Podcasts about CTO

Explore the intersection of high-level technology leadership and profound personal resilience with Bryan Muehlberger, CTO at The Health Institute and Founder of Lumiyo & CXO/RISE. Drawing from his expansive career driving innovation at iconic brands like Red Bull and Beachbody, Bryan shares practical wisdom on leading with heart, mentoring new executives, and tapping into creative problem-solving. He also opens up about navigating unimaginable tragedy, detailing how grief, faith, and human connection transformed his perspective on leadership and community impact.Guest Links:Bryan's LinkedInLumiyoThe Health InstituteCXO/RISECredits: Host: Lisa Nichols, Executive Producer: Jenny Heal, Marketing Support: Joe Szynkowski & Dylan Price, Podcast Engineer: Portside MediaSomething Extra with Lisa Nichols

Four days after Anthropic released Claude for Financial Advisors, it was the main topic on our AI mastermind call. I get why. Asking one tool to pull your appointments, your Schwab balances, and your Orion data, and having it all ready before your first meeting, is hard to ignore.But there's a gap between a tool being available and a tool being ready to use in your firm.Most advisors are thinking about what this can do. Far fewer are thinking about where client data goes once it starts moving between systems, who is reviewing what the AI produces, or whether their compliance program holds up when a regulator asks how the firm uses AI.So I sat down with Triad's CTO, Quin Kilgore, and our Chief Compliance Officer, Shannon McGinnis, for an early take. Quin brings the tech stack side. Shannon brings the side that keeps you out of trouble. I play the advisor who wants to turn it on today.If you've been tempted to connect a new AI tool to your client data, or you're not sure your firm has an AI policy at all, listen to this before you do.3 Insights From This Week's Episode…#1.) Treat AI Like A New EmployeeShannon shares a simple framework for thinking about AI inside an advisory firm. Just as you wouldn't hire someone, give them access to every system, and walk away, bringing AI deeper into your business introduces questions around policies, training, access, oversight, and ongoing supervision.#2). Archiving Your AI Isn't EnoughIt's easy to assume that if your AI tool keeps logs, you're covered. Shannon explains why the logs are only the starting point, and what regulators expect you to do with them.#3.) The Weak Link Between Your Trusted ToolsSchwab and Orion grew up in a regulated industry. Claude didn't. We look at what changes when a newer, general-purpose tool gets connected to the systems holding your clients' data, and the question most advisors aren't asking about the connection itself.FREE GIVEAWAYGet Your Free PDF of Our AI Implementation Checklist by Visiting: https://bradleyjohnson.com/188-ai-implementation-guide/SPONSORED BY BELAYIf you're an advisor and you're still scheduling your own appointments, sending your own follow-up emails, or dealing with other tasks keeping you from bringing on other clients, you're the bottleneck. BELAY helps busy leaders find world-class Virtual Assistants who can take tasks off their plate, protect their time, and help them stay focused on the work that actually moves the business forward. Learn more about BELAY and find the right assistant for your business here: http://belaysolutions.com/dbdlSHOW NOTEShttps://bradleyjohnson.com/188FOLLOW BRAD JOHNSON ON SOCIALXInstagramLinkedInFOLLOW DBDL ON SOCIAL:YouTubeTwitterInstagramLinkedInFacebookDISCLOSURE DBDL podcast episode conversations are intended to provide financial advisors with ideas, strategies, concepts and tools that could be incorporated into their business and their life. No statements made in the episode are offered as, and shall not constitute financial, investment, tax or legal advice. Financial professionals are responsible for ensuring implementation of anything discussed related to business is done so in accordance with any and all regulatory, compliance responsibilities and obligations. The Triad member statements reflect their own experience which may not be representative of all Triad Member experiences, and their appearances were not paid for. Triad Wealth Partners, LLC is an SEC Registered Investment Adviser. Please visit Triadwealthpartners.com for more information. Triad Wealth Partners, LLC and Triad Partners, LLC are affiliated companies. TP10265950813See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Dave Mabe, systematic trading coach and former CTO of Trade Ideas, shares how he transitioned from discretionary trading to running 25 automated strategies, and the single biggest backtesting mistake most traders make. Full episode + show notes: https://tradingnut.com/dave-mabe/ Key moments - [05:44] Having a full-time job and trading on the side forces you to focus on a small window, preventing you from strategy-hopping. - [11:53] Discretionary traders can look at situations and figure out when to 'swing the bat hard,' but it is a very difficult skill to master. - [14:30] My very first backtest of my discretionary strategy actually performed better than my manual execution. - [18:12] The only thing harder than coming up with your first profitable trading strategy is coming up with your second. - [19:08] The biggest mistake in backtesting is trying to shoot for the final version of a strategy on your very first run. - [20:01] Your first backtest should have no filtering; instead, add custom columns for every indicator and optimize using that dataset. - [27:44] Scale wide by adding more automated trades across smaller timeframes, rather than just increasing your position size. - [30:54] A good backtesting software must be able to scan the entire US equities market and allow custom columns. - [39:42] When using Claude to write backtest code, give it a specific chart setup where a trade should occur to let it self-verify. Watch: https://www.youtube.com/watch?v=FJIx4aNF49o --- Sponsors BlackBull Markets - 100% deposit bonus & free TradingView. NZ-regulated broker. https://tradingnut.com/go/blackbull/ FTMO - Buy a challenge and we'll help you pass it. https://tradingnut.com/go/ftmo/ --- More from Trading Nut Strategy Vault - Free guest setups organised by method. https://tradingnut.com/go/vault/ TraderZen - Trade control unlocked. Finally stick to your plan while you trade. https://tradingnut.com/go/traderzen/ Dynamic Traders Club - Free 7-minute rule strategy — mechanical intraday setup. https://tradingnut.com/go/dtc/ --- Educational only - not financial advice. Risk: https://tradingnut.com/disclaimers/ Some links are affiliate partnerships.

Every retailer wants AI agents, and very few have data an agent can use. Abhi Sachdeva, co-founder and CTO of EKYAM, has spent 18 of his 24 years in technology inside retail, running tech at 1-800-Flowers, Tory Burch and QVC. He joins Rick to explain what data readiness requires and where retailers keep getting it wrong.A data lake stores rows. It doesn't tell an agent what those rows mean or which system to trust, so the agent picks up the first answer it finds and presents it with full confidence. Ask a retailer for its customer count or gross margin and different teams will often give different numbers. People know which dashboard to believe. Agents don't. Abhi walks through what he thinks a semantic layer needs to include, and the piece he says most companies leave out, which is the history of how the data changed over time.Rick and Abhi also get into buy versus build. Abhi's view is that retailers should never build the control layer (governance, gateways, LLM monitoring, cost tracking) because it changes every time a new model ships. What retailers should own are their definitions and workflows. He thinks the big consulting firms are right to push master data cleanup and operating model work, and wrong when they spend 12 to 18 months building a bespoke agent data layer the client then has to maintain alone.On ownership, Abhi puts meaning with the business (merchandising, finance, supply chain), enforcement with technology, and a small product data team between them. By his estimate, about 5 in 100 retailers he talks to have that team in place. He also explains what happens when a marketing manager starts uploading customer CSVs to a chatbot on their own, and why the answer is governance rather than a ban.Abhi's company is built on the bet that data readiness is never finished. Weigh that accordingly. The last decade of retailers building and rebuilding their own checkouts suggests he has a point.This episode is sponsored by EKYAM.#AIAgents #Retail #Ecommerce

Kevin Novak is the CISO at Old National Bank. In this episode, he joins host Scott Schober and Dylan DeAnda, field CTO at Doppel, to discuss proactive resilience, including why cyber defense can no longer stop at the castle walls, how to strengthen external defense, and more. This episode of CISO Confidential is brought to you by Doppel. Learn more about our sponsor at https://doppel.com.

-Introduction What does a mid-sized brokerage have that it cannot actually use? A specialized book and the data behind it, and none of the underwriting, capital, filing, and admin machinery it would take to turn that book into a program it owns. The ten largest firms solved this by buying MGAs. Cole Riccardi built Authentic to rent that machinery to everyone else, and nine of the top ten retail brokerages now work with him on digital programs anyway. In this conversation he explains why he refuses to call Authentic a software company, what he got wrong in the first two years, and why he went after Main Street before he went after the hard market. Guest Bio Cole Riccardi founded Authentic in 2022 after five years at Aquiline Capital Partners, where he invested across the insurance value chain and became convinced that nobody was building affinity insurance programs at scale. He built the company with a co-founder and COO who spent fifteen years as an actuary at AmTrust and a CTO who came out of six years at Amazon, and the platform was architected as multi-carrier, multi-tenant, and multi-product from the first day. Authentic runs out of New York and Dayton, Ohio, is backed by FirstMark Capital, and now sits behind six carriers with programs spanning Main Street, energy, and inland marine. Key Topics -Programs should not be limited to the top ten - The largest brokerages own their own MGAs. Riccardi's argument is that everyone else already has the book and the data, and only lacks the machinery, which is a rentable problem. -An MGA that gives its software away - Riccardi is emphatic that Authentic never wanted to sell software and makes its money in the transaction flow. The technology is what makes it a better trading partner, not the product. -Affinity is product creation, not just distribution - For PushPress, Authentic went through the entire book and built four or five custom class codes that mapped to CrossFit gyms and similar facilities, rather than rating them off generic fitness codes. -Main Street first, on purpose - Rather than chase hard-market classes where premium comes fast, Authentic built a sizable Main Street book to prove that the configured programs were what moved business over, not market scarcity. -Same infrastructure, new practices - A Chubb energy underwriter joined with a thesis about independent power producers being underserved on casualty. That program launched in May, and an inland marine practice was announced the day of this recording. -The captive he overbuilt - Riccardi is direct that as a first-time founder he overcomplicated the early years by building custom captive cells on the back end. The captive still exists and takes smaller slices, but it is no longer the pitch. -Fast nos and the right to win - Authentic has no trucking underwriter and no capacity partner who will raise a hand for it, so brokers asking about long-haul trucking get turned down quickly rather than strung along. Notable Quotes "We never wanted to sell software. I always wanted to be in the transaction flow. We make money as an MGA today. That's always how we will make money." "Nine of the top ten work with us today on digital programs." "We've actually gone after Main Street to start and built up a pretty sizable book there. We wanted to do that so we could prove to ourselves that our technology mousetrap mattered and we weren't just writing business because it was hard to find markets elsewhere." "Can we challenge the status quo that MGAs need twelve points to operate? If you're built on modern systems, I don't think you do." Resources Guest: Authentic: https://authenticinsurance.com/ Authentic, For Brokers: https://authenticinsurance.com/brokers Cole Riccardi on LinkedIn: https://www.linkedin.com/in/cole-riccardi-63a243127/ Host & Organization: Joshua R. Hollander on LinkedIn: https://www.linkedin.com/in/joshuarhollander/ Horton International (USA): https://www.horton-usa.com/ Insurtech Leadership Podcast (LinkedIn Showcase): https://www.linkedin.com/showcase/insurtech-leadership-show Subscribe & Review If you enjoyed this episode, subscribe on your favorite platform and leave a review. The Insurtech Leadership Podcast is available on YouTube, Podbean, Apple Podcasts, and Spotify.

This week, Alex and Emily are joined by Keoni Mahelona of Te Hiku Media to discuss Indigenous language technology. They discuss problems that arise when Big Tech tries to bring its "solutions" into this space, and best practices for research and development that center the expertise, cultural norms and goals of Indigenous communities.Keoni Mahelona is the CTO of Te Hiku Media and a Native Hawaiian from the island of Kaua'i. He's a driving force behind the development of sovereign digital technologies that aim to protect and promote Indigenous languages, knowledge, culture, and now democracy.References:Microsoft/IDB paper: "The Performance of Artificial Intelligence in the Use of Indigenous American Languages"Also referenced:Kaitiakitanga LicenseJoanna Radin paper on Indigenous data collectionFresh AI Hell:"Apple Watch's new always-on don't-call-it-transcription feature"Nevada County prosecutors use "AI" and insert errors into criminal casesUW Grand Rounds: "If AI Can Write the Note, Who Does the Thinking?"The "Pro-Human Assembly"Check out future streams on Twitch. Meanwhile, send us any AI Hell you see.Find our book The AI Con here, and MAIHT3k merch here.Subscribe to our newsletter via Buttondown.Follow us!EmilyBluesky: emilymbender.bsky.socialMastodon: dair-community.social/@EmilyMBenderAlexBluesky: alexhanna.bsky.socialMastodon: dair-community.social/@alexTwitter: @alexhannaMusic by Toby Menon.Artwork by Naomi Pleasure-Park. Production by Ozzy Llinas Goodman.

If robotic automation is becoming inevitable in utility-scale solar, where does that thesis run into the realities of actually building a power plant?That's the question behind today's episode.We start with Nick de Vries, CTO of Silicon Ranch, to ask what an owner actually wants from automation. Nick's answer sets the standard: the technology matters only if it improves safety, quality, repeatability, and the performance of an asset Silicon Ranch expects to own for decades.Then Matt Campbell, CEO of Terabase, takes us into the field, where his team is applying manufacturing principles to solar construction. That's where an important constraint emerges. Robots can move and place modules remarkably well, but some of the work that experienced crews make look simple becomes much harder when every action has to be repeatable.One deceptively small interface keeps surfacing: the point where the module actually attaches to the structure.That leads Nico to Frédéric Laure, VP of Diversification at ARaymond North America, to understand why fastening, tolerances, torque, and component design become different problems once machines enter the workflow.The bigger lesson isn't simply about robots or fasteners. It's about constructability.If what we are seeking is manufacturing-level throughput in the field, more of the variability may need to be engineered out long before construction begins.Listen in to hear where solar automation is working, where the real-world constraints still live, and what the industry may need to redesign next.Paid partnership: This Tactical Tuesday was produced with support from ARaymond.Are there other technologies you've scouted on the frontlines of the Clean Energy Revolution that you think we should be covering here on SunCast? Hit us up - team@suncast.me with your feedback & recommendations.You'll find more resources and learn about SunCast's guest(s), recommendations, book links, and more than 850 other founder stories and startup advice at www.mysuncast.com.You can learn more about partnering with SunCast here: https://mysuncast.com/sponsorsYou can connect with me, Nico Johnson, on:Twitter - https://www.twitter.com/nicomeoLinkedIn - https://www.linkedin.com/in/nickalusSubscribe to Valence, our weekly Linkedin Newsletter, and learn the elements of compelling storytelling: https://www.linkedin.com/newsletters/valence-content-that-connects-7145928995363049472/(00:00) Rethinking Utility-Scale Solar Construction and Robotics(02:37) An Owner's Perspective on Jobsite Automation(05:21) Lessons from 20 Years of Module Automation(07:20) Targeting True Limiting Factors in Solar Design(09:41) Installation Speed vs Long-Term Fastener Reliability(13:39) Eliminating Field Variability Through Upstream Engineering(15:39) Overcoming Construction Labor Shortages with Field Factories(18:40) Standardizing Mega-Projects Through Digitalization(20:33) Solving the Module Attachment Automation Bottleneck(23:45) Integrating the Supply Chain for Robotic Assembly(28:46) Redesigning Fasteners for Automated Construction(32:54) Clip-Based Attachment Systems vs Threaded Bolts(36:37) Designing for Assembly Across the Clean Energy Ecosystem(40:22) Key Takeaways: Repeatability, Interfaces, and Upstream Fixes

Did you like this episode? Dislike it?

