wang [00:12:53]: Is there, is there such an example that you, uh, a benchmark inspired in architectural improvement? Like, uh, I'm just kind of. Jumping on that because you just.Jeff Dean [00:13:02]: Uh, I mean, I think some of the long context capability of the, of the Gemini models that came, I guess, first in 1.5 really were about looking at, okay, we want to have, um, you know,Shawn Wang [00:13:15]: immediately everyone jumped to like completely green charts of like, everyone had, I was like, how did everyone crack this at the same time? Right. Yeah. Yeah.Jeff Dean [00:13:23]: I mean, I think, um, and once you're set, I mean, as you say that needed single needle and a half. Hey, stack benchmark is really saturated for at least context links up to 1, 2 and K or something. Don't actually have, you know, much larger than 1, 2 and 8 K these days or two or something. We're trying to push the frontier of 1 million or 2 million context, which is good because I think there are a lot of use cases where. Yeah. You know, putting a thousand pages of text or putting, you know, multiple hour long videos and the context and then actually being able to make use of that as useful. Try to, to explore the über graduation are fairly large. But the single needle in a haystack benchmark is sort of saturated. So you really want more complicated, sort of multi-needle or more realistic, take all this content and produce this kind of answer from a long context that sort of better assesses what it is people really want to do with long context. Which is not just, you know, can you tell me the product number for this particular thing?Shawn Wang [00:14:31]: Yeah, it's retrieval. It's retrieval within machine learning. It's interesting because I think the more meta level I'm trying to operate at here is you have a benchmark. You're like, okay, I see the architectural thing I need to do in order to go fix that. But should you do it? Because sometimes that's an inductive bias, basically. It's what Jason Wei, who used to work at Google, would say. Exactly the kind of thing. Yeah, you're going to win. Short term. Longer term, I don't know if that's going to scale. You might have to undo that.Jeff Dean [00:15:01]: I mean, I like to sort of not focus on exactly what solution we're going to derive, but what capability would you want? And I think we're very convinced that, you know, long context is useful, but it's way too short today. Right? Like, I think what you would really want is, can I attend to the internet while I answer my question? Right? But that's not going to happen. I think that's going to be solved by purely scaling the existing solutions, which are quadratic. So a million tokens kind of pushes what you can do. You're not going to do that to a trillion tokens, let alone, you know, a billion tokens, let alone a trillion. But I think if you could give the illusion that you can attend to trillions of tokens, that would be amazing. You'd find all kinds of uses for that. You would have attend to the internet. You could attend to the pixels of YouTube and the sort of deeper representations that we can find. You could attend to the form for a single video, but across many videos, you know, on a personal Gemini level, you could attend to all of your personal state with your permission. So like your emails, your photos, your docs, your plane tickets you have. I think that would be really, really useful. And the question is, how do you get algorithmic improvements and system level improvements that get you to something where you actually can attend to trillions of tokens? Right. In a meaningful way. Yeah.Shawn Wang [00:16:26]: But by the way, I think I did some math and it's like, if you spoke all day, every day for eight hours a day, you only generate a maximum of like a hundred K tokens, which like very comfortably fits.Jeff Dean [00:16:38]: Right. But if you then say, okay, I want to be able to understand everything people are putting on videos.Shawn Wang [00:16:46]: Well, also, I think that the classic example is you start going beyond language into like proteins and whatever else is extremely information dense. Yeah. Yeah.Jeff Dean [00:16:55]: I mean, I think one of the things about Gemini's multimodal aspects is we've always wanted it to be multimodal from the start. And so, you know, that sometimes to people means text and images and video sort of human-like and audio, audio, human-like modalities. But I think it's also really useful to have Gemini know about non-human modalities. Yeah. Like LIDAR sensor data from. Yes. Say, Waymo vehicles or. Like robots or, you know, various kinds of health modalities, x-rays and MRIs and imaging and genomics information. And I think there's probably hundreds of modalities of data where you'd like the model to be able to at least be exposed to the fact that this is an interesting modality and has certain meaning in the world. Where even if you haven't trained on all the LIDAR data or MRI data, you could have, because maybe that's not, you know, it doesn't make sense in terms of trade-offs of. You know, what you include in your main pre-training data mix, at least including a little bit of it is actually quite useful. Yeah. Because it sort of tempts the model that this is a thing.Shawn Wang [00:18:04]: Yeah. Do you believe, I mean, since we're on this topic and something I just get to ask you all the questions I always wanted to ask, which is fantastic. Like, are there some king modalities, like modalities that supersede all the other modalities? So a simple example was Vision can, on a pixel level, encode text. And DeepSeq had this DeepSeq CR paper that did that. Vision. And Vision has also been shown to maybe incorporate audio because you can do audio spectrograms and that's, that's also like a Vision capable thing. Like, so, so maybe Vision is just the king modality and like. Yeah.Jeff Dean [00:18:36]: I mean, Vision and Motion are quite important things, right? Motion. Well, like video as opposed to static images, because I mean, there's a reason evolution has evolved eyes like 23 independent ways, because it's such a useful capability for sensing the world around you, which is really what we want these models to be. So I think the only thing that we can be able to do is interpret the things we're seeing or the things we're paying attention to and then help us in using that information to do things. Yeah.Shawn Wang [00:19:05]: I think motion, you know, I still