Ali Dasdan, CTO @ Dropbox, joins the pod to share insights on the company's large-scale AI adoption and research as a foundation of successful engineering leadership. First, Ali shares insights on why he continues publishing research despite his role as a CTO and how research / curiosity can drive better trust within your organization. He shares about how his research ultimately guided decision making regarding redrawing Dropbox's entire architecture and the importance of creating a record of the company's progression as the rearchitecture occurred. The bulk of the conversation centers around how Dropbox is adopting AI, including tools like Nova and Dash, within both EPD and non-EPD departments. ABOUT ALI DASDANAli Dasdan is a C-suite technology executive with 25+ years building, scaling, and leading global engineering, product, and technical-operations organizations of up to ~1,000 people across a dozen industries, in public companies, high-growth startups, and enterprises in Silicon Valley and London.Today he is CTO of Dropbox, serving over 700 millions of users on an exabyte-scale platform with $2.5B+ revenue, where he also leads the AI strategy and AI enablement. Previously EVP and CTO of ZoomInfo and VP of Engineering at Atlassian (Confluence Cloud, Trello, Jira Work Management). As CTO (at Turn, Vida, Poynt, ZoomInfo, and Dropbox) he defines and owns company-wide technology strategy; in divisional roles (eBay, Atlassian, Tesco, Yahoo) he led engineering for revenue-critical parts of the business. SHOW NOTES:Why Ali continues to publish research as a CTO (1:51)How research results can alter decision making as an eng leader (3:30)The connection between research & organizational trust (4:29)Redrawing the org's architecture to learn more about it (6:00)What kinds of info help eng leaders better understand the business / technology (8:12)Building a record of the org's progress & development (10:13)Where Dropbox is at in their AI adoption journey (11:58)AI @ Dropbox in EPD vs. non-EPD functions (13:36)Frameworks for implementing AI at the highest level (15:23)The engagement / ownership model @ Dropbox (16:58)Insights on Nova, Dropbox's code review AI workflow (18:59)Incorporating human-driven feedback into the AI loop (21:12)What the immediate future looks like for automated developer tools (23:54)How AI has directly impacted productivity @ Dropbox (27:15)Addressing bottlenecks related to AI adoption (30:35)Using AI infrastructure to assist in non-EPD functions (32:33)Dash as a contextual AI platform (35:39)Ali's experience with specialized models vs. foundation models (38:10)Rapid fire questions (39:48) This episode wouldn't have been possible without the help of our incredible production team:Patrick Gallagher - Producer & Co-HostJerry Li - Co-HostNoah Olberding - Associate Producer, Audio & Video Editor https://www.linkedin.com/in/noah-olberding/Dan Overheim - Audio Engineer, Dan's also an avid 3D printer - https://www.bnd3d.com/Ellie Coggins Angus - Copywriter, Check out her other work at https://elliecoggins.com/about/ Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

Bryan Duarte lost his sight at 18 in a motorcycle accident. He went on to earn a PhD in computer science studying one question: can touch carry what sight usually carries? Today he's co-founder and CTO of HapWare, building ALEYE. It's a wristband paired with a camera that reads facial expressions, gestures and body language. It plays each one as a pattern of vibrations on the wearer's wrist. In Part 1, Bryan talks about what the wristband feels like after a few hours and why learning the patterns is like learning a new ringtone. He also explains why HapWare stopped trying to tell a slight smile from a big one. Audio Description runs into the same question. A script can say "She smiles." The performance decides whether that smile is flirty, sarcastic or barely there. ALEYE tells the wearer someone smiled and leaves the meaning to them.

Terry sits down with Cesile Johnson, CTO of Core Bank, for a conversation about what it actually takes to modernize a bank from the inside — the legacy systems nobody wants to touch, the guardrails that have to exist before AI gets anywhere near customer data, and the reality of building a career in a room where she was often the only woman in it.Cesile walks through how her team took delivery velocity on one initiative from 5% to 85%, what real AI governance looks like inside a regulated bank versus the "vibe coding" hype, and why the trades conversation (welders and electricians, not just four-year degrees) matters as much to the future workforce as anything happening in tech. She's also direct about fraud, budget reality, crisis leadership, and what people consistently get wrong about what a CTO actually does.In this episode:What "modernizing a legacy system" really means when you can't just rip and replaceHow Core Bank built AI guardrails before deploying AI, not afterTaking one team's delivery velocity from 5% to 85%, and what actually changedWhy the trades — not just four-year degrees — are part of the talent pipeline conversationCrisis leadership and what accountability looks like when something breaksThe real budget math behind fighting fraud, including the scale of elder fraud losses nationallyBeing the only woman in the room, and what it took to get to CTOWhat people misunderstand about the CTO role, and how it's evolvedAdvice for the next generation of leaders coming up in banking and techChapters: 0:00 Intro 0:22 Technology Is the Bank 14:11 Legacy Systems and Real Modernization 23:38 Guardrails for AI in Banking 30:55 From 5% to 85%: AI, Velocity, and Vibe Coding 39:42 Trades, Degrees, and Vanishing Entry Jobs 48:14 Crisis Leadership and Accountability 1:01:25 AI in Banking: Fraud and the Budget Reality 1:10:51 The Only Woman in the Room 1:20:16 What People Get Wrong About the CTO Role 1:28:58 Legacy, Team Culture, and Advice for the Next GenNew episodes of the Kill Chain Podcast — like, comment, and subscribe. Learn more at fleetdefender.comWant to learn more about securing your fleets, platforms, or mission critical systems? Contact us at FleetDefender.com.

En este episodio de Road to CTO, conversamos con Emili Planas, que ha sido CTO en Mediapro durante muchísimos años, en uno de los grupos audiovisuales e infraestructuras de retransmisión más importantes de Europa y del mundo.Emili cuenta con una trayectoria atípica y fascinante: pasó de estudiar cine y trabajar como técnico en unidades móviles a liderar la arquitectura tecnológica y las operaciones en directo de eventos deportivos masivos como LaLiga. Bajo su dirección, MediaPro se convirtió en pionera a nivel mundial en implantar la producción remota en competiciones Tier 1, cambiando cómo se produce y distribuye el contenido audiovisual.A lo largo de la charla analizamos el salto histórico de los equipos físicos (hardware dedicado) al software y la nube, la gestión de latencias críticas en retransmisiones y apuestas, la toma de decisiones sin miedo al error y cómo aplicar la narrativa cinematográfica a la ingeniería de sistemas. Un episodio del que saldrás con más, te lo aseguramos.Support the show

A little trick I play on my clients when they think I'm not paying attention often leads us to uncover repeating mistakes and can help you avoid repeating yours. More in this week's episode.Grab a copy of my books, Capitalizing Your Technology and  The Tech Executive Operating System.Subscribe to the best newsletter for tech executives.For any questions or comments, reach out to me directly: aviv@avivbenyosef.com

What happens when the technology designed to make work easier becomes too complex, restrictive or difficult to manage? In this episode of Softcat's Explain IT podcast, Helen Gidney, Softcat's Head of Architecture, is joined by Scott Manchester, Chief Product and Technology Officer at Nerdio; James Millington, Field CTO for Healthcare, EMEA at IGEL; and Mark Forster, CTO at Softcat.Together, they explore how organisations can create digital workspaces that are secure, manageable and genuinely productive for employees. The conversation covers the changing role of BYOD, the rise of shadow IT and BYOAI, and the growing expectations employees bring from consumer technology into the workplace.The panel discusses the competing pressures facing IT leaders, including investment in artificial intelligence, cyber security, data management and endpoint modernisation. They consider how virtualisation, centralised data and immutable endpoints can reduce risk while still delivering the seamless experience users expect.The conversation highlights that when employees cannot work effectively, the cost extends beyond IT support to lost productivity, talent attrition and disrupted business relationships.The episode covers practical advice for organisations navigating a fast-changing digital workplace: find a trusted advisor, understand your users and consider security, cost, complexity and experience together.The Explain It podcast is released monthly. Subscribe or follow in your app so you don't miss the next one, and if this episode was useful, a review helps other IT leaders find the show.This podcast is produced by The Podcast Coach. Hosted on Acast. See acast.com/privacy for more information.

Ready to level up your agency in person? Get your LevelUp 2026 tickets here

Ce mardi 29 septembre, Didier Girard, CTO et directeur général de SFEIR, et Erik Campanini, associé chez Alixio Group, sont revenus sur le nouvel agent IA de Meta et les moyens accordés au déploiement de l'IA dans le budget de 2027, dans l'émission Tech&Co Business présentée par Frédéric Simottel. Tech&Co Business est à voir ou écouter le mardi sur BFM Business.

Mardi 29 septembre, Frédéric Simottel a reçu Dr Thomas Di Giacomo, directeur technologie et produit chez SUSE, Amaury Delplancq, directeur général EMEA chez Dataiku, Alix Mirshams, directeur général d'Opteamis, Vincent Desclaux, directeur général Europe chez Palo IT, Didier Girard, CTO et directeur général de SFEIR, ainsi qu'Erik Campanini, associé chez Alixio Group, dans l'émission Tech&Co Business sur BFM Business. Retrouvez l'émission le samedi et réécoutez la en podcast.

In this episode, Dave and Jamison answer these questions: Hi Dave and Jamison, I've been a Java backend dev for over 10 years. After getting laid off from a job I loved, I spent 3 tough months job hunting. I eventually landed a full-stack role at a non-tech company where I am the entire front-end department. The team is a mess: A CTO who hasn't coded in years and calls me directly to assign tasks, all labeled “highest priority.” When I asked the CTO and CEO if we had QA, they literally laughed at me. Prod releases are “tested” by the CTO before telling me to push immediately. I'm at a crossroads: Stick it out for 6–12 months to gain AI experience, then quit? Try to move into management, and rebuild this mess into a functioning engineering organization. What should I do? Hey, Dave and Jamison. I'm a non-native Japanese speaker working in Tokyo with about 6 years of experience in backend(3 years in Tokyo) and haven't used English in work. I started listening to this podcast for learning English. And I soon found this podcast is very funny and I found a lot of problems actually exist around the world, like AI hype, why people hate HR, low-performing team members, useless product managers, etc. For some reason, I want to leave Japan and find a job using English. I think the best approach is to find an English job in Tokyo first. And then transfer to an English-speaking country. I want to know as a non-native English-speaking engineer, what English level I need to get a senior engineer job using English. What would you consider if you were the manager who is interviewing a non-native English speaker? Thank you guys, I love this podcast, I can learn English, get some answers to the questions I'm confused about too, and laugh every week.

Everyone's arguing about whether AI will replace 90% of your job. Steve Orrin, federal CTO at Intel, says that's the wrong number to even be worried about. We discuss how a $25 million deepfake of a bank's own CFO exposed a whole new category of cyberattack, why sprinkling "AI pixie dust" on a problem doesn't actually solve anything, and why, thirty years into being a CTO, his best advice still comes down to surrounding yourself with people smarter than you. All of this right here, right now, on the Modern CTO Podcast!

Zack Korman on why the AI safety debate has been captured by an extinction narrative, and why the incident everyone is calling a watershed moment looks a lot more like a preventable security failure.In this episode I sit down with Zack Korman, co-founder of Embroidery, where he uses AI to monitor AI agents, and former CTO of the cybersecurity company Pistachio. I first came across Zack on X during the controversy over fake SOC 2 reports, and he has since become one of the sharpest critics of how AI safety is being framed, funded, and investigated.We get into the effective altruism roots of the existential risk movement, why he argues METR's review of the Hugging Face incident was not independent oversight, and what an actual incident response firm would have done differently. Zack makes the case that alignment is one control among many, that there is no "safe" path in AI, only trade-offs, and that the real risk for most organizations is enterprise environments that were never ready for agents in the first place.In this episode:● How building an AI insider threat product pulled a developer and CTO into the cybersecurity community, and into picking fights on X● Effective altruism, longtermism, and how a focus on preventing extinction came to dominate AI safety● Why "just be careful" misses the point when every path involves trading one risk for another● Dario Amodei's Pacing the Frontier, independent auditors, and why Zack calls bringing in METR "bring your friend to work day"● The funding and relationships connecting METR, Redwood Research, Coefficient Giving, and the labs● True believers versus IPO hype, and why genuine belief does not make someone right● Why many AI doomers believed we were all going to die before they ever learned about computers● The Hugging Face investigation, context drop in AI-analyzed transcripts, and treating knowable facts as unknowable● What Unit 42 or Mandiant would have demanded before putting their name on the report● Eight layers of failure, from a single package proxy at egress to missing monitoring, classifiers, and kill switches● Alignment failure versus containment failure, and why alignment belongs inside defense in depth● The real risks: threat actors misusing AI, enterprise agent deployments, and new attack chains● Why security's risk-averse, laggard culture may be its biggest vulnerabilityChapters:0:00 Intro0:54 Zack's Background1:43 What Turned a CTO Into an AI Safety Critic3:31 Effective Altruism and the Extinction Narrative5:53 No Safe Path, Only Trade-offs6:40 METR, Redwood, and Bring Your Friend to Work Day10:23 True Believers, Hype, or Both12:49 How Doomers Find Their Way Into AI14:39 Taking AI Risk Seriously When It Is Ideological16:56 What an IR Firm Would Have Asked21:10 Eight Layers of Failure in the Hugging Face Incident22:49 Alignment Failure Versus Containment Failure24:32 Alignment as One Layer of Defense in Depth26:22 Why Enterprise Environments Are Not Ready for Agents29:10 Security's Laggard Culture29:19 ClosingConnect with Zack Korman:X: https://x.com/ZackKormanEmbroidery: https://embroidery.ioWebsite: https://zkorman.comResilient Cyber: https://www.resilientcyber.ioSubscribe for more conversations with security practitioners and leaders.#aisafety #aisecurity #agenticai #incidentresponse #effectivealtruism #cybersecurity

Interview with Christopher Crawley - The Value of SecOps This week, we talk to Chris Crawley about his new book, The Value of Cybersecurity Operations. Chris has been training teams on security operations for years, but found that students often struggled to justify the importance of secops training and even the need for secops in general. This book was written to more broadly arm practitioners with the information they need to communicate why security operations are necessary to leadership. The book is available in several forms: eBook, hardcover, softcover, and an audiobook read by Chris himself! You can pick all versions up here: https://shop.montance.com/collections/all and Security Weekly listeners get 50% off with code sw-50! The Evolving Exploitation of Trust with Josh Davies, Security Strategist at Fortra Fortra Intelligence and Research Experts (FIRE) is Fortra's emergent threat intelligence identity, created to unify research teams across multiple security disciplines and share intelligence with the wider threat intelligence community. FIRE co-ordinator and security strategist Josh Davies joins us to share insights from original FIRE research like Calphishing, Mirage2FA, RatPressto phishkit and heavily obfuscated campaigns that evade defences. While also digging into trends from big data analysis within phishing, fraud, social engineering and credential theft. Segment Resources: Art of Security Podcast Fortra Research This segment is sponsored by Fortra. Access FIRE Research here: https://securityweekly.com/fortrabh Why Agentic AI Security Requires a Defense-In-Depth Strategy with Kaushik Shanadi, CTO and Co-Founder of Helmet Security Many organizations are approaching agentic AI security by stitching together individual controls, assuming that identity, network, endpoint, or application security alone is enough. Each address only one part of the problem. As AI agents become more autonomous, move across systems, and take actions on behalf of users, organizations need a true defense-in-depth strategy that combines every layer of the security stack. Kaushik can discuss how each security layer provides a different piece of the puzzle and how individually, none of the controls are sufficient. But together, they create the layered governance needed to securely deploy agentic AI. For more information about Helmet Security, please visit https://securityweekly.com/helmetbh Above Security on Why Legacy Tools Don't Cut It for Modern Insider Risk with Tricia Howard, Head of Marketing at Above Security As organizations rush to adopt agentic AI, agents are often given broad access to critical business systems and sensitive data. Legacy insider risk management tools weren't built for this, and organizations have spent years on insider risk posturing and investments that leave security teams inundated with false positives while the real risk slips through undetected. In this segment, Tricia Howard draws on her decade in cybersecurity, including years watching user and entity behavior analytics (UEBA) overpromise and underdeliver, to unpack why understanding human – and now AI – intent is the missing piece. Segment Resources: Blog Post: Redefining Insider Threat Blog Post: What I wanted UEBA to be, Years Ago To learn more about Above Security and insider risk management in the era of agentic AI, visit: https://securityweekly.com/abovebh The Identity Crisis Your Security Team Didn't See Coming: Governing AI Agents with Darren Guccione, CEO and Co-Founder of Keeper Security AI agents outnumber human identities in enterprise environments, and traditional security software was never built to govern them. 89% of cybersecurity decision-makers recently surveyed by Keeper Security said that AI adoption has made identity management harder. In this Black Hat USA conversation with Cyber Risk TV, Darren Guccione, CEO and Co-Founder of Keeper Security, shares a practical framework for extending identity governance to non-human identities and AI agents before the access gap becomes a breach. Segment Resources: Privileged Access Management Identity Security at Machine Speed To learn more about Keeper Security, visit: https://securityweekly.com/keeperbh Visit https://www.securityweekly.com/esw for all the latest episodes! Show Notes: https://securityweekly.com/esw-478

Another Microsoft SharePoint flaw now exploited Details and doubts emerge regarding OpenAI hack of Australian Health portal Kiteworks urges customers to stop using platform Get the show notes here: https://cisoseries.com/cybersecurity-news-new-sharepoint-exploit-australian-openai-hack-developments-kiteworks-urges-stoppage/ Huge thanks to our episode sponsor, Intezer Today's tip: pull your low-severity alerts and count how many got opened. That's where attackers hide. Intezer investigates every alert, boring ones included, in under a minute. MGM Resorts' CTO has talked about nation-state threats turning up right there. Intezer. The AI SOC trusted by Salesforce, MGM Resorts and Nvidia. See it yourself at intezer.com/headlines.

Interview with Christopher Crawley - The Value of SecOps This week, we talk to Chris Crawley about his new book, The Value of Cybersecurity Operations. Chris has been training teams on security operations for years, but found that students often struggled to justify the importance of secops training and even the need for secops in general. This book was written to more broadly arm practitioners with the information they need to communicate why security operations are necessary to leadership. The book is available in several forms: eBook, hardcover, softcover, and an audiobook read by Chris himself! You can pick all versions up here: https://shop.montance.com/collections/all and Security Weekly listeners get 50% off with code sw-50! The Evolving Exploitation of Trust with Josh Davies, Security Strategist at Fortra Fortra Intelligence and Research Experts (FIRE) is Fortra's emergent threat intelligence identity, created to unify research teams across multiple security disciplines and share intelligence with the wider threat intelligence community. FIRE co-ordinator and security strategist Josh Davies joins us to share insights from original FIRE research like Calphishing, Mirage2FA, RatPressto phishkit and heavily obfuscated campaigns that evade defences. While also digging into trends from big data analysis within phishing, fraud, social engineering and credential theft. Segment Resources: Art of Security Podcast Fortra Research This segment is sponsored by Fortra. Access FIRE Research here: https://securityweekly.com/fortrabh Why Agentic AI Security Requires a Defense-In-Depth Strategy with Kaushik Shanadi, CTO and Co-Founder of Helmet Security Many organizations are approaching agentic AI security by stitching together individual controls, assuming that identity, network, endpoint, or application security alone is enough. Each address only one part of the problem. As AI agents become more autonomous, move across systems, and take actions on behalf of users, organizations need a true defense-in-depth strategy that combines every layer of the security stack. Kaushik can discuss how each security layer provides a different piece of the puzzle and how individually, none of the controls are sufficient. But together, they create the layered governance needed to securely deploy agentic AI. For more information about Helmet Security, please visit https://securityweekly.com/helmetbh Above Security on Why Legacy Tools Don't Cut It for Modern Insider Risk with Tricia Howard, Head of Marketing at Above Security As organizations rush to adopt agentic AI, agents are often given broad access to critical business systems and sensitive data. Legacy insider risk management tools weren't built for this, and organizations have spent years on insider risk posturing and investments that leave security teams inundated with false positives while the real risk slips through undetected. In this segment, Tricia Howard draws on her decade in cybersecurity, including years watching user and entity behavior analytics (UEBA) overpromise and underdeliver, to unpack why understanding human – and now AI – intent is the missing piece. Segment Resources: Blog Post: Redefining Insider Threat Blog Post: What I wanted UEBA to be, Years Ago To learn more about Above Security and insider risk management in the era of agentic AI, visit: https://securityweekly.com/abovebh The Identity Crisis Your Security Team Didn't See Coming: Governing AI Agents with Darren Guccione, CEO and Co-Founder of Keeper Security AI agents outnumber human identities in enterprise environments, and traditional security software was never built to govern them. 89% of cybersecurity decision-makers recently surveyed by Keeper Security said that AI adoption has made identity management harder. In this Black Hat USA conversation with Cyber Risk TV, Darren Guccione, CEO and Co-Founder of Keeper Security, shares a practical framework for extending identity governance to non-human identities and AI agents before the access gap becomes a breach. Segment Resources: Privileged Access Management Identity Security at Machine Speed To learn more about Keeper Security, visit: https://securityweekly.com/keeperbh Show Notes: https://securityweekly.com/esw-478

This and all episodes at: https://aiandyou.net/ . In all the news about and focus on startups like OpenAI and Anthropic, there's a long-established player that doesn't get as many soundbites but which deserves our attention: IBM. It's important to understand what their current strategy and thinking is. And I have just the person to do that: Ruchir Puri, the Chief Scientist of IBM Research, and Vice-President of IBM Corporate Technology and Technical Community. He led IBM Watson as its CTO and Chief Architect from 2016-19 and is a Fellow of the IEEE, and has been an ACM Distinguished Speaker, an IEEE Distinguished Lecturer, and was honored with John Von-Neumann Chair at Institute of Discrete Mathematics at Bonn University, Germany. Notably from our point of view, he is an inventor of over 70 US patents and has authored over 120 scientific papers, which lands him squarely in the technical perspective that we value so much on this show.  We talk about the continuing importance of the foundation of computer science, which is algorithms, the tension between the deterministic nature of algorithms and the more touchy-feely variability of generative AI interfaces, what that means for reliability and verifiability, how best to think about Gen AI, the history of IBM's big AI projects like Watson and Deep Blue, and IBM's ambitions in quantum computing and the role of AI in that. All this plus our usual look at today's AI headlines! Transcript and URLs referenced at HumanCusp Blog.        

Analysts Don Kellogg and Roger Entner are joined by Yago Tenorio, SVP and CTO at Verizon, to discuss the coming of 6G and the future of networks, as well as exciting real world use cases happening right now.00:00 Episode intro 00:24 6G Innovation Forum overview 01:21 Lessons learned from 5G 03:59 What players are needed for 6G? 05:24 Actual test cases for 6G 07:34 Spectrum issues and ISAC 08:48 Future network infrastructure and monetization 12:44 6G Innovation Forum developments 13:20 Episode wrap-upTags: telecom, telecommunications, wireless, prepaid, postpaid, cellular phone, Don Kellogg, Roger Entner, Yago Tenorio, Verizon, 6G, 5G, FWA, drones, crowd detection, AI, LLMs, wearables, ISAC, network, spectrum, Google, devices, data centers, latency

Interview with Christopher Crawley - The Value of SecOps This week, we talk to Chris Crawley about his new book, The Value of Cybersecurity Operations. Chris has been training teams on security operations for years, but found that students often struggled to justify the importance of secops training and even the need for secops in general. This book was written to more broadly arm practitioners with the information they need to communicate why security operations are necessary to leadership. The book is available in several forms: eBook, hardcover, softcover, and an audiobook read by Chris himself! You can pick all versions up here: https://shop.montance.com/collections/all and Security Weekly listeners get 50% off with code sw-50! The Evolving Exploitation of Trust with Josh Davies, Security Strategist at Fortra Fortra Intelligence and Research Experts (FIRE) is Fortra's emergent threat intelligence identity, created to unify research teams across multiple security disciplines and share intelligence with the wider threat intelligence community. FIRE co-ordinator and security strategist Josh Davies joins us to share insights from original FIRE research like Calphishing, Mirage2FA, RatPressto phishkit and heavily obfuscated campaigns that evade defences. While also digging into trends from big data analysis within phishing, fraud, social engineering and credential theft. Segment Resources: Art of Security Podcast Fortra Research This segment is sponsored by Fortra. Access FIRE Research here: https://securityweekly.com/fortrabh Why Agentic AI Security Requires a Defense-In-Depth Strategy with Kaushik Shanadi, CTO and Co-Founder of Helmet Security Many organizations are approaching agentic AI security by stitching together individual controls, assuming that identity, network, endpoint, or application security alone is enough. Each address only one part of the problem. As AI agents become more autonomous, move across systems, and take actions on behalf of users, organizations need a true defense-in-depth strategy that combines every layer of the security stack. Kaushik can discuss how each security layer provides a different piece of the puzzle and how individually, none of the controls are sufficient. But together, they create the layered governance needed to securely deploy agentic AI. For more information about Helmet Security, please visit https://securityweekly.com/helmetbh Above Security on Why Legacy Tools Don't Cut It for Modern Insider Risk with Tricia Howard, Head of Marketing at Above Security As organizations rush to adopt agentic AI, agents are often given broad access to critical business systems and sensitive data. Legacy insider risk management tools weren't built for this, and organizations have spent years on insider risk posturing and investments that leave security teams inundated with false positives while the real risk slips through undetected. In this segment, Tricia Howard draws on her decade in cybersecurity, including years watching user and entity behavior analytics (UEBA) overpromise and underdeliver, to unpack why understanding human – and now AI – intent is the missing piece. Segment Resources: Blog Post: Redefining Insider Threat Blog Post: What I wanted UEBA to be, Years Ago To learn more about Above Security and insider risk management in the era of agentic AI, visit: https://securityweekly.com/abovebh The Identity Crisis Your Security Team Didn't See Coming: Governing AI Agents with Darren Guccione, CEO and Co-Founder of Keeper Security AI agents outnumber human identities in enterprise environments, and traditional security software was never built to govern them. 89% of cybersecurity decision-makers recently surveyed by Keeper Security said that AI adoption has made identity management harder. In this Black Hat USA conversation with Cyber Risk TV, Darren Guccione, CEO and Co-Founder of Keeper Security, shares a practical framework for extending identity governance to non-human identities and AI agents before the access gap becomes a breach. Segment Resources: Privileged Access Management Identity Security at Machine Speed To learn more about Keeper Security, visit: https://securityweekly.com/keeperbh Show Notes: https://securityweekly.com/esw-478

AI zjada Open Source? Koniec programistów, czas na makerów | Brew #74A może po prostu zmienia się sama definicja tworzenia software'u? O tym i nie tylko w nowym Brew.

AI is no longer just a productivity tool - it is becoming an operational risk, a security layer, and a new kind of business identity. Ben Wilcox, CISO and CTO at ProArc, explains why the old rules for governance, auditing, and trust are breaking fast, and what leaders need to do before agents start making real decisions on their own. Ben and I unpack the shift from clicks to prompts, from deterministic software to probabilistic behavior, and from human-only workflows to environments where machine identities may outnumber people. If you are responsible for security, data, cloud costs, or AI strategy, this conversation shows why visibility and guardrails matter more than ever. You'll discover: Why data governance has suddenly become a board-level priority How agent behavior changes the way security, logging, and audit trails must work Why “every click is a waste” is reshaping product design and user experience The hidden cost of AI usage, token spend, and why sticker shock is coming Why some AI use cases should be scaled back, not expanded How to choose the right business use case before you pilot anything at scale Ben shares practical guidance for getting ahead of AI sprawl, including how to identify what AI is already in your environment, how to assign both business and technical ownership to an agent, and how to define the right boundaries before something breaks in production. He also breaks down why security has to enter the conversation earlier, or it becomes friction instead of an enabler. The bigger message: AI adoption is moving too fast for guesswork. If your organization wants to move quickly without creating chaos, you need standards, observability, and a clear model for trust before you scale. Essential listening for CISOs, CTOs, IT leaders, and anyone trying to turn AI from hype into durable business value. I hope you enjoy it!