want to shout out, I think Gemini, still the only native video understanding model that's out there. So I use it for YouTube all the time. Nice.Jeff Dean [00:19:15]: Yeah. Yeah. I mean, it's actually, I think people kind of are not necessarily aware of what the Gemini models can actually do. Yeah. Like I have an example I've used in one of my talks. It had like, it was like a YouTube highlight video of 18 memorable sports moments across the last 20 years or something. So it has like Michael Jordan hitting some jump shot at the end of the finals and, you know, some soccer goals and things like that. And you can literally just give it the video and say, can you please make me a table of what all these different events are? What when the date is when they happened? And a short description. And so you get like now an 18 row table of that information extracted from the video, which is, you know, not something most people think of as like a turn video into sequel like table.Alessio Fanelli [00:20:11]: Has there been any discussion inside of Google of like, you mentioned tending to the whole internet, right? Google, it's almost built because a human cannot tend to the whole internet and you need some sort of ranking to find what you need. Yep. That ranking is like much different for an LLM because you can expect a person to look at maybe the first five, six links in a Google search versus for an LLM. Should you expect to have 20 links that are highly relevant? Like how do you internally figure out, you know, how do we build the AI mode that is like maybe like much broader search and span versus like the more human one? Yeah.Jeff Dean [00:20:47]: I mean, I think even pre-language model based work, you know, our ranking systems would be built to start. I mean, I think even pre-language model based work, you know, our ranking systems would be built to start. With a giant number of web pages in our index, many of them are not relevant. So you identify a subset of them that are relevant with very lightweight kinds of methods. You know, you're down to like 30,000 documents or something. And then you gradually refine that to apply more and more sophisticated algorithms and more and more sophisticated sort of signals of various kinds in order to get down to ultimately what you show, which is, you know, the final 10 results or, you know, 10 results plus. Other kinds of information. And I think an LLM based system is not going to be that dissimilar, right? You're going to attend to trillions of tokens, but you're going to want to identify, you know, what are the 30,000 ish documents that are with the, you know, maybe 30 million interesting tokens. And then how do you go from that into what are the 117 documents I really should be paying attention to in order to carry out the tasks that the user has asked? And I think, you know, you can imagine systems where you have, you know, a lot of highly parallel processing to identify those initial 30,000 candidates, maybe with very lightweight kinds of models. Then you have some system that sort of helps you narrow down from 30,000 to the 117 with maybe a little bit more sophisticated model or set of models. And then maybe the final model is the thing that looks. So the 117 things that might be your most capable model. So I think it has to, it's going to be some system like that, that is really enables you to give the illusion of attending to trillions of tokens. Sort of the way Google search gives you, you know, not the illusion, but you are searching the internet, but you're finding, you know, a very small subset of things that are, that are relevant.Shawn Wang [00:22:47]: Yeah. I often tell a lot of people that are not steeped in like Google search history that, well, you know, like Bert was. Like he was like basically immediately inside of Google search and that improves results a lot, right? Like I don't, I don't have any numbers off the top of my head, but like, I'm sure you guys, that's obviously the most important numbers to Google. Yeah.Jeff Dean [00:23:08]: I mean, I think going to an LLM based representation of text and words and so on enables you to get out of the explicit hard notion of, of particular words having to be on the page, but really getting at the notion of this topic of this page or this page. Paragraph is highly relevant to this query. Yeah.Shawn Wang [00:23:28]: I don't think people understand how much LLMs have taken over all these very high traffic system, very high traffic. Yeah. Like it's Google, it's YouTube. YouTube has this like semantics ID thing where it's just like every token or every item in the vocab is a YouTube video or something that predicts the video using a code book, which is absurd to me for YouTube size.Jeff Dean [00:23:50]: And then most recently GROK also for, for XAI, which is like, yeah. I mean, I'll call out even before LLMs were used extensively in search, we put a lot of emphasis on softening the notion of what the user actually entered into the query.Shawn Wang [00:24:06]: So do you have like a history of like, what's the progression? Oh yeah.Jeff Dean [00:24:09]: I mean, I actually gave a talk in, uh, I guess, uh, web search and data mining conference in 2009, uh, where we never actually published any papers about the origins of Google search, uh, sort of, but we went through sort of four or five or six. generations, four or five or six generations of, uh, redesigning of the search and retrieval system, uh, from about 1999 through 2004 or five. And that talk is really about that evolution. And one of the things that really happened in 2001 was we were sort of working to scale the system in multiple dimensions. So one is we wanted to make our index bigger, so we could retrieve from a larger index, which always helps your quality in general. Uh, because if you don't have the page in your index, you're going to not do well. Um, and then we also needed to scale our capacity because we were, our traffic was growing quite extensively. Um, and so we had, you know, a sharded system where you have more and more shards as the index grows, you have like 30 shards. And then if you want to double the index size, you make 60 shards so that you can bound the latency by which you respond for any particular user query. Um, and then as traffic grows, you add, you add more and more replicas of each of those. And so we eventually did the math that realized that in a data center where we had say 60 shards and, um, you