Introduction  What happens when you build an MGA from scratch with today's AI capabilities, rather than trying to layer them onto legacy technology?  In this episode, Robin Merttens speaks with Theo Butt, CEO of Kinetic Insurance Services, Andy Roberts, CTO of Kinetic Insurance Services and Guillaume Bonnissent, CEO of Quotech, about Convex's new hybrid carrier-MGA platform, the technology being built behind it and what an AI-native underwriting operation could look like in practice.  Theo explains why Convex created Kinetic to extend into lines of business that are complementary to its existing underwriting appetite. The model is designed to give entrepreneurial underwriters some of the advantages of building their own MGA, while providing access to cornerstone capacity, technology, operations, claims and compliance infrastructure. The aim is simple: allow underwriters to spend more time building profitable portfolios and less time building the machinery around them.  Andy takes us inside Kinetic's technology strategy. Starting with a greenfield environment has given the team an opportunity to design around agentic AI from day one, with the ambition of automating as much non-revenue-generating work as possible. That could mean workflows running from the moment an email arrives through to quote generation, while keeping the underwriter in control of the decisions that require human judgement.  But building an AI-native insurance business is not simply about automation. Andy explains why explainability, auditability and human oversight have to be designed into agentic workflows from the beginning, rather than added later. Every action needs to be traceable, with controls over what agents can do and clear visibility into how and why work has been completed.  Guillaume explores what could come next. As large language models become the interface through which employees access information and complete tasks, he believes underwriters could increasingly start their day in their company's LLM rather than moving between email, underwriting systems and other applications. The technology to do this already exists, he argues, making adoption, governance and cost control increasingly important challenges.  The conversation also looks at one of the biggest advantages and risks of building on the technological frontier. Kinetic does not have the legacy infrastructure that can make deploying AI at scale difficult, but the pace of change means today's new technology can quickly become tomorrow's legacy. Building for flexibility and continuous change may therefore be just as important as choosing the right technology today.  In this episode you'll learn:  Why Convex chose a hybrid carrier-MGA model for Kinetic  How the model aims to give entrepreneurial underwriters more time to focus on underwriting  Why Convex intends to provide cornerstone capacity to Kinetic's underwriting cells  How Kinetic is building agentic AI into its workflows from the outset  Why explainability, auditability and human oversight are critical when deploying AI in a regulated market  How automation could change the roles of underwriting and operations teams  Why flexible technology matters when new underwriting teams need to be launched quickly  How large language models could change the way underwriters access their systems and information  Why AI adoption may now be a bigger challenge than technological capability  How legacy infrastructure can make deploying agentic AI at scale more difficult  Why the accelerating pace of technological change is forcing insurers to rethink what it means to future-proof their architecture  If you like what you're hearing, please leave us a review on whichever platform you use or contact Robin Merttens on LinkedIn.  Sign up to the InsTech newsletter for a fresh view on the world every Wednesday morning.

Parag Agrawal is the Co-Founder and CEO of Parallel, building web search infrastructure for AI agents. Parallel has announced $230M in funding from Sequoia, Khosla, and First Round Capital. Previously, he was CEO of Twitter, succeeding Jack Dorsey after serving as CTO and becoming the company's first Distinguished Engineer. AGENDA: 03:00 What Breaks When Agents Search the Web 1,000x More? 07:00 Speed, Cost or Accuracy: What Do Agents Really Need? 13:00 Will Tiny Models Catch Today's Most Powerful AI? 16:00 Is Model Routing a Commodity? 21:00 Is Amazon Making a Mistake by Blocking AI Agents? 23:00 Does the Ads Business Model Die in a World of Agents? 27:00 Can You Pay Publishers Without Killing Your Margins? 33:00 Why Parag Wants a Race to the Bottom on Price 43:00 What Stops Your AI Agent Breaking the Rules to Get Results? 45:00 Why AI Hacks Should Embarrass the Labs 52:00 What Parag Saw Working With Elon Musk  

Cortney Harding steps in as guest co-host this week, joining Charlie to unpack a genuinely strong Meta Connect, new occluded VR glasses aimed squarely at the Vision Pro, a camera-free Ray-Ban, and Muse AI's surprise climb to the top of the app stores. From there, the conversation turns to Disney's newly created CTO role, the quiet fate of its big-money bets on Epic Games and Sora, and Jeffrey Katzenberg's open letter arguing entertainment has always been a technology business. They close out the news with a sharp exchange on Amazon's push into drone delivery and a pointed critique of DraftKings' use of AI to target its most engaged, and most vulnerable, customers.Simon Horsman and Daniel Wood, co-founders of Quilty.app, join to talk about building an AI-powered script coverage tool for screenwriters who have no access to Hollywood's traditional gatekeepers. The two walk through how Quilty scores a script the way a studio story analyst would, the leaderboard system built to surface strong writing, and a striking real-world comparison Simon shares from his years running a film fund: a big-budget historical epic north of $220 million versus a hybrid AI production covering similar ground for under a million dollars. The conversation closes with how Quilty is trying to connect writers directly to producers in a system that's traditionally shut most people out.Key Moments: [01:35] A confident, glitch-free Meta Connect and what it signals [11:35] Disney creates a company-wide CTO role, hired from Character AI [13:35] The quiet status of Disney's Epic Games and Sora deals [15:35] Jeffrey Katzenberg's open letter on tech and entertainment [19:35] The backlash brewing over Amazon's drone delivery push [21:35] DraftKings, AI, and the ethics of targeting engaged gamblers [24:35] Simon Horsman and Daniel Wood join to introduce Quilty.app [30:35] A real budget comparison: $220M studio epic vs. under $1M hybrid AI film [34:35] How Quilty scores scripts and connects writers to producersBrought to you by Zappar and Mattercraft, the leading visual development environment for immersive 3D web experiences. Start building at mattercraft.io. Hosted on Acast. See acast.com/privacy for more information.

Sai Vemprala, CTO of General Robotics, talks about why robotics still faces a major gap between research and real-world deployment—and how the company's Grid platform aims to close it. Their conversation explores robot-agnostic intelligence, simulation, reusable skills, deployment architecture, enterprise data ownership, and the growing role of agentic AI. Rather than relying on one universal “robot brain,” Grid combines specialized capabilities—including perception, grasp prediction, and motion planning—into workflows that can be adapted to different robots and tasks. Sai also explains how simulation and deployment data can work together in a feedback loop, and how agents could help build, evaluate, and deploy robotic skills. ### Register now for RoboBusiness 2026: https://cvent.me/w0eRN9?RefId=podcast Listen to the podcast for a special discount code.

Earlier this month, world model company Runway introduced GWM Worlds 2, a research preview that “turns high-fidelity video and audio generation into real-time interactive simulation.” Runway calls this an “autoregressive diffusion” model; with autoregressive describing how it generates over time.One new feature in particular caught our eye: WorldPrompt, a proposed input format for specifying a generated world and the actions within it. It allows you to fix some aspects of a simulated environment — including the first frame — and then create a series of timestamped events. The events, or actions, can even be prompted in real-time.To understand the implications of WorldPrompt, we spoke to Kamil Sindi, Runway's CTO, and Robin Kahlow, its Principal Research Scientist for generative video and multimodal AI. We also have exclusive comments from Anastasis Germanidis, co-founder & co-CEO of Runway, courtesy of a podcast swyx and Vibhu did with him.Who's building real-time interactive world models?First, some context about world models that can generate interactive video and audio in real-time.Runway is reportedly valued at $5.3 billion, based on its most recent fund raise of $315 million in February. Its first release, GWM Worlds, was launched last December.Alongside Runway, there are several other notable projects in this domain: Google DeepMind's Genie 3 (which also generates at 720p and 24 fps), Odyssey-2 Pro, and World Labs' RTFM (Real-Time Frame Model). We've summarized their differences in the following table:Given the complexity and massive latency demands of real-time video and audio generation (which we'll get into below), all of the projects listed above have limitations. For instance, Google notes that Genie 3 “can currently support a few minutes of continuous interaction, rather than extended hours.”But as our interviews with Runway show, real progress is being made.The central idea of WorldPromptWorldPrompt, a new feature in GWM Worlds 2, helps differentiate Runway from its competition. You can think of it as a control layer for characters, cameras and the environment. As Kahlow put it, it's a way to “control all the different subjects in the world” — similar to a computer game.“Like, if there's an NPC [Non-Player Character] somewhere, the NPC might walk up to you and say something. So you could achieve the same thing with this kind of model, where you can have very detailed control over everything in the scene.”As the name suggests, WorldPrompt is a prompting mechanism — not a programming language. So, unlike virtual world games like Minecraft or Roblox, GWM Worlds 2 doesn't offer scripting capabilities or the ability to control state. But there's a power to that, as Sindi pointed out.“You can create promptable worlds on-demand with video and audio in sync, across all these different domains and environments. That's not a distant-future hypothetical thing,” he said.But there are also limitations to prompting a world model. We asked how reliably the model would follow an instruction to create, for example, a law of gravity or a certain ability in a character?“Yeah, so it's a research preview,” Kahlow replied. “So it's not perfect, of course, and there are still flaws. It really depends on how difficult the action is. I would say movement works quite reliably.”Sindi added that more training plus scaling the data and models is resulting in “better following.”How a video model becomes a real-time runtimeDespite the current limitations of GWM Worlds 2 — especially if you compare it to pre-designed and scriptable worlds like Minecraft or Roblox — the true promise of world models like Runway is that they'll eventually lead to fully self-generated, real-time games and experiences. Which is an extremely hard engineering problem, as Kahlow reminded us.“There are two challenges. One is making the model not generate a whole clip at once. So instead, you want it to generate frame by frame while you're looking at it. And the other challenge is actually making the generation fast, so you can play it in real time.”GWM Worlds 2 offers real-time interactive worlds streamed in continuous 720p video at 24 frames per second (fps) and audio at 48,000 Hz.Runway achieved this firstly by taking its foundational audio-video generation model and fine-tuning it to the new WorldPrompt format, so the model can follow that. It then post-trains the model to generate autoregressively.“And after that, we work on making it real-time through distillation methods,” Kahlow added.Co-CEO Anastasis Germanidis offered more technical details in our podcast with him. He told us that the process starts from “bidirectional diffusion that basically generates an entire video at once and [makes] it autoregressive.” This allows the model to “generate one frame or a few frames at a time.”Germanidis described two possible forms of distillation in order to make it real-time: distilling a larger model into a smaller one or reducing its diffusion steps. As a general example, he said a model might go from around 50 denoising steps to four, with some quality loss but potentially comparable results.The challenges of real-time generationGermanidis admitted that there were issues with how it generates real-time interactive video.“The biggest challenge with autoregressive models is error accumulation,” he said. “You're feeding generated frames back into the model to generate the next frames, and if there are any small errors, they accumulate over time.”Sindi told us there are also challenges dealing with “infinite generations” of content.“There's all these challenges around what context to keep, what to discard that's not important. And so there's all these optimizations we have to think about, so we're not blowing up our GPU memory.”Another current limitation is long-term memory. “The model does not have perfect memory,” Kahlow said. “That's still an open research problem.”Causality and correctnessWhile performance is the primary challenge for Runway at this time, its world model also has to produce plausible consequences when a user takes different actions.Germanidis used the example of simulating football; he pointed out that online video training data contains more successful goals than failed goal attempts, so a video model might render the first more convincingly.“If I take this action versus this action, you want it to generate equally realistic outcomes,” he told us. “That's, I think, the big gap between video models and world models: that idea of counterfactual generation.”Sindi told us that evaluation gets harder the more complex interactions get.“If you have this multi-prompt, multi-character, multi-scene [environment], how do you really understand what was causal and what was not?”To try and solve that, Runway has some automated verifiable tests. But since GWM Worlds 2 is a research preview, Kahlow noted that doing tests yourself is also advisable — “trying out your model to see what doesn't work is really important.”More than gaming — there are agent use cases tooGaming is the obvious use case for what Runway is building, but there are others. Kahlow mentioned robotics — for example using a simulated environment to test how a robot works.Another, more intriguing, use case is to use it to test agents at scale.“Having thousands of simulated environments is much less challenging if you have a suitable model like GWM Worlds,” Kahlow said.But how does an agent know what's changed in the world — is there a structured state that it can read, or is it just the generated video and audio that it's consuming and understanding?“So there's no structured state here,” Kahlow replied. “It's just observing the same thing you might observe in real life, just [in this case] from cameras.”Sindi noted that GWM Worlds can also be used for “synthetic data generation for agents.”Finally, Germanidis suggested there's potential to use these world models alongside reasoning models.“You're maybe using some reasoning [for] planning of the scene, and then you're passing it into the diffusion head that's actually generating the pixels.”Anastasis Germanidis* LinkedIn: https://www.linkedin.com/in/agermanidis/* X: https://x.com/agermanidisTimestamps00:00:00 Introduction00:05:17 Runway's Origins and the Bet on Generative Video00:12:23 The Stable Diffusion Story00:18:44 Gen-2, Controllability, and the Weekend Hack00:23:02 From Video Generation to World Models00:28:03 Learning From the World, Not Just Language00:35:04 Sora, Runway's Existential Crisis, and Gen-300:39:39 Why Real-Time Video Is Inevitable00:43:06 Interface World Models: Software Without Code00:50:25 The Fully Neural Operating System00:55:11 World Models for Robotics01:02:32 Robot Policies and World Action Models01:07:47 The Lucid Dream Test01:11:41 Video Agents and Omni Models01:23:12 Artists, AI, and Creative Workflows01:27:14 Physical AI and the Future of World ModelsTranscriptIntroduction: Runway, Creative AI, and the Early ThesisSwyx [00:00:00]: Okay, we're here with, Anastassios from Runway, with, me and Vibhu in the studio. Welcome.Anastasis [00:00:08]: Good to be here.Swyx [00:00:09]: Congrats on all your success and progress with Runway. You're opening offices all over the world. Did you envision this when you first started out?Anastasis [00:00:16]: Not quite. I think even when we started, we had this idea that, It was more a matter of when, not if, we were seeing the early generative models of 2016, 2017, and just extrapolating, assuming, we resolution, quality increases predictably over time. There's gonna be a point where most of content will be generated, and that was maybe the initial thesis of Runway was we will need, as a result of those generative models, rethink how creative tools are made. and as we built out the research behind, our generative models, it then became clear that they were useful far beyond that as well.Anastasis' Background: Art, Simulation, and Machine LearningSwyx [00:00:57]: And it is more obvious now with, like, the real-world stuff and the world models that we'll talk about later. I'm just kinda curious how you go from a background in, like, Zocdoc and, computer vision into Runway. Like, take us back to that early conversations with Chris and, whoever else is on your founding team.Anastasis [00:01:14]: I was always splitting through those two worlds. One was the I had my own art practice. I was making a lot of interactive art, I think for a long time. and then on the other side, I was working in startups, and I was working as a ML engineer, as a backend engineer at different companies. I've always been interested in, coding and computation, and especially interested in simulation and brought it back into my early artwork as well. And at the same time, I was interested inSwyx [00:01:43]: The personal site has a few, right?Anastasis [00:01:44]: Yeah.Swyx [00:01:45]: Is there one that we should pull up? Just in case there's something that's like. I just like to go down memory lane.Anastasis [00:01:50]: Yeah.Swyx [00:01:50]: Okay, what is this?Anastasis [00:01:51]: So this was, a project that I made, I think back in 2015, where I built this software that would give, voice instructions to people in a gallery space. So it would coordinate interactions between people. And so it will first give you an identity, like you're an, architect, you're 30 years old, and, you like sports. and then it would match you with another person, and you have this completely generated interaction. language models were not quite there at the time, and so it was it was a mix of some templates and some, like, some Markov chain-generated text, and it would just completely simulate these small talk conversations between, everyone in the gallery space. so was always very fascinated on the one hand with, generative models and, like, the early machine learning work that was being at that time. But at the same time, there was this separate thread of simulation and what it means. Like, what can we learn about humans by creating those very simple models of their interactions and their behavior?Early Generative Art: pix2pix, GANs, and Uncanny ValleyVibhu [00:02:56]: Did you generate the prompts or, the 30-year-old, whatever? Was it you generating them? How'd you, how'd youAnastasis [00:03:03]: Exactly. So the program would just generate- those, from. Yeah, a lot of it would be Mad Libs style of justVibhu [00:03:10]: YesAnastasis [00:03:10]: You have lists of different professions, lists of different,Vibhu [00:03:14]: HobbiesAnastasis [00:03:15]: Personality types, lists of different, ages, things like that. And then it would just combine those things together. And then maybe the next project we go is, Uncanny Valley, Uncanny Road, which wasSwyx [00:03:27]: GansAnastasis [00:03:27]: One of the first projects that, we built with, one of my two co-founders, Chris. This was taking, pix2pixHD, which was one of the early image-to-image models that NVIDIA released back in 2016 or 2017. and it was a model that would take a semantic map of a scene and then generate a photorealistic, let's call it, output. very early days, so it was not very high-fidelity outputs, but it w I think was the first image-generation model that could generate at 1K resolution. And it was all trained on self-driving datasets. So the semantic categories it would support were only, things you would encounter on the road. So it would be pedestrians, traffic signs,Vibhu [00:04:16]: StoplightsAnastasis [00:04:17]: Bikes, stoplights. And so that was one of our first indications that we built this and people were making all this, like, very surreal imagery of, yeah, a million plus a million pedestrians or a million traffic signs or, like, gigantic humans. And it was a indication that you could take a model that was trained on this very boring dataset, essentially, of, like, not that many interesting things happen when you're on the road, and then you can repurpose it and go very out of distribution and make something that was artistically compelling. And that was It's a summary of the thesis of Runway in some ways, that you can take the same generative models, and if you look at them from another direction, if you build interesting tools around them and you give them to artists, they're gonna do things that you don't expect.Vibhu [00:05:02]: Very cool. I like the, UX of it. You're just given an empty canvas, try whatever, do whatever. And then the other one, like, you see everyone with wired headphones? Like, that's, that's a sign that it's, it's veryAnastasis [00:05:16]: The AppleVibhu [00:05:17]: YeahAnastasis [00:05:17]: Apple, your version.Vibhu [00:05:17]: Original ads. Yeah. Take us to today. You've been doing this for seven years at Runway. How have we got to this? Like, how do we go from driving simulator data to all this? And you cover the whole stack of generative media?From Creative Tools to a Research LabAnastasis [00:05:33]: Interestingly, we're almost back in, we're, we're full circle. We're, we're now applying our models and beyond creative tools into real-world scenarios. But it was a, it was a long journey. It was very early on we realized the first version of Runway was a way to easily use the, all the open source model of the day, things like pix2pix to. and give them to artists. That was the initial idea, is those models are too difficult to use if you're not a machine learning engineer. Like, what happens when you give them to artists? Very quickly, we realized we needed to build a research org, inside of Runway, and that happened maybe on year one. And, a lot of the mandate there was. The image-generation models of the time, the video generation models of the time, or there were barely any video generations all the time, but they were not quite there where they could be productionized and brought into tools that would be part of creative workflows. so we need to push the frontier of the research. And so maybe the first four years of Runway, research was almost happening on the background until there was a moment in 2022, with latent diffusion, with, DALL-E 2, where, there was that step function change, and you guys maybe remember around the time.Swyx [00:06:49]: I started in this space because of latent diffusion and Stable Diffusion.Anastasis [00:06:54]: Yeah.Swyx [00:06:54]: Because I was like, “Wow, this is not only, like, feasible, it is doable on consumer hardware.”Anastasis [00:07:01]: Exactly, yeah.Vibhu [00:07:01]: I think the delta is also huge. Like, I learned pix2pix. Like, this was intro to ML, the TensorFlow, like, Jupyter, Google Colab notebooks were like this, and then you have a sudden step function change, with diffusion and whatnot. Any other ones since that. Like, there were clear examples of what early diffusion were to get to here. Any other changes in key technology research?Green Screen, Rotoscoping, and Early RunwayAnastasis [00:07:26]: Between, 2018 when we started and 2022?Vibhu [00:07:29]: Yeah.Anastasis [00:07:29]: So one of the early work that we did in Runway was solving segmentation, image and video segmentation. It was a very important problem because most VFX involves essentially separatingSwyx [00:07:42]: RotoscopeAnastasis [00:07:42]: Subjects. Yeah, rotoscoping. Extremely manual process. Nobody enjoys doing that. and so a lot of the early days of Runway was building this tool. It was called Green Screen, and it was for a long time the main thing that people were using Runway for. It ended up being used in, Everything Everywhere All at Once and a bunch of other high-visibility films and series. But that was essentially, Runway for a long time was a post-production tool until latent diffusion and generat- Gen-1, Gen-2, happened.Swyx [00:08:12]: Cool. let's, let's go past that moment. You've come a long way. Then you started releasing your own models. Maybe describe that journey as well.Scaling Video Models and the Bet on 1,000 A100sAnastasis [00:08:20]: Yeah, so we go to the other point, yeah, in mid-2022 when it became clear that we're doing research at a fairly small scale of compute, and it became clear that, like, scaling laws would apply to, image and video gen in the same way that we're applying to language generation. So we made a big bet, and I think at so at the time, we signed this deal to build a cluster of a thousand A100s, which at the time we were a Series B startup. That was a almost, slightly irrational decision maybe, but we really believed that if we trained a video model at a large scale, we would get, like, a great model at the end. And at the time, the goal or we set the goal around fall of 2022 of what is, what does the latent diffusion, Stable Diffusion moment look like for video? And at the time, the best model of the time was called CogVideo. it was one of the early video models. It was very 256 by 256 resolution, very not very high quality. and so we decided we're gonna build out this cluster, and we're gonna just invest in, like, in building out our own video model. it became clear as we're training Gen-1 that it was difficult to get to fully. we wanted to build text-to-video, but it became clear to us that an easier starting point would be to start from video to video. Because when you have a stronger conditioning, it's, it's an easier problem to restylize an existing video versus generate the video from scratch. And so we released Gen-1 first back in, it was January of, 2023. Yeah.Vibhu [00:10:04]: It's just a fun visual podcast, honestly. Like, if we can see February 2023, what was the state of stuff?Gen-1: Video-to-Video and Depth ConditioningAnastasis [00:10:10]: It's so interesting ‘cause at the time when you see those results, you think this is so incredible, and this is like, it's almost like image generation or video generation is solved. And then you look back a few years after, and it's like, it's It's just like you get used to the results very quickly, with those models. But at the time when we started seeing those results, it was, it felt quite incredible, and the level of, like, quality that you could get. And, so the Gen-1 was a depth-conditioned video model, so it would turn. it would take a input video, it would predict. it would it would first convert it into the depth map, and then we would generate, pixels with a latent diffusion model.Swyx [00:11:01]: Yeah, very effective.Vibhu [00:11:02]: Yeah. I didn't realize how distracting the blog post would be. Sorry.Anastasis [00:11:05]: Yeah, but, one of my favorite examples of on those, on Gen-1 was both, if you go up to mode three or mode two, there was this storyboard use case where people would makeVibhu [00:11:18]: OohAnastasis [00:11:18]: WouldVibhu [00:11:20]: You can mess around with theAnastasis [00:11:20]: Make a city out of books or out of boxes, and then they would shoot a video with their phone and then translate it into a photo-photorealistic output. There was all these ways in which those models were starting to be used for storyboarding and also for really. and then if you go to mode four, like, of taking untextured 3D scenes and then turning them into photorealistic output. So we saw a lot of use cases early on where people that were familiar, were power VFX editors would just take a blender, render, and then they would get translated in with Gen-1 or create a scene in Unity and then take a capture a video of it and then translate into, restylize it. So I still think video to video is powerful. I think we had a recent video-to-video model as well, and it's one of my favorite ways of using those models is essentially using them to use ground truth video as, like, the initial inspiration and then translate into different styles or different outputs.Stable Diffusion, Stability AI, and Open SourceSwyx [00:12:23]: But I think we're gonna go into, like, the rest of Runway and catch people up to speed today. I did wanna cover the, let's call it the Stable Diffusion controversy, or, what happened with Stability AI, whatever. I think there was a two sides of the story. I think there's part of that is a normal thing of, like, people, join and leave companies, but what is the, retrospective now that, there's been some years behind it?Anastasis [00:12:49]: Yeah, it's a very, it's a very long story to go into. I think it wouldSwyx [00:12:53]: Which I remember you wrote a really long post about.Anastasis [00:12:56]: We would probably cover the whole hour to go into it in more detail. But, essentially, there was the latent diffusion paper that came in, I think that was at the end of, 2021. And then Patrick Esser, who was one of the researchers behind, latent diffusion, and he worked at Runway at the time, he built latent diffusion in collaboration with Robin Rumbach and a few other folks back, in the in, CompVis, which was, a labSwyx [00:13:26]: Like a research group, yeah.Anastasis [00:13:27]: And, after releasing the early latent diffusion model, they, essentially they were. the goal was to keep working on versions of the model, scale it up, incorporate new data, incorporate new tasks. And Stable Diffusion was the same model, but trained on more compute, and then with a few more tricks, like a classifier-free guidance paper came at some point, I think in the early 2022. And thatSwyx [00:13:52]: Which, like, was a big prompting improvement.Anastasis [00:13:55]: Yeah.Swyx [00:13:55]:?Anastasis [00:13:56]: That improved results. it was trained on better data, so like, the esthetic subset of LAION, but it was effectively, the same underlying architecture. And there was that big training run, that, happened on Stability's cluster. Stability financed that run. And looking back at that story, I think it was the work to build and train that model was done. It was a, it was a research project. It was done as part of, like, continuation of the latent diffusion work. It then, I think it the model became very successful, and it, I think there were the. And I think as a result of its success, other companies tried to, figure out the commercialization path for it. But for us, it was very important that we try to, we make sure that we. It was meant to be an open source research project, and so the we decided that we should continue releasing versions of it, since that was the original goal of Stable Diffusion, and that led to releasing Stable Diffusion 1.5. There was maybe a day of, a bit of, miscommunication there, but ultimately that was resolved very quickly within hours. so yeah, there wasSwyx [00:15:12]: OkayAnastasis [00:15:12]: Not a niceSwyx [00:15:13]: I just wanted to. you have toAnastasis [00:15:15]: Yeah.Swyx [00:15:15]: You're one of the main players in that journey, and so it's nice to hear from the source of, like, what happened. Yeah.Anastasis [00:15:22]: Yeah. I think it's all, it's all in the past nowSwyx [00:15:26]: YeahAnastasis [00:15:26]: I would say. and, like, both companies, Stability took its own path, Runway took its own path.Swyx [00:15:32]: Yeah. There's still. James Cameron is backing the new Stability, whatever they're doing with the Hollywood studios.Anastasis [00:15:38]: Right.Swyx [00:15:38]: I don't know what they are doing. I think one thing that impresses me, and I'm happy to move on, is that back in the that time, let's say, like 2021, 2022, there was this community of people that you were involved in that was researching all this stuff, right? And, like, from everyone I talked to who was active then, it seemed like it was fairly obvious that somebody would do the hero training run that would produce Stable Diffusion. So, like, I guess the question is, like, you had the you were you had made investments. You were you had the foresight. Is it accurate to say, like, that is reflective of, like, what people were thinking at the time? Or was it still very much like, “Well, we'll use it as, like, a post-production tool or something. I don't know.”? Like, where in the sentiment were we that maybe you can think back to, like, what the community was like back then?The Early Creative AI CommunityAnastasis [00:16:28]: I reminisce and I think very fondly those early years, from like 2018 to 2022, because it was a very small community that, as you said, were very convinced that this was gonna be a big thing. And at the time, anyone who. Because it was such a small circle and, everyone who would, like, be part of that circle and, like, make projects with it would, immediately get, go viral. so likeSwyx [00:16:55]: And you didn't know who they are, right? They're just some name on a, GitHub or Hugging Face somewhere.Anastasis [00:16:59]: Exactly, yeah. So I remember one of the first big viral moments of creative AI was, there was the neural style transfer paperSwyx [00:17:09]: HuhAnastasis [00:17:09]: ThatSwyx [00:17:10]: Something dreaming?Anastasis [00:17:11]: I think it was called neural style transfer.Swyx [00:17:14]: Okay.Anastasis [00:17:14]: There was also Deep Dream, the puppy sliceSwyx [00:17:16]: YesAnastasis [00:17:16]: Which was, also really cool. but, yeah, there was this project that, Jim Kogan, who was an early advisor of Runway and one of those,Swyx [00:17:25]: Marketing guysAnastasis [00:17:26]: Big, creative AI, folks, he literally just, like, showed a video of himself taking the New York Subway and going over the Williamsburg Bridge and then stylized it with, I think in the style of Van Gogh or, like, one, painter. And that was. Like, at the time, that was, like, so cool and it went viral and it was completely revelation to people that you could do this with generative models. And that was only, it was less than. It was maybe 10 years ago. So just, like, as an indication of, like, how quickly things have gone.Vibhu [00:18:02]: It's pretty crazy. Like, even since then, you've got people at every level of the stack. You've got devs, creatives, artists, hobbyists. You've got everyone using it. And for people that tried stuff early, they'll remember how hard it was to use regular diffusion, right? Like, nowadays, you can use your favorite ChatGPT image gen or whatever, give a sentence, get a beautiful output. But diffusion was like, the whole ultra HD, 4K, high resolution. Like, prompting these things was very different. anything you learned on the tooling side, like from the offerings you guys have now, so like creatives, devs, you really took the. Research and brought it to everyone to use. anything interesting there to share?From Gen-2 to Controllable Video GenerationAnastasis [00:18:44]: We had to build the entire model serving infrastructure for video diffusion models. There was nothing else, already, like, because we had Gen-2 was the first text-to-video model, I think, out in the market. So many things that we learn over time. I think the I think the biggest one was, like, we. it was very clear early on that text-to-video was not gonna be the answer. Like, you. Like, people wanted a lot more control than that, and so we invested in, like, control building on top of those models very quickly. how do you use the camera trajectory as control? How do you use an initial input frame as control? So that was a very early learning for us. With text-to-video was, like Gen-2 was an amazing, step function improvement in the quality of video models, but it was used much more in an exploratory way because there was nothing to ground it to. There was no reference that you could bring into it. There was no. You couldn't really control the camera motion. You couldn't control the object motion. And so the first year, in 2023, was really all about what are all the interesting ways in which we can condition those models? And it was a lot of just post-training rounds on top of the base model to figure out, like, what, -- how do people wanna control them? And so there was, like, this quick succession of the we it was called Motion Brush, which was you could, like, you could draw arrows and dictate where things should move in the scene.Vibhu [00:20:09]: That's so cool.Anastasis [00:20:09]: There was camera control that was you could just describe, like, how you want the camera to move in the scene. And because we work with filmmakers from the most of the history of Runway, we immediately got this feedback and got this, decided that this was worth investing in. And so control ability became a big theme, I think, very early on as we were building, as we were building those models. Something fun that I haven't really talked about too much was just how Gen-2 came to be out of Gen-1. So it was a bit strange because we announced Gen-2 two months after Gen-1 andHow Gen-2 Came From a Weekend HackVibhu [00:20:43]: We're accelerating.Anastasis [00:20:44]: It was before Gen-1 was even generally available. But Gen-1 was a depth-to-video model, so it would take a depth map and it would convert it into RGB. and we couldn't get, text or image-to-video to work directly, and that's why we started from depth to video. but, and we had discussions of like, okay, we need to spend the next six months investing in text-to-video, maybe increasing the compute scale or the model scale, like train a larger model. And I had this weekend project idea, which was, what if I take a model that, starts from text input and converts to depth maps and then use Gen-1 to convert the depth maps Into RGB?Vibhu [00:21:29]: It would probably work.Anastasis [00:21:30]: And so Gen-2 was that.Vibhu [00:21:32]: Oh. The hackathon pipeline.Swyx [00:21:35]: The weekend hackathon pipeline.Anastasis [00:21:36]: Yeah.Vibhu [00:21:37]: But it looks good.Anastasis [00:21:38]: And it worked pretty well. there were if you, with the knowledge that it has this, like, two-stage pipeline, you can tell in some cases that the structure of the video looks a bit off because you had to generate the depth first before you go into the output video. But it worked and it allowed us to bring this to our, to users very quickly. But it's now it's interesting because, like, people are coming back to this almost two-stage approach. Like, if you look at the Reve text-to-image model that came a few months ago, it had this planner model that would generate bounding boxes before it fed that into the diffusion transformer.Swyx [00:22:19]: Yeah, Ideogram also the same day.Anastasis [00:22:22]: Yeah.Swyx [00:22:22]: I remember that was very strange that both of them came out the same day with the same exact innovation.Anastasis [00:22:26]: It's a small community, I think.Swyx [00:22:28]: I'm like, this is like, this is completely coincidental, right?Anastasis [00:22:32]: People talk. So yeah, there's, there's definitely something