know, 20 copies of each shard, we now had 1200 machines, uh, with disks. And we did the math and we're like, Hey, one copy of that index would actually fit in memory across 1200 machines. So in 2001, we introduced, uh, we put our entire index in memory and what that enabled from a quality perspective was amazing. Um, and so we had more and more replicas of each of those. Before you had to be really careful about, you know, how many different terms you looked at for a query, because every one of them would involve a disk seek on every one of the 60 shards. And so you, as you make your index bigger, that becomes even more inefficient. But once you have the whole index in memory, it's totally fine to have 50 terms you throw into the query from the user's original three or four word query, because now you can add synonyms like restaurant and restaurants and cafe and, uh, you know, things like that. Uh, bistro and all these things. And you can suddenly start, uh, sort of really, uh, getting at the meaning of the word as opposed to the exact semantic form the user typed in. And that was, you know, 2001, very much pre LLM, but really it was about softening the, the strict definition of what the user typed in order to get at the meaning.Alessio Fanelli [00:26:47]: What are like principles that you use to like design the systems, especially when you have, I mean, in 2001, the internet is like. Doubling, tripling every year in size is not like, uh, you know, and I think today you kind of see that with LLMs too, where like every year the jumps in size and like capabilities are just so big. Are there just any, you know, principles that you use to like, think about this? Yeah.Jeff Dean [00:27:08]: I mean, I think, uh, you know, first, whenever you're designing a system, you want to understand what are the sort of design parameters that are going to be most important in designing that, you know? So, you know, how many queries per second do you need to handle? How big is the internet? How big is the index you need to handle? How much data do you need to keep for every document in the index? How are you going to look at it when you retrieve things? Um, what happens if traffic were to double or triple, you know, will that system work well? And I think a good design principle is you're going to want to design a system so that the most important characteristics could scale by like factors of five or 10, but probably not beyond that because often what happens is if you design a system for X. And something suddenly becomes a hundred X, that would enable a very different point in the design space that would not make sense at X. But all of a sudden at a hundred X makes total sense. So like going from a disk space index to a in memory index makes a lot of sense once you have enough traffic, because now you have enough replicas of the sort of state on disk that those machines now actually can hold, uh, you know, a full copy of the, uh, index and memory. Yeah. And that all of a sudden enabled. A completely different design that wouldn't have been practical before. Yeah. Um, so I'm, I'm a big fan of thinking through designs in your head, just kind of playing with the design space a little before you actually do a lot of writing of code. But, you know, as you said, in the early days of Google, we were growing the index, uh, quite extensively. We were growing the update rate of the index. So the update rate actually is the parameter that changed the most. Surprising. So it used to be once a month.Shawn Wang [00:28:55]: Yeah.Jeff Dean [00:28:56]: And then we went to a system that could update any particular page in like sub one minute. Okay.Shawn Wang [00:29:02]: Yeah. Because this is a competitive advantage, right?Jeff Dean [00:29:04]: Because all of a sudden news related queries, you know, if you're, if you've got last month's news index, it's not actually that useful for.Shawn Wang [00:29:11]: News is a special beast. Was there any, like you could have split it onto a separate system.Jeff Dean [00:29:15]: Well, we did. We launched a Google news product, but you also want news related queries that people type into the main index to also be sort of updated.Shawn Wang [00:29:23]: So, yeah, it's interesting. And then you have to like classify whether the page is, you have to decide which pages should be updated and what frequency. Oh yeah.Jeff Dean [00:29:30]: There's a whole like, uh, system behind the scenes that's trying to decide update rates and importance of the pages. So even if the update rate seems low, you might still want to recrawl important pages quite often because, uh, the likelihood they change might be low, but the value of having updated is high.Shawn Wang [00:29:50]: Yeah, yeah, yeah, yeah. Uh, well, you know, yeah. This, uh, you know, mention of latency and, and saving things to this reminds me of one of your classics, which I have to bring up, which is latency numbers. Every programmer should know, uh, was there a, was it just a, just a general story behind that? Did you like just write it down?Jeff Dean [00:30:06]: I mean, this has like sort of eight or 10 different kinds of metrics that are like, how long does a cache mistake? How long does branch mispredict take? How long does a reference domain memory take? How long does it take to send, you know, a packet from the U S to the Netherlands or something? Um,Shawn Wang [00:30:21]: why Netherlands, by the way, or is it, is that because of Chrome?Jeff Dean [00:30:25]: Uh, we had a data center in the Netherlands, um, so, I mean, I think this gets to the point of being able to do the back of the envelope calculations. So these are sort of the raw ingredients of those, and you can use them to say, okay, well, if I need to design a system to do image search and thumb nailing or something of the result page, you know, how, what I do that I could pre-compute the image thumbnails. I could like. Try to thumbnail them on the fly from the larger images. What would that do? How much dis bandwidth than I need? How many des seeks would I do? Um, and you can sort of actually do thought experiments in, you know, 30 seconds or a minute with the sort of, uh, basic, uh, basic numbers at your fingertips. Uh, and then as you sort of build software using higher level libraries, you kind of want to develop the same intuitions for how long does it take to, you know, look up something in this particular kind of.Shawn Wang [00:31:21]: I'll see you next time.Shawn Wang [00:31:51]: Which is a simple byte conversion. That's nothing interesting. I wonder if you have any, if you were to update your...Jeff Dean [00:31:58]: I mean, I think it's really good to think about calculations you're doing in a model, either for training or inference.Jeff Dean [00:32:09]: Often a good way to view that is how much state will you need to bring in from memory, either like on-chip SRAM or HBM from the accelerator. Attached memory or DRAM or over the network. And then how expensive is that data motion relative to the cost of, say, an actual multiply in the matrix multiply unit? And that cost is actually really, really low, right? Because it's order, depending on your precision, I think it's like sub one picodule.Shawn Wang [00:32:50]: Oh, okay. You measure it by energy. Yeah. Yeah.Jeff Dean [00:32:52]: Yeah. I mean, it's all going to be about energy and how do you make the most energy efficient system. And then moving data from the SRAM on the other side of the chip, not even off the off chip, but on the other side of the same chip can be, you know, a thousand picodules. Oh, yeah. And so all of a sudden, this is why your accelerators require batching. Because if you move, like, say, the parameter of a model from SRAM on the, on the chip into the multiplier unit, that's going to cost you a thousand picodules. So you better make use of that, that thing that you moved many, many times with. So that's where the batch dimension comes in. Because all of a sudden, you know, if you have a batch of 256 or something, that's not so bad. But if you have a batch of one, that's really not good.Shawn Wang [00:33:40]: Yeah. Yeah. Right.Jeff Dean [00:33:41]: Because then you paid a thousand picodules in order to do your one picodule multiply.Shawn Wang [00:33:46]: I have never heard an energy-based analysis of batching.Jeff Dean [00:33:50]: Yeah. I mean, that's why people batch. Yeah. Ideally, you'd like to use batch size one because the latency would be great.Shawn Wang [00:33:56]: The best latency.Jeff Dean [00:33:56]: But the energy cost and the compute cost inefficiency that you get is quite large. So, yeah.Shawn Wang [00:34:04]: Is there a similar trick like, like, like you did with, you know, putting everything in memory? Like, you know, I think obviously NVIDIA has caused a lot of waves with betting very hard on SRAM with Grok. I wonder if, like, that's something that you already saw with, with the TPUs, right? Like that, that you had to. Uh, to serve at your scale, uh, you probably sort of saw that coming. Like what, what, what hardware, uh, innovations or insights were formed because of what you're seeing there?Jeff Dean [00:34:33]: Yeah. I mean, I think, you know, TPUs have this nice, uh, sort of regular structure of 2D or 3D meshes with a bunch of chips connected. Yeah. And each one of those has HBM attached. Um, I think for serving some kinds of models, uh, you know, you, you pay a lot higher cost. Uh, and time latency, um, bringing things in from HBM than you do bringing them in from, uh, SRAM on the chip. So if you have a small enough model, you can actually do model parallelism, spread it out over lots of chips and you actually get quite good throughput improvements and latency improvements from doing that. And so you're now sort of striping your smallish scale model over say 16 or 64 chips. Uh, but as if you do that and it all fits in. In SRAM, uh, that can be a big win. So yeah, that's not a surprise, but it is a good technique.Alessio Fanelli [00:35:27]: Yeah. What about the TPU design? Like how much do you decide where the improvements have to go? So like, this is like a good example of like, is there a way to bring the thousand picojoules down to 50? Like, is it worth designing a new chip to do that? The extreme is like when people say, oh, you should burn the model on the ASIC and that's kind of like the most extreme thing. How much of it? Is it worth doing an hardware when things change so quickly? Like what was the internal discussion? Yeah.Jeff Dean [00:35:57]: I mean, we, we have a lot of interaction between say the TPU chip design architecture team and the sort of higher level modeling, uh, experts, because you really want to take advantage of being able to co-design what should future TPUs look like based on where we think the sort of ML research puck is going, uh, in some sense, because, uh, you know, as a hardware designer for ML and in particular, you're trying to design a chip starting today and that design might take two years before it even lands in a data center. And then it has to sort of be a reasonable lifetime of the chip to take you three, four or five years. So you're trying to predict two to six years out where, what ML computations will people want to run two to six years out in a very fast changing field. And so having people with interest. Interesting ML research ideas of things we think will start to work in that timeframe or will be more important in that timeframe, uh, really enables us to then get, you know, interesting hardware features put into, you know, TPU N plus two, where TPU N is what we have today.Shawn Wang [00:37:10]: Oh, the cycle time is plus two.Jeff Dean [00:37:12]: Roughly. Wow. Because, uh, I mean, sometimes you can squeeze some changes into N plus one, but, you know, bigger changes are going to require the chip. Yeah. Design be earlier in its lifetime design process. Um, so whenever we can do that, it's generally good. And sometimes you can put in speculative features that maybe won't cost you much chip area, but if it works out, it would make something, you know, 10 times as fast. And if it doesn't work out, well, you burned a little bit of tiny amount of your chip area on that thing, but it's not that big a deal. Uh, sometimes it's a very big change and we want to be pretty sure this is going to work out. So we'll do like lots of carefulness. Uh, ML experimentation to show us, uh, this is actually the, the way we want to go. Yeah.Alessio Fanelli [00:37:58]: Is there a reverse of like, we already committed to this chip design so we can not take the model architecture that way because it doesn't quite fit?Jeff Dean [00:38:06]: Yeah. I mean, you, you definitely have things where you're going to adapt what the model architecture looks like so that they're efficient on the chips that you're going to have for both training and inference of that, of that, uh, generation of model. So I think it kind of goes