into this approach. And, now, like every single like, video generation model in production uses a complex prompt completion pipeline under the hood. I think that's no secret that there is. ThatSwyx [00:22:48]: Humans are terrible at prompting.Prompt Rewriting, Camera Control, and the Seed of World ModelsVibhu [00:22:51]: I think across the board.Anastasis [00:22:51]: Yes.Vibhu [00:22:52]: But yeah, I think like the original Sora one blog post even told you that what happens after your input is rewriting your prompt. It's much more descriptive about what you would want.Anastasis [00:23:02]: Exactly. I, And there was the DALL-E 3 paper beforehand that, was the first public, description of the fact that synthetic captions and really detailed captions work really well. And then Sora built on that. Yeah, so it was 2023. We were releasing all these updates to Gen-2, like the camera control, Motion Brush. And there was something very interesting about camera control because it was the first time that you felt that instead of, like, you were creating video, you were creating a short video, you were navigating inside the world. And I think camera control was maybe the seed of some of the ideas that we had around world models and really opening up that research direction. We realized, it was this era and this series of, Gen-1 and Gen-2 models really proved to ourselves, yeah, this is theSwyx [00:23:56]: Cool.Anastasis [00:23:57]: So this is not the original camera control. This was the updated camera control on top of Gen-3. But yeah, I think it made those models usable to filmmakers, I would say. The so camera control was very popular. And so we realized, there is one way of seeing those models, which is, you're just as content creation machines, and there is the other way, which is you're. As you're predicting video in order to predict video well, you need to simulate the world in an increasing and increasing capacity. And if scaling laws apply on video, just like they apply on language models, then as we scale the compute that we put into those models, then they're gonna be able to simulate physics, they're gonna be able to simulate human actions and dynamics increasingly well and predictably well. That was the thesis about around our efforts on world models, and we spin up this research group to just focus on the world models and how do we turn the video generation models that we're building into something broader and something that would be useful beyond, also content creation as well.Swyx [00:25:04]: And that was roughly when?Anastasis [00:25:06]: Yeah, so that was inSwyx [00:25:06]: OhAnastasis [00:25:07]: In late 2023.Vibhu [00:25:08]: Interesting. like, I think, a lot of people have been saying a lot of video gen model companies have all pivoted to world models these days, but like, 2023, you're posting it. oneWorld Models: From Video Generation to SimulationSwyx [00:25:21]: It's, it's debatable whether it's a pivot.Vibhu [00:25:23]: Yeah.Swyx [00:25:23]: Like, arguablyVibhu [00:25:24]: YeahSwyx [00:25:24]: That's what you always had to do anyway, right?Anastasis [00:25:26]: It's in a way an expansionVibhu [00:25:28]: YeahAnastasis [00:25:28]: Of the applicationsVibhu [00:25:29]: YeahAnastasis [00:25:29]: Of the models as they become more capable.Vibhu [00:25:31]: The early signs, it seems like the original models you guy had, guys had, people would say it's very not bitter lesson pilled, right? You're adding, rewriting prompts, you're having all these one-off things, but that's just the state of the tech as it was versus the future of as you said, you can scale it up as, we can scale up to world models.Anastasis [00:25:50]: Yeah. So it just became. And if you looked at the outputs of Gen-2Vibhu [00:25:56]: YeahAnastasis [00:25:56]: It was not. I think it was not obvious to people that this would scale to become a general simulator of the world. Like, you had very limited movement, you had, very low fidelity or low resolution, like obvious mistakes in human anatomy, like all kinds of limitations. But it was just, the idea was that's just GPT-two, and GPT-two, it can barely generate, like, coherent sentences. Similar, Gen-2 can barely create coherent video, but if you scale it up, you're gonna. There is no reason why it shouldn't work in a way. It's, And I think that was. That's, that's always the mindset of Runway is like this extrapolation of, like, if, like, even when we started in 2018 and you looked at the results of the day, you need to look more at the trend of, like, where we were in 2018 versus when we were at the, when the first GAN came out in twenty, four 2014 or twenty, fifteen. And, you started from, like, thirty-two by thirty-two images of faces, and then by the time in 2018, you could generate, street images at the 1K resolution. And it was the same with world models, very early signs of something much bigger.Swyx [00:27:08]: Yeah. I was gonna say, like, it's diffusing into focus. Like, if you look at our visible output from year to year, it looks like a diffusion process itself.Anastasis [00:27:17]: Yeah.Vibhu [00:27:17]: Especially watching the early, like, old blog posts, you can really see the choppiness, the details.Anastasis [00:27:24]: Yeah. Like human civilization starting from random noise and thenVibhu [00:27:27]: YeahAnastasis [00:27:27]: Denoising intoSwyx [00:27:28]: Yeah. Just run it a hundred years.Anastasis [00:27:30]: Civilization.Swyx [00:27:30]: Yeah.Vibhu [00:27:31]: That's how you're on track, you're still noising, right?Swyx [00:27:34]: Yeah. I like the way that you guys phrased it when you, announced it in June, which is, oh, that you had a video essay. “The human mind is no longer the center of AI. Our world is.” Right? Which is, let's, let's call it the past five years of LLM-based AI is very much like trying to emulate human preferences and human speech. But now that's, like, mostly solved. I think that's, like, some of the context of your essay, which you also wrote around the time. And now it's like the focus is on modeling the world accurately.Scaling Laws for Video and Why Predicting Pixels MattersAnastasis [00:28:03]: Exactly, yeah. So the way we see it is, there is that, initial mission statement of DeepMind, which is, solve intelligence and then use it to solve everything else. But I think it's starting from everything else, could be valuable of, like, starting from. there is just so much complexity, and detail in the world that in order to. That it's, it's hard to learn directly from just human descriptions of the world. Like, we're assuming that, like, language models learn from everything that humans have written about the world, like our own understanding as of, the twenty twenties. And there is just so much that we don't know and so much that's not captured by existing text, about both the low level dynamics of the world, like we're not describing in detail. if I tell you to describe, like, how do you tie your shoes, that's a very difficult thing to describe in words, but it's very obvious thing to demonstrate. And so I think there's been. And there's, more of X paradox, like we're constantly underestimating all the complexity that goes into very, like, things that we do subconsciously as humans, and we don't even necessarily always have the words to describe them. And so in my mind, the simulating the world and simulating, physics, simulating the dynamics of the world has always been underestimated, compared to, we place too much emphasis on the things that are easy to talk about. but there is just all this complexity and richness of the world that if we just try and train directly on that observational data instead of training on how people describe the world, we would learn something new that we wouldn't otherwise know.Swyx [00:29:54]: You think that the present architectural paradigm is fine? You don't need, like, another layer, like JEPA, like another famous, New York AI leader would say?Anastasis [00:30:05]: We're a very pragmatic research lab. If, we have evidence that an approach works better than the approach that we're taking, then we have no qualms to taking it. We just have seen no indication that video prediction itself doesn't scale. And even if you look now, not just our work, but the work of others, you're seeing in robotics some of the most promising work, starts from video prediction models, and then you adapt them to also the action models, for example. so there is very little evidence that you need something else and that your time is better spent on a novel architectural change compared to improving data and improving the, and scaling the current approach. And so, We don't have any indication that. the, there is that counterargument that I think there was a tweet by Yann LeCun a few days ago that, understanding the dynamics of the world is very different than, generating, cute videos.Swyx [00:31:05]: And your answer is no, they're the same thing.Anastasis [00:31:07]: Yeah, they're the same thing.Swyx [00:31:08]: My cat videos are the same as understanding physics.Anastasis [00:31:11]: Right, because if you wanna generate. video models can cheat and, like, they could you could give, like, successive dif shots of the scene in a way that doesn't require you to simulate difficult physics. There is like, all these different ways in which you can hide the deficiencies of the model, and it's important not to be too tricked by the performance of the current video models. It's easy to, cherry-pick examples and think that video models are further advanced than they are. So there is a lot more work that we need to do to improve those models. But in my mind, very similar to language, and, like, we've. you go from barely coherent sentences to something that, could hold a conversation with a human to something that could can operate autonomously for a day and, like, create entire code bases. And the main difference, there is some architecture improvements along the way, but the main thing is scale. And so it's the same bet for video, and we have no indications that this is saturating. Like, we have benchmarks that we use for measuring the physics of those models, and we see those predictably improve as we scale those models. So there is. If you want to Google up, Physics-IQ, is one of those benchmarks that measures how well does the model perform at solid mechanics or fluid dynamics or optics.Vibhu [00:32:32]: I'm curious if you've seen any emergence, any scaling law around this.Swyx [00:32:37]: Yeah, he's saying there is a scaling law, right?Anastasis [00:32:39]: Exactly.Vibhu [00:32:40]: Yeah,Anastasis [00:32:40]: So the way those models, those benchmarks work is you. the researchers have gone and, like, captured, a few videos that are representative of different physical phenomena, and then you can take the first frame and then pass it through an image-to-video model and then generate a rollout that shows what should happen next. So you have, a ball hanging from the ceiling, and then you use that as input, and then you the model predicts how the ball should fall on the ground. and this measures. we have an intuitive understanding of physics. I know, you can imagine what will happen next if I drop this bottle. So it's measuring that same intuitive physics understanding of those models, and we've measured that at different model scales, and we see, and compute scales, and we see that the score on physics IQ predictably improves. There's other, tricks and techniques that you can make to improve the score even further, but even scale alone helps, in the model learning better physics.Swyx [00:33:40]: My main sympathy with Yann LeCun is the, Plato's cave allegory, right? Like, you're, you're, like, learning on the output of a thing, not the internal process of a thing, and it's very noisy. And, if only you could observe the internals of a thing. It's hard to observe the internals of a human mind, but you can very much observe, or at least we have a whole branch of science and physics that we're ignoring on how to model Physics and movement and, gravity and, other interactions. and we're just, like, throwing away all of that and just saying just scale data, which is very much the lesson of unsupervised learning, but it feels wrong. that's the main idea.Anastasis [00:34:21]: I think the history of machine learning is, at large, it feels wrong.Swyx [00:34:25]: Yeah. It's a bitter lesson, right? Yeah. It's, it's, it's the simple answer to that.Vibhu [00:34:29]: I guess, how much can you scale? So, like, even on, let's say, the video generation side, like, there's one side of video understanding. Video generation, are we still gonna have tools where it's like, I wanna generate two hours, twenty hours? there's a infra way to do it in batches and stitch it together, but, like, do we just keep scaling? Do we just continue long generation consistency, all that at scale? And, like, tying it into where we're at now from we looked at Runway two to four point fiveGen-3, Sora, and Runway's Scaling InflectionAnastasis [00:34:58]: Yeah.Vibhu [00:34:58]: Like, technically, what advancements have we made to today, and then where do you see things still going?Anastasis [00:35:04]: So part of the answer is definitely scale. and that was. We learned that lesson in a big way for with Gen-3. So Gen-3 was the model we released the year after, like in 2024. That was a few months after Sora was released. so yeah, there's an interesting story of that came to be as well. Gen-3 for us was, the first time that we really needed to build. we had to learn all the lessons that the language model world learned in two in three years in the span of a few months. one of the biggest changes of Sora was using diffusion transformers instead of convnets. So a lot of the early, latent diffusion models were all, convnets for the diffusion model part. And the diffusion transformer paper came at some point in 2023, and it showed scaling laws for image, diffusion transformers. And we realized at that point that we needed to invest in infrastructure for model parallelism, for really scaling training to larger than, a few billion parameter models. And we spent maybe the, most of the fall of 2023 building out our infrastructure for distributed training. And we had a lot of false starts and a lot of failure in trying to scale, image and video diffusion transformers. And at that point, February 2024, Sora comes out, and the results areAnastasis [00:36:35]: Very much superior to what Gen-2 could produce. There were a lot of, a lot of chatter on Twitter about Runway. Runway's done. like, there is no way Runway will catch up. And if you remember, also OpenAI in the early twenty-It felt very, like it's aSwyx [00:36:56]: To the moonAnastasis [00:36:57]: It's a formidable opponent now, but at that point, it, they were on the top of their game. nobody could even get close to them. There was maybe Gemini was just the first version of Gemini had just released. So when OpenAI came with Sora and it was such a big jump of like quality, it gave me, there was like an existential crisis for a few hours. But that, I think the amazing thing about Runway and like I think the, we've been around eight years now, which is almost we're dinosaur in AI, and we had to like, we had there was a lot of those moments we had to learn, adapt very quickly and build out skill set in the team that we didn't have. And so, if you ask anyone what is their favorite time at Runway that was there during that time, it was that push in like three months to get to a model better than Sora. and it, we scaled 10x the model scale, the model size and the, compute that we were training on. we figured out model parallelism. We had zero expertise in that. And then we came out with Gen-3 during that summer. So that was a big turning point, I think, for the company where the research org grew very quickly, and we really started pursuing this vision of the general world model, in earnest, I think after Gen-3 was out.Swyx [00:38:12]: Yeah. that's the amazing thing about building when you're building. There's no stack to. You have to invent everything yourself. You have to be completely full stack. Now I think like there are inference specialists like Fal or whatever that can help with like, model serving, and I think you guys work with them as well. but yeah, like it's, it. But at the time, it was just. It's very interesting to think about what you do when Sora comes out and people are questioning whether your company should still exist.Distillation, Turbo Models, and Real-Time VideoAnastasis [00:38:41]: Yeah. And yeah, there was no, there was no VLM of diffusion models. Like, we had to build the whole model serving infrastructure and make things efficient. And a few months after we released Gen-3, we released the Turbo version, which I think was the first step-distilled model in production.Swyx [00:38:56]: That was a whole trend that we covered as well. Yeah.Anastasis [00:38:59]: So that allowed us, to serve those models at the larger scale, ‘cause I think the first version of Gen-3 was quite, expensive to serve.Swyx [00:39:09]: I think the whole like trend in like consistency models, Lightning and, Turbo and all these things somehow didn't really stick around. I don't know if you have any reflections on this. Because at the time, I was like, “Well, everything should start with a distilled model first, and then you can upscale,” right? It. your bigger models just turn into fancy upscalers, but like you should always draft with a smaller model and faster model, right? Because you can get it so quickly, like near real-time.Anastasis [00:39:39]: Yeah. I would not be so sure to say that didn't stick around. I think that, it's, it's likely to. that there is a lot of step-distilled models that are actively used in production. there is still a gap in quality compared to the, non-distilled model. but in my mind, we're still. there is a two to three year offset from language models. So the things that, So it's just a matter of time before there is better distillation techniques. we use. Right now we have a real-time model core character that I think is the largest deployment of real-time video models, that's a step-distilled model, and it's actively being used. It's a very specific use case compared to a general video model. So this is aSwyx [00:40:27]: Very cool, by the way.Anastasis [00:40:27]: This is avatars stuff, right?Swyx [00:40:28]: Consistency, character.Anastasis [00:40:30]: Yeah. So this is a talking avatar, model. we were able to. we optimized the hell out of it, and it generates at 24 FPS, and it's a, it's a step-distilled autoregressive video model. So if we look at our world model direction, a big component of it is starting from the bidirectional diffusion that generates entire video at once and making autoregressive shows. So you generate one frame or a few frames at a time. so there's a lot that goes into that pipeline of getting to a real-time model. It's first you need to make it into a causal autoregressive model, and then you just turn it into. You need to do some additional step distillation to get it to be real-time. and I think that part is just starting. I'll be very surprised if we're, two years from now, we don't primarily use real-time models. To me, real-time video generation is just inevitable that, it has much better user experience, it's much cheaper to serve, and, the quality gap between the base model and the real-time model is only gonna close as we figure out better, distillation techniques. And we made a lot of progress there internally on maintaining the quality of the base model when we distill them.Swyx [00:41:49]: How much of this is transferable? So is it the same base model? Like if you're doing diffusion across the whole sequence and you're converting it to step autoregressive distillation, is this like distillation where you still need to train both, you can use the same base and converter? What's that process like to go from regular model to something that's real-time on a technical level?Anastasis [00:42:11]: So the nice thing about diffusion models is you have, two axes of distillation. So there is the. You can distill to a smaller model, which resembles what you do in LLMs, or you can distill in terms of taking less steps, less diffusion steps. So you could take a model that generates in fifty steps and generate in four steps and get to, You have some performance, degradation, but very often you get comparable outputs. So you can even take the large frontier model and distill it with step distillation and get to a real-time performance, and that's what we've seen. So, depending on the use case, in some cases we might also serve with a smaller model, but in a lot of use cases, we just use theSwyx [00:42:56]: Step distillationAnastasis [00:42:56]: The frontier model, and we're able to make it work in real-time.Swyx [00:42:59]: I think this might be a good time to cut over to his laptop to show off some of the real-time stuff that you're doing.Interface World Models and Neural SoftwareAnastasis [00:43:06]: This is one of the research updates that we did recently. so we've been working and f in getting our general world models to, different applications. one of them that we think is very compelling is using general world models as essentially, an interface, a universal interface to software. This is a version of our world model that's called an interface world model. and the idea is that it essentially, replaces, the, front end of a software application. It renders the pixels directly of an interface and is trained to predict what happens next as a result of, a click or another interaction you have with the interface. So this is all pixels. it's there is no HTML, CSS, React that's powering this interface. This is directly at the output of our real-time, video generation model, and it takes clicks directly as input.Swyx [00:44:09]: And drags, click and drag.Anastasis [00:44:12]: Right. So it supportsSwyx [00:44:13]: Ooh.Anastasis [00:44:14]: Yeah, clicks. It supports drags. it also supports scrolling. and the amazing thing about this is that you can effectively describe in the prompt how you want different elements, like what do you want the behavior of different elements to be. So it's almost you're you can turn, an interface from, markup language description of, like, an HTML interface, and instead you can just describe the interface. if I press this button, I expect this to happen. If I press this button, this should happen. And it's useful, we believe, both for prototyping, for, like, just testing, like, what different interactions would feel like. you can also add audio to it. So it's a video audio generation model. So you get you essentially can describe both what the visual outcome should be of your click and also what the if there is a sound effect that comes out of it. So we believe that's gonna be a much more flexible way of building software. Just render. It just, in why generate the code that generates the pixels? Just generate the pixels directly.Anastasis [00:45:18]: It's the end-to-end philosophy applying applied to front ends.Anastasis [00:45:25]: So we think there is a few interesting use case. So you can build creative tools on top of it.Anastasis [00:45:32]: We think that, for any use case that involves a lot of exploration or, like, educational use case where you wanna learn about a new concept and you want some visualization and like, and open-ended exploration, we think those this is a very powerful, approach. you can imagine new forms of, design, industrial design software that could emerge as a result of those models. And this is all, generated in real-time as well. So, you can build a lot of interesting camera transitions and forms of interaction that are very difficult to build otherwise. And one way in which we evaluate this is what if you try to generate the same interface with Claude by just, prompting Claude, “Here's an image reference of my interface that I made in Figma or that I created somewhere else. create this particular interaction,” which in this case it's, drag that object, upwards. and beyond it being slower, it's also very difficult to capture some interactions by just fully, with just LLMs. So we think that this is likely to be the way that a lot of the future, like, software in the future will be created. and one of the additional benefits is personalization might be a lot easier done with those models. Like, you can essentially try out different prompts based on who is visiting the interface. You can, more easily, prompt engineer the interface to have larger size, text for more accessibility reasons, or you can make this or, like, if you have a particular aesthetic preferences. So we're very excited about this approach. It's early days, and I think we'll need to, make it more cost-effective as well to serve those models ‘cause, running a real-time video model versus just purely rendering HTML, there's -- the computational needs are much higher. but we do see a lot of potential in this approach to building front-end interfaces.Swyx [00:47:47]: So we covered this similar thing with Flipbook before with our, Ethan Hara episode with Groq, video. And yeah, I think it's very engaging visually. I think it's maybe very good for education, but it's it does sound expensive. I think there's an upper bound to how expensive it will be, though, right? Like, the inference cost will go down over time. You'll figure out ways to optimize it. Effectively, when it pauses, you don't you're not receiving human input. You don't have to generate anything, right? So.Anastasis [00:48:14]: Yeah, you could also. Like, in this case, you have ambient motion, so there is parts of the screen that might. if you're let's say you wanna, visit Paris and then you get this interface that allows you to explore.Swyx [00:48:29]: People walking. Yeah.Anastasis [00:48:29]: You have people walking or, like, things happening. But, it's, it's a no Yeah, it makes it more expensive because you need to run the model all the time. Maybe you have some looping mechanism so you don't need to do that. But all those things, I think, is stuff we'll need to figure out.Toward a Fully Neural Operating SystemSwyx [00:48:44]: Yeah.Anastasis [00:48:44]: I think our first consideration is let's make this clearly find some use cases where it's clearly a much more compelling interaction compared to traditional interfaces. And then it's a matter of time before it becomes more cost-effective to serve.Swyx [00:48:58]: Yeah. When it comes to the people walking, I think the approach that makes the most sense to me is Nick.Anastasis [00:49:04]: Nick.Swyx [00:49:04]: Oh, God. I keep messing up their name. With Chris Manning and Fanny Yan. I don't know if you've come across them, where they. Mapped to some game engine. I think it's Unity or something, or Godot. And they you can script some NPC behavior behind that and train on that. Whereas here, you can really imagine whatever you want. Like, that is a UI, right? Like, and it feels, like, more tractable, I guess, to, create a world model of software that is interactable because we have many of examples of that, and you can, do your fancy RL environment stuff on that than it is scaling up to embodied and real-world physical use cases. But this is a nice first step.Vibhu [00:49:43]: Or, there's the opposite of you have, like, one B models, three 50 million parameter language models. It just gets so small that they're just predicting, like, fishes moving.Swyx [00:49:53]: Small models are now 120 B, so.Vibhu [00:49:57]: Ultra mini on device.Vibhu [00:49:58]: But, no, I think it, like, it puts it into perspective, at least the car one for me, like, the applications, right? The amount of work to do that, sure, you only make one model year car per year, but applying this, it's also a cost-saving to have to manually make all this, right? So it opens up a lot of possibilities, too. I'm curious if you extend this out two, three years, so where do you see things going even further?Anastasis [00:50:25]: Effectively, the end game of something like interface world models is you have, a fully neural operating system. So I think, Andrej Karpathy has written about that quite a while back. But it's, You, I think to me it's, it's a bit, it's a bit odd that, we have, for example, with an interaction with an LLM of today, you have this LLM that can talk to you about anything. It can You can take the conversation in any direction. You can It's very general, so it can solve all those different tasks, but you interact with it through a very rigid interface. And so to me, it's just a matter of time before the interface itself becomes learnable and becomes, part of the whole loop of, like, you're not just delivering. You're delivering an application end-to-end, and that means you're delivering the language model, but you're also delivering the render and the pixels and that's also a learnable component. And the concept of applications might not necessarily. I think we'll need to figure out new abstractions for software. the concept of application comes from this idea that you need, separate code bases to describe, to, for, to power each individual, tool and each individual application. But you might think of something a lot more unified if you're. if you have, a video model that's generating the interface as you go. so it can take context from an LLM and allow you to combine different functionalities that traditionally would live in different applications. So it's a, it's a way to solve, software end-to-end, effectively. We also see this as a powerful way to train computer use agents as well. so this is, one way to see this as. And in general, with world models, there is those two directions. One is world models for humans and world models forSwyx [00:52:24]: AgentsAnastasis [00:52:24]: To train agents.Swyx [00:52:25]: Yeah.Anastasis [00:52:25]: And so for every new work of, world models that we do, we have this both uses become possible. So this is a powerful synthetic data generator for training computer use models. It could become, a live, RL environment that you could use to do online RL with a computer use agent, and you can get wide diversity of different interactions, kinds of interfaces, just generated on the fly that, to improve the how robust the, your agent, becomes. So that's the same also with the world models that we're working on for a robotics use case as well.Long Context, Error Accumulation, and Autoregressive VideoSwyx [00:53:02]: Is there a research breakthrough that you're Waiting for that would unlock the next set of use cases that you really wanna pursue?Anastasis [00:53:10]: Long context is a very important one, so being able to maintain consistency for long periods of time, and that depends on the use case. So for our characters model, for example, or for the interface world model, it's easier to maintain long sessions of interaction. If you go into more open-ended worlds that you navigate and you take arbitrary actions in, we, like, there is more the context at which you can and duration which you can generate becomes limited much more quickly.Swyx [00:53:40]: Yeah.Anastasis [00:53:40]: So we see more degradation and error accumulation happening. so the biggest challenge with autoregressive models is error accumulation, is you're feeding generative frames back into the model to generate the next The next frames. And if there is any small errors, they accumulate over time. That's not a new problem. It's a problem that LLMs also have, and we've seen the ability to generate now really long outputs. So it's a solved problem, but it's definitely still a challenge.Swyx [00:54:08]: Yeah. And what is the state of the art? so for Grok, it would be like 10 to 20 seconds of context going in there for video.Anastasis [00:54:16]: With our characters models, we're able to generate up to 30 minutes of video autoregressively.Swyx [00:54:21]: Yeah. But that's just for the avatars.Anastasis [00:54:24]: Yeah. So if we look at, GWM Worlds, which is more our open-ended world exploration model, it's, it's on the order of a few minutes, which is Yeah, soSwyx [00:54:35]: Probably enough for people because you have to cut to the next scene anyway, right?Anastasis [00:54:40]: Yeah, it's not, it's not the ideal game experience if you have to restart every few minutes. So I think. But, I think it's. Yeah, for certain kinds of game experiences, you can work around it.