both ways. Um, you know, sometimes you can take advantage of, you know, lower precision things that are coming in a future generation. So you can, might train it at that lower precision, even if the current generation doesn't quite do that. Mm.Shawn Wang [00:38:40]: Yeah. How low can we go in precision?Jeff Dean [00:38:43]: Because people are saying like ternary is like, uh, yeah, I mean, I'm a big fan of very low precision because I think that gets, that saves you a tremendous amount of time. Right. Because it's picojoules per bit that you're transferring and reducing the number of bits is a really good way to, to reduce that. Um, you know, I think people have gotten a lot of luck, uh, mileage out of having very low bit precision things, but then having scaling factors that apply to a whole bunch of, uh, those, those weights. Scaling. How does it, how does it, okay.Shawn Wang [00:39:15]: Interesting. You, so low, low precision, but scaled up weights. Yeah. Huh. Yeah. Never considered that. Yeah. Interesting. Uh, w w while we're on this topic, you know, I think there's a lot of, um, uh, this, the concept of precision at all is weird when we're sampling, you know, uh, we just, at the end of this, we're going to have all these like chips that I'll do like very good math. And then we're just going to throw a random number generator at the start. So, I mean, there's a movement towards, uh, energy based, uh, models and processors. I'm just curious if you've, obviously you've thought about it, but like, what's your commentary?Jeff Dean [00:39:50]: Yeah. I mean, I think. There's a bunch of interesting trends though. Energy based models is one, you know, diffusion based models, which don't sort of sequentially decode tokens is another, um, you know, speculative decoding is a way that you can get sort of an equivalent, very small.Shawn Wang [00:40:06]: Draft.Jeff Dean [00:40:07]: Batch factor, uh, for like you predict eight tokens out and that enables you to sort of increase the effective batch size of what you're doing by a factor of eight, even, and then you maybe accept five or six of those tokens. So you get. A five, a five X improvement in the amortization of moving weights, uh, into the multipliers to do the prediction for the, the tokens. So these are all really good techniques and I think it's really good to look at them from the lens of, uh, energy, real energy, not energy based models, um, and, and also latency and throughput, right? If you look at things from that lens, that sort of guides you to. Two solutions that are gonna be, uh, you know, better from, uh, you know, being able to serve larger models or, you know, equivalent size models more cheaply and with lower latency.Shawn Wang [00:41:03]: Yeah. Well, I think, I think I, um, it's appealing intellectually, uh, haven't seen it like really hit the mainstream, but, um, I do think that, uh, there's some poetry in the sense that, uh, you know, we don't have to do, uh, a lot of shenanigans if like we fundamentally. Design it into the hardware. Yeah, yeah.Jeff Dean [00:41:23]: I mean, I think there's still a, there's also sort of the more exotic things like analog based, uh, uh, computing substrates as opposed to digital ones. Uh, I'm, you know, I think those are super interesting cause they can be potentially low power. Uh, but I think you often end up wanting to interface that with digital systems and you end up losing a lot of the power advantages in the digital to analog and analog to digital conversions. You end up doing, uh, at the sort of boundaries. And periphery of that system. Um, I still think there's a tremendous distance we can go from where we are today in terms of energy efficiency with sort of, uh, much better and specialized hardware for the models we care about.Shawn Wang [00:42:05]: Yeah.Alessio Fanelli [00:42:06]: Um, any other interesting research ideas that you've seen, or like maybe things that you cannot pursue a Google that you would be interested in seeing researchers take a step at, I guess you have a lot of researchers. Yeah, I guess you have enough, but our, our research.Jeff Dean [00:42:21]: Our research portfolio is pretty broad. I would say, um, I mean, I think, uh, in terms of research directions, there's a whole bunch of, uh, you know, open problems and how do you make these models reliable and able to do much longer, kind of, uh, more complex tasks that have lots of subtasks. How do you orchestrate, you know, maybe one model that's using other models as tools in order to sort of build, uh, things that can accomplish, uh, you know, much more. Yeah. Significant pieces of work, uh, collectively, then you would ask a single model to do. Um, so that's super interesting. How do you get more verifiable, uh, you know, how do you get RL to work for non-verifiable domains? I think it's a pretty interesting open problem because I think that would broaden out the capabilities of the models, the improvements that you're seeing in both math and coding. Uh, if we could apply those to other less verifiable domains, because we've come up with RL techniques that actually enable us to do that. Uh, effectively, that would, that would really make the models improve quite a lot. I think.Alessio Fanelli [00:43:26]: I'm curious, like when we had Noam Brown on the podcast, he said, um, they already proved you can do it with deep research. Um, you kind of have it with AI mode in a way it's not verifiable. I'm curious if there's any thread that you think is interesting there. Like what is it? Both are like information retrieval of JSON. So I wonder if it's like the retrieval is like the verifiable part. That you can score or what are like, yeah, yeah. How, how would you model that, that problem?Jeff Dean [00:43:55]: Yeah. I mean, I think there are ways of having other models that can evaluate the results of what a first model did, maybe even retrieving. Can you have another model that says, is this things, are these things you retrieved relevant? Or can you rate these 2000 things you retrieved to assess which ones are the 50 most relevant or something? Um, I think those kinds of techniques are actually quite effective. Sometimes I can even be the same model, just prompted differently to be a, you know, a critic as opposed to a, uh, actual retrieval system. Yeah.Shawn Wang [00:44:28]: Um, I do think like there, there is that, that weird cliff where like, it feels like we've done the easy stuff and then now it's, but it always feels like that every year. It's like, oh, like we know, we know, and the next part is super hard and nobody's figured it out. And, uh, exactly with this RLVR thing where like everyone's talking about, well, okay, how do we. the next stage of the non-verifiable stuff. And everyone's like, I don't know, you know, Ellen judge.Jeff Dean [00:44:56]: I mean, I feel like the nice thing about this field is there's lots and lots of smart people thinking about creative solutions to some of the problems that we all see. Uh, because I think everyone sort of sees that the models, you know, are great at some things and they fall down around the edges of those things and, and are not as capable as we'd like in those areas. And then coming up with good techniques and trying those. And seeing which ones actually make a difference is sort of what the whole research aspect of this field is, is pushing forward. And I think that's why it's super interesting. You know, if you think about two years ago, we were struggling with GSM, eight K problems, right? Like, you know, Fred has two rabbits. He gets three more rabbits. How many rabbits does he have? That's a pretty far cry from the kinds of mathematics that the models can, and now you're doing IMO and Erdos problems in pure language. Yeah. Yeah. Pure language. So that is a really, really amazing jump in capabilities in, you know, in a year and a half or something. And I think, um, for other areas, it'd be great if we could make that kind of leap. Uh, and you know, we don't exactly see how to do it for some, some areas, but we do see it for some other areas and we're going to work hard on making that better. Yeah.Shawn Wang [00:46:13]: Yeah.Alessio Fanelli [00:46:14]: Like YouTube thumbnail generation. That would be very helpful. We need that. That would be AGI. We need that.Shawn Wang [00:46:20]: That would be. As far as content creators go.Jeff Dean [00:46:22]: I guess I'm not a YouTube creator, so I don't care that much about that problem, but I guess, uh, many people do.Shawn Wang [00:46:27]: It does. Yeah. It doesn't, it doesn't matter. People do judge books by their covers as it turns out. Um, uh, just to draw a bit on the IMO goal. Um, I'm still not over the fact that a year ago we had alpha proof and alpha geometry and all those things. And then this year we were like, screw that we'll just chuck it into Gemini. Yeah. What's your reflection? Like, I think this, this question about. Like the merger of like symbolic systems and like, and, and LMS, uh, was a very much core belief. And then somewhere along the line, people would just said, Nope, we'll just all do it in the LLM.Jeff Dean [00:47:02]: Yeah. I mean, I think it makes a lot of sense to me because, you know, humans manipulate symbols, but we probably don't have like a symbolic representation in our heads. Right. We have some distributed representation that is neural net, like in some way of lots of different neurons. And activation patterns firing when we see certain things and that enables us to reason and plan and, you know, do chains of thought and, you know, roll them back now that, that approach for solving the problem doesn't seem like it's going to work. I'm going to try this one. And, you know, in a lot of ways we're emulating what we intuitively think, uh, is happening inside real brains in neural net based models. So it never made sense to me to have like completely separate. Uh, discrete, uh, symbolic things, and then a completely different way of, of, uh, you know, thinking about those things.Shawn Wang [00:47:59]: Interesting. Yeah. Uh, I mean, it's maybe seems obvious to you, but it wasn't obvious to me a year ago. Yeah.Jeff Dean [00:48:06]: I mean, I do think like that IMO with, you know, translating to lean and using lean and then the next year and also a specialized geometry model. And then this year switching to a single unified model. That is roughly the production model with a little bit more inference budget, uh, is actually, you know, quite good because it shows you that the capabilities of that general model have improved dramatically and, and now you don't need the specialized model. This is actually sort of very similar to the 2013 to 16 era of machine learning, right? Like it used to be, people would train separate models for lots of different, each different problem, right? I have, I want to recognize street signs and something. So I train a street sign. Recognition recognition model, or I want to, you know, decode speech recognition. I have a speech model, right? I think now the era of unified models that do everything is really upon us. And the question is how well do those models generalize to new things they've never been asked to do and they're getting better and better.Shawn Wang [00:49:10]: And you don't need domain experts. Like one of my, uh, so I interviewed ETA who was on, who was on that team. Uh, and he was like, yeah, I, I don't know how they work. I don't know where the IMO competition was held. I don't know the rules of it. I just trained the models, the training models. Yeah. Yeah. And it's kind of interesting that like people with these, this like universal skill set of just like machine learning, you just give them data and give them enough compute and they can kind of tackle any task, which is the bitter lesson, I guess. I don't know. Yeah.Jeff Dean [00:49:39]: I mean, I think, uh, general models, uh, will win out over specialized ones in most cases.Shawn Wang [00:49:45]: Uh, so I want to push there a bit. I think there's one hole here, which is like, uh. There's this concept of like, uh, maybe capacity of a model, like abstractly a model can only contain the number of bits that it has. And, uh, and so it, you know, God knows like Gemini pro is like one to 10 trillion parameters. We don't know, but, uh, the Gemma models, for example, right? Like a lot of people want like the open source local models that are like that, that, that, and, and, uh, they have some knowledge, which is not necessary, right? Like they can't know everything like, like you have the. The luxury of you have the big model and big model should be able to capable of everything. But like when, when you're distilling and you're going down to the small models, you know, you're actually memorizing things that are not useful. Yeah. And so like, how do we, I guess, do we want to extract that? Can we, can we divorce knowledge from reasoning, you know?Jeff Dean [00:50:38]: Yeah. I mean, I think you do want the model to be most effective at reasoning if it can retrieve things, right? Because having the model devote precious parameter space. To remembering obscure facts that could be looked up is actually not the best use of that parameter space, right? Like you might prefer something that is more generally useful in more settings than this obscure fact that it has. Um, so I think that's always attention at the same time. You also don't want your model to be kind of completely detached from, you know, knowing stuff about the world, right? Like it's probably useful to know how long the golden gate be. Bridges just as a general sense of like how long are bridges, right? And, uh, it should have that kind of knowledge. It maybe doesn't need to know how long some teeny little bridge in some other more obscure part of the world is, but, uh, it does help it to have a fair bit of world knowledge and the bigger your model is, the more you can have. Uh, but I do think combining retrieval with sort of reasoning and making the model really good at doing multiple stages of retrieval. Yeah.Shawn Wang [00:51:49]: And reasoning through the intermediate retrieval results is going to be a, a pretty effective way of making the model seem much more capable, because if you think about, say, a personal Gemini, yeah, right?Jeff Dean [00:52:01]: Like we're not going to train Gemini on my email. Probably we'd rather have a single model that, uh, we can then use and use being able to retrieve from my email as a tool and have the model reason about it and retrieve from my photos or whatever, uh, and then make use of that and have multiple. Um, you know, uh, stages of interaction. that makes sense.Alessio Fanelli [00:52:24]: Do you think the vertical models are like, uh, interesting pursuit? Like when people are like, oh, we're building the best healthcare LLM, we're building the best law LLM, are those kind of like short-term stopgaps or?Jeff Dean [00:52:37]: No, I mean, I think, I think vertical models are interesting. Like you want them to start from a pretty good base model, but then you can sort of, uh, sort of viewing them, view them as enriching the data. Data distribution for that particular vertical domain for healthcare, say, um, we're probably not going to train or for say robotics. We're probably not going to train Gemini on all possible robotics data. We, you could train it on because we want it to have a balanced set of capabilities. Um, so we'll expose it to some robotics data, but if you're trying to build a really, really good robotics model, you're going to want to start with that and then train it on more robotics data. And then maybe that would. It's multilingual translation capability, but improve its robotics capabilities. And we're always making these kind of, uh, you know, trade-offs in the data mix that we train the base Gemini models on. You know, we'd love to include data from 200 more languages and as much data as we have for those languages, but that's going to displace some other capabilities of the model. It won't be as good at, um, you know, Pearl programming, you know, it'll still be good at Python programming. Cause we'll include it. Enough. Of that, but there's other long tail computer languages or coding capabilities that it may suffer on or multi, uh, multimodal reasoning capabilities may suffer. Cause we didn't get to expose it to as much data there, but it's really good at multilingual things. So I, I think some combination of specialized models, maybe more modular models. So it'd be nice to have the capability to have those 200 languages, plus this awesome robotics model, plus this awesome healthcare, uh, module that all can be knitted together to work in concert and called upon in different circumstances. Right? Like if I have a health related thing, then it should enable using this health module in conjunction with the main base model to be even better at those kinds of things. Yeah.Shawn Wang [00:54:36]: Installable knowledge. Yeah.Jeff Dean [00:54:37]: Right.Shawn Wang [00:54:38]: Just download as a, as a package.Jeff Dean [00:54:39]: And some of that installable stuff can come from retrieval, but some of it probably should come from preloaded training on, you know, uh, a hundred billion tokens or a trillion tokens of health data. Yeah.Shawn Wang [00:54:51]: And for listeners, I think, uh, I will highlight the Gemma three end paper where they, there was a little bit of that, I think. Yeah.Alessio Fanelli [00:54:56]: Yeah. I guess the question is like, how many billions of tokens do you need to outpace the frontier model improvements? You know, it's like, if I have to make this model better healthcare and the main. Gemini model is still improving. Do I need 50 billion tokens? Can I do it with a hundred, if I need a trillion healthcare tokens, it's like, they're probably not out there that you don't have, you know, I think that's really like the.Jeff Dean [00:55:21]: Well, I mean, I think healthcare is a particularly challenging domain, so there's a lot of healthcare data that, you know, we don't have access to appropriately, but there's a lot of, you know, uh, healthcare organizations that want to train models on their own data. That is not public healthcare data, uh, not public health. But public healthcare data. Um, so I think there are opportunities there to say, partner with a large healthcare organization and train models for their use that are going to be, you know, more bespoke, but probably, uh, might be better than a general model trained on say, public data. Yeah.Shawn Wang [00:55:58]: Yeah. I, I believe, uh, by the way, also this is like somewhat related to the language conversation. Uh, I think one of your, your favorite examples was you can put a low resource language in the context and it just learns. Yeah.Jeff Dean [00:56:09]: Oh, yeah, I think the example we used was Calamon, which is truly low resource because it's only spoken by, I think 120 people in the world and there's no written text.Shawn Wang [00:56:20]: So, yeah. So you can just do it that way. Just put it in the context. Yeah. Yeah. But I think your whole data set in the context, right.Jeff Dean [00:56:27]: If you, if you take a language like, uh, you know, Somali or something, there is a fair bit of Somali text in the world that, uh, or Ethiopian Amharic or something, um, you know, we probably. Yeah. Are not putting all the data from those languages into the Gemini based training. We put some of it, but if you put more of it, you'll improve the capabilities of those models.Shawn Wang [00:56:49]: Yeah.Jeff Dean [00:56:49]:

Donald Trump kündigte Ende 2025 eine massive Aufrüstung der US-Marine an: neue, gigantische Schlachtschiffe, eine eigene "Trump-Klasse" mit maximaler Feuerkraft. Er nennt sie die "goldene Flotte". Doch sind solche Stahlkolosse im Zeitalter von Drohnen und Hyperschallwaffen überhaupt noch zeitgemäß? In der aktuellen Folge spricht Stefan Niemann mit dem Admiral a.D. Manfred Nielson über Trumps Pläne und die Frage, wie eine Seemacht im 21. Jahrhundert aussehen sollte. Nielson ist der Meinung: "Moderne Einheiten müssen heute nicht über Größe überzeugen, sondern durch Technologie". Dabei geht es auch um den strategischen Wettbewerb mit China und um die Rolle der US-Navy in den kommenden Jahrzehnten. Ein historischer Rückblick zeigt, dass die Debatte über militärische Stärke nicht neu ist. Schon im US-Wahlkampf 2012 reagierte Barack Obama auf Kritik an der schrumpfenden amerikanischen Flotte mit dem Hinweis, dass moderne militärische Fähigkeiten nicht durch bloßes Zählen von Schiffen zu erfassen seien - ein Argument, das heute aktueller denn je erscheint.Zuvor blickt Kai Küstner wie gewohnt auf den russischen Krieg gegen die Ukraine. Nachdem die Ukraine im vergangenen Jahr mit der Produktion von Drohnen in Deutschland begonnen hatte, erwartet Präsident Selenskij nun eine erste Auslieferung bereits in den nächsten Tagen. Überhaupt verzahnen sich die Rüstungsindustrien der Ukraine und ihrer europäischen Partner zunehmend: Kyjiw vollzieht einen Strategiewechsel und will künftig Waffen exportieren. Außerdem geht es um den Test einer deutsch-britischen Hyperschallrakete und um die Frage, ob der anfangs so gepriesene ukrainischen Marschflugkörper “Flamingo" hinter den Erwartungen zurückblieb. Auch die Lage an den Fronten und die anhaltenden russischen Luftangriffe auf die ukrainische Energieinfrastruktur werden eingeordnet.Lob und Kritik, alles bitte per Mail an streitkraefte@ndr.de Interview mit Admiral a.D. Manfred Nielson https://www.ndr.de/nachrichten/info/audio-412636.html Flamingo - erfüllt Erwartungen bislang nicht: https://kyivindependent.com/ukraine-strikes-russias-oreshnik-launch-site-in-kapustin-yar-with-flamingo-missiles-general-staff-says/ https://www.tagesspiegel.de/internationales/da-stimmt-was-nicht-ist-der-gross-beworbene-ukrainische-marschflugkorper-flamingo-ein-flop-15207640.html Trump kündigt Bau neuer Kriegsschiffe an: https://www.tagesschau.de/ausland/amerika/trump-kriegsschiffe-100.html Alle Folgen von "Streitkräfte und Strategien" https://www.ndr.de/nachrichten/info/podcast2998.html Podcast-Tipp: 15 Minuten. Der tagesschau-Podcast am Morgen https://www.ardaudiothek.de/sendung/urn:ard:show:b84b465ae5abcd64/

Save 10% on a Las Vegas Advisor 2026 membership and book with code MTM.  https://www.lasvegasadvisor.com/shop/products/lva-membership-platinum/ Episode Description This week the visitor and gaming numbers came in for December, 2025 giving us a picture for the year as a whole. While many metrics were down significantly in 2025, what can we take away from the year and how damaging will it be to the future of Las Vegas. Can the city bring back the everyman and why is gaming revenue not falling as quickly? In other news Four Queens has arrived with the perfect Year of the Horse gift. We also discuss: Caesars garage in ruin, Luger's secret salad, reimagining Flamingo's garden, Nevada Landing's fake website, saving with Las Vegas Advisor, the Hard Rock glass and why Flamingo's 1996 commercial gives us nostalgia. Episode Guide 0:00 Caesars garage an actual ruin? 0:30 Mirage/Hard Rock glass update 1:43 Peter Luger's "secret salad" 2:42 Zoox recovers from mysterious shutdown 4:07 Losing a $450K sidebet 5:25 Harrah's Laughlin Legionnaires 6:44 Flamingo Hilton 1996 ad 7:35 The perfect tiki bar space for Vegas? 8:39 Four Queens insane Year of the Horse gift 10:03 Las Vegas Advisor 10% off - 2026 books now available 11:25 Nevada Landing's retro website 14:10 Vegas 2025 year end numbers 15:31 Visitors, occupancy & room rates down for 2025 16:40 Can Vegas bring back the everyman? 19:00 Looking forward to 2026? Each week tens of thousands of people tune into our MtM Vegas news shows at http://www.YouTube.com/milestomemories. We do two news shows weekly on YouTube with this being the audio version. Never miss out on the latest happenings in and around Las Vegas! Enjoying the podcast? Please consider leaving us a positive review on your favorite podcast platform! You can also connect with us anytime at podcast@milestomemories.com.  You can subscribe on Apple Podcasts, Google Podcasts, Spotify or by searching "MtM Vegas" or "Miles to Memories" in your favorite podcast app. Don't forget to check out our travel/miles/points podcast as well!

Air Week: January 26-February 1, 2026 George Goldner, Pt. 3 – Gone & End Records It’s part 3 of our 3 part series on record man, George Goldner. He is said to have had the “golden ear” for hit records and songwriter Jerry Leiber even complimented his talent for picking hit songs by saying that Goldner had, “the musical taste of a fourteen-year-old-girl.” Born to Jewish immigrants in 1919, Goldner’s first love was Latino dance music and he began his career by opening night clubs and starting Tico Records, a Latino label in 1948. By 1953, he was interested in Rhythm & Blues and began releasing records under the Rama subsidiary. In early 1954, he set up Gee Records and scored a huge hit in early ’56 with The Teenagers, “Why Do Fools Fall In Love.” By mid-’57, due to his gambling debts, Goldner sold Tico, Rama and Gee to alleged mobster Morris Levy. This week, we will take a close look at Goldner’s last R&B labels that he would run independently: Gone & End Records. Both new labels did well with Gone scoring hits with NY vocal group, the Dubs and Goldner-arranged instrumental “7-11 (Mambo No. 5)” by the Gone All Stars featuring Buddy Lucas on tenor sax. End soon followed with million-sellers from The Chantels, The Imperials and The Flamingos. Both labels proved that Goldner still had the magic ear for picking the music teenagers wanted to hear and buy, but eventually both labels would face the same fate as Goldner’s early record companies. You’ll get the full story of Gone and End Records and the finale of George Goldner on this week’s “Juke In The Back.” LISTEN BELOW