Meta rilancia la propria strategia puntando alla cosiddetta Personal SuperIntelligence attraverso l’assistente virtuale Muse da usare su smartphone, computer, occhiali e su un nuovo dispositivo, ancora in forma di prototipo. Con Roberto Pezzali, esperto di tecnologia della redazione di Dday.it parliamo in particolare di Ray-Ban Meta Audio e Meta VR Glasses.Ci occupiamo di sicurezza dei dispositivi fisici. L’italiana Exein (che ha annunciato un finanziamento da 270 milioni di dollari con una valutazione di 1,7 miliardi) guarda alla transizione verso la Physical AI come naturale evoluzione della propria tecnologia pensata per proteggere un dispositivo. Ne parliamo con Giovanni Alberto Falcione, CTO e co fondatore di Exein.STMicroelectronics ha presentato un progetto che si svilupperà nei prossimi anni alle porte di Milano (nel centro ricerche di Castelletto) per creare un nuovo concetto di laboratorio di ricerca basato su un Digital Twin: un ecosistema digitale connesso capace di integrare tutti le fasi: modellizzazione, emulazione, simulazione, fino alla progettazione e al collaudo di nuovi prodotti. Enrico Pagliarini ne ha parlato con Fabio Gualandris, responsabile Ricerca, Produzione e Qualità di STMicroelectronics.E come sempre le notizie di tecnologia e innovazione più importanti della settimana.

On SlatorPod, AudioStack Co-Founder and CTO Peadar Coyle discusses what it takes to deploy voice AI at enterprise scale, from localization and domain expertise to human oversight.Peadar explains how AudioStack enables brands, governments, and media companies to produce localized audio across accents, regions, languages, and individual products. Rather than betting on a single voice model, the company works with multiple providers, treating models as building blocks and selecting them according to language, accent, quality, and use case.The CTO shares how AudioStack moved away from developing its own text-to-speech models after deciding its competitive advantage lay in orchestrating technologies, delivering high-quality audio, incorporating brand guidelines, and applying domain-specific knowledge.Peadar also argues that humans will remain part of AI audio workflows as automated systems can increasingly identify potential quality problems, but brands still want editorial control over pronunciation, tone, and reputation-sensitive content.On synthetic voices, he highlights consent, vendor due diligence, transparency, and regulation, with AudioStack introducing metadata identifying AI-generated content.Looking ahead, AudioStack is investing in agentic capabilities, faster and more intelligent workflows, and a combination of proprietary technology and third-party models.

If AI is better than your developers, why are your best engineers still writing code?Software engineering is the only engineering discipline where we make the engineer also turn the wrenches. Nicky Pike, Field CTO at Coder, argues that AI is finally letting your best engineers become architects while agents do the mechanic work — and lays out exactly what has to change about how you hire, govern, and provision your teams to make it real.Nicky has 20+ years in developer infrastructure — Xbox Live, private cloud at CVS Health, Cloud Foundry at VMware/Pivotal, and now the 'first mile' problem at Coder. He's a self-described 'platform guy, not a developer' who's watched this pattern play out through four platform shifts and sees the same arc happening faster. What you'll learn• Why the developer laptop is the worst possible endpoint for running AI agents• The identity split that fixes 'AI slop' overnight — human vs. agent permissions• Why Uber blew through their entire AI budget in four months (and how to avoid it)• How the software bottleneck moved from the inner loop to the outer loop — and why GitHub can't keep up• What to actually tell junior devs asking 'am I cooked?' ConnectNicky Pike on LinkedInCoder Chapters• 0:00 Introduction• 1:15 From Air Force crew chief to platform CTO — the 'works on my machine' problem• 2:44 Why this is the most interesting moment in tech since the cloud• 4:46 AI is Home Depot for software — the citizen developer reframe• 8:33 The homegrown-app problem coming back to bite us• 10:16 How Minecraft kids built Coder — and the 'first mile' problem• 13:00 Why the developer laptop is your biggest AI security hole• 16:17 AI is a Mensa student on caffeine — the identity fix that changes everything• 19:31 Uber blew through their AI budget in 4 months — the tokenomics trap• 20:36 Sovereign AI, open weights, and self-hosting the future• 29:02 The bottleneck moved — why GitHub can't keep up• 31:46 The airplane magazine trap — why bolting AI on will destroy you• 36:33 Software engineers have been doing two jobs this whole time• 40:27 GTG 1002 and why the disease is also the cure• 43:02 What happens to open source when AI writes the code• 48:19 What to tell junior devs asking 'am I cooked?' -- This episode of IT Visionaries is brought to you by Meter - the company building better networks. Businesses today are frustrated with outdated providers, rigid pricing, and fragmented tools. Meter changes that with a single integrated solution that covers everything wired, wireless, and even cellular networking. They design the hardware, write the firmware, build the software, and manage it all so your team doesn't have to.That means you get fast, secure, and scalable connectivity without the complexity of juggling multiple providers. Thanks to meter for sponsoring. Go to meter.com/itv to book a demo.---IT Visionaries is made by the team at Mission.org. Learn more about our media studio and network of podcasts at mission.org. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

What happens when the CTO and CISO roles come together in one person?  KEY MOMENTS 2:14 New risks come up or amplify? 3:12Copilot Rollout Risks 6:50 How to prevent this scenario? 9:20 AI Opportunity Mindset 13:23 As a CISO CTO, how do they use AI on a regular basis? 14:50 Starting use cases of AI 18:55 What ProArch does? 21:30 Conclusion  In this episode, Punit Bhatia speaks with Ben Wilcox about balancing innovation and security in the age of AI. They explore whether AI introduces new risks or amplifies existing ones, and what organizations must consider when scaling tools like Microsoft Copilot.  The discussion highlights the importance of data governance, access controls, and proper configuration to avoid unintended privacy risks. Ben also shares practical AI use cases from productivity gains to building agents and explains how organizations can start their AI journey responsibly. Overall, this episode offers a balanced view of AI as both an opportunity and a risk, with actionable insights for adopting it securely and effectively.  ABOUT THE GUEST Ben Wilcox is the CISO and CTO at ProArch, where he leads cybersecurity strategy and cloud innovation for enterprise clients. With deep expertise in risk management, compliance, and AI-driven security, Ben helps organizations navigate evolving threats while building resilient, future-ready systems.  ABOUT THE HOST    Punit Bhatia is one of the leading privacy experts who works independently and has worked with professionals in over 30 countries. Punit works with business and privacy leaders to create an organization culture with high privacy awareness and compliance as a business priority. Selectively, Punit is open to mentor and coach privacy professionals.    Resources & Links  Guest Links Ben Wilcox • LinkedIn: https://www.linkedin.com/in/ben-wilcox/ • Website: https://www.proarch.com  Grow Skills (Privacy Courses & Insights) • Courses: https://growskills.store/courses/ • Insights: https://growskills.store/insights/ • Website: https://growskills.store/  FIT4Privacy • Website: https://www.fit4privacy.com • Podcast: https://www.fit4privacy.com/podcast • Blog: https://www.fit4privacy.com/blog • YouTube: http://youtube.com/fit4privacy  Punit Bhatia • Website: https://www.punitbhatia.com  Books  • Be Ready for GDPR • AI & Privacy How to Find Balance • Intro to GDPR • Be an Effective DPO 

The AI systems guiding Ukrainian combat drones aren't running on expensive Nvidia chips. They're running on a Raspberry Pi Zero - a $15 hobby computer - and that single detail tells you more about how Ukraine's drone war is actually being fought than any headline about AI weapons. Craig Smith sits down with Alex Liannoy, CTO of NORDA Dynamics, for a technically specific and operationally candid conversation about what autonomous drone guidance looks like from inside the team building it. NORDA builds the avionics and targeting modules that plug into drones made by approximately 40 of Ukraine's hundreds of manufacturers, solving the two problems that make drone warfare hard at range: GPS denial from jamming, and the radio horizon that causes pilots to lose contact with their drone in the final 300 to 400 meters before ground level. NORDA's terminal guidance module takes over at up to one kilometer from the target, completing the approach autonomously while a human pilot's earlier lock defines the destination. The most counterintuitive finding in the episode is about what keeps humans in the loop, and it isn't ethics. Alex's argument is pragmatic: military doctrine requires proof that a drone hit its intended target. If you're monitoring the drone to verify the hit, you have enough connection to control it, which makes full autonomy less necessary than it might appear. That said, he's clear that the technology is "very close", one pilot already controls up to 32 drones simultaneously in some deployed systems, and target detection AI for notifying pilots of candidate targets is already in the field. The episode closes with a hardware finding that should recalibrate anyone's assumptions about AI compute requirements: after engineers mocked NORDA for using Raspberry Pi 4, they moved to the smaller, less powerful Raspberry Pi Zero. Both run real-time computer vision in active combat. The lesson Alex draws is precise: good engineering overcomes chip limitations, bad engineering can overload even a high-end processor. Subscribe to Eye on A.I. for weekly conversations with the people building and deploying the future of AI. 

Send us Fan MailAndrew Rabinovich is the former CTO of Upwork and has spent 25 years in machine learning and computer vision, long before AI was fashionable. After a PhD in computer vision at UCSD, he worked at Google. He then spent five years at Magic Leap building mixed reality devices before founding Headroom, a video conferencing company that Upwork acquired. Recorded live from the floor of Ai4 2026 in August, this lightning round explores what happens when you actually measure how much of a real job an AI agent can finish.Andrew and host Dan Turchin dig into Upwork's shift from a matching marketplace to a work delivery platform, the meta agent that breaks a job into atomic units of work, how the platform proves a freelancer is human, and more. What You'll LearnHow Upwork is evolving from a matchmaking marketplace into a work operating system, and how its meta agent decomposes a single job into atomic units of workWhat Upwork's AI productivity index revealed when agents were tested against real jobs already completed by human freelancersWhy every agent on the platform is paired with a human expert, and how that pairing changes both the speed and the quality of the resultHow proof of humanness works in practice, from identity verification to AI video interviews, and what happens when the system gets it wrongWhy agents may soon become the demand side of the marketplace, calling human experts in real time to verify what they cannot

Send us Fan MailOn this episode of Embedded Insiders, Eystein Stenberg, Co-founder and CTO of Northern.tech, joins the podcast to talk about how challenging IoT products have been to manage over the past few years. During the time of this recording, Northern.tech released the 2026 State of Industrial IoT Device Lifecycle Management report, and we discuss what that latest industry data reveals. Next, Nick Comfoltey, the Vice President of Home & Industrial IoT End Markets at GlobalFoundries, joins contributing editor Rich Nass to discuss power in AI. AI is front and center in so many conversations these days, and while we focus on processing, we often forget the need for sufficient power. For more information, visit embeddedcomputing.com

What happens when enterprises spend trillions of dollars on AI but the systems underneath it still cannot provide the context those models need to make reliable decisions? In this episode of Tech Talks Daily, I reconnect with Christian Monberg, CTO at Zeta Global, to examine what separates AI experimentation from production systems that organizations can actually trust. Our previous conversation focused on how businesses could use AI to scale marketing without losing the human connection with customers. This time, we move deeper into the technology underneath those experiences. Christian explains why disconnected tools and fragmented data remain barriers to AI adoption, and why Zeta rebuilt its data architecture using Palantir Foundry. We discuss the role of context graphs in connecting customer identity, business objectives, previous decisions, campaign history and outcomes so AI systems can understand more than isolated pieces of information. We also examine one of the biggest questions surrounding agentic AI: when should businesses allow an AI agent to take action? Christian shares what enterprises need around explainability, permissions, observability and learning loops before AI systems can safely move from recommendation to execution. With global AI spending expected to reach $2.59 trillion in 2026, the conversation ultimately comes back to a simple question: how can technology leaders prove that their AI investments are producing measurable business value?

In this episode of the Power Producers Podcast, David Carothers sits down with Nikhil Kansal, Co-founder and CTO of Cara, to discuss how Florida Risk Partners is using AI to eliminate administrative friction, improve the client experience, and create more capacity for producers to focus on revenue-generating activities. David walks through the actual workflows FRP has built around Cara, HubSpot, its phone system, raters, quote comparisons, proposals, and automated follow-up. Rather than discussing AI theoretically, the conversation focuses on what is happening inside the agency today, including 24/7 inbound lead capture, automated underwriting data collection, CRM creation, faster quoting, proposal generation, coverage comparisons, and follow-up sequences. The discussion also explores how these efficiencies can change the economics of small commercial insurance. By removing unnecessary human labor from repetitive tasks, FRP has been able to profitably handle accounts that previously would not have justified the producer's time while preserving human interaction for conversations where expertise actually matters. For middle-market producers, the larger lesson is about capacity. AI shouldn't replace the producer. It should eliminate the low-value work surrounding the producer so more time can be spent prospecting, advising clients, analyzing risk, building relationships, and closing larger accounts. Connect with:  ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠David Carothers LinkedIn⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠  ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Nikhil Kansal Linkedin ⁠ ⁠Cara Visit Websites:  ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Killing Commercial⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠  ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Crushing Content⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠  ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Power Producers Podcast⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠  ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Policytee⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠  ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠The Dirty 130⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠  ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠The Extra 2 Minute⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠   

CISA rides out a Cyber Storm. The EU struggles to share cyber threat information. Nightmare Eclipse drops another Defender zero-day. TASK#STOMP steals business documents. North Korean operatives fake their way through job interviews. A genetics lab pays $700,000 over a phishing breach. A zero-day in Meta's Muse AI assistant opens the door to privilege hijacking. Marc Woolward, Senior Advisor to Humanix and former CTO for Goldman Sachs, discussing social engineering and vishing attacks. Infiltrating Team PCP. Remember to leave us a 5-star rating and review in your favorite podcast app. Miss an episode? Sign-up for our daily intelligence roundup, Daily Briefing, and you'll never miss a beat. And be sure to follow CyberWire Daily on LinkedIn. CyberWire Guest Today we are joined by Marc Woolward, former CTO for Goldman Sachs and Senior Advisor to Humanix, discussing social engineering/vishing attacks and how ShinyHunters continues to use this method on other industries. Selected Reading CISA's Cyber Storm X tests cybersecurity preparedness across transportation, water and wastewater sectors (Industrial Cyber) Another worry for water systems: infostealer exposure (CyberScoop) Poor data sharing undermining EU cyber defences, auditors say (Reuters) New TASK#STOMP Windows Backdoor Enables Continuous Document Theft (Hackread) New Windows Defender zero-day blocks Microsoft antivirus updates (Bleeping Computer) Japan Dismantles First North Korean Laptop Farm as US and Allies Detail Wider Scheme (SecurityWeek) Ambry Genetics Pays $700K HIPAA Fine in Phishing Breach (GovInfo Security) Meta Muse AI app flaw lets local malware redirect dictation traffic (The Register) An Undercover Google Analyst Infiltrated a Notorious Supply-Chain Hacking Gang (WIRED) Share your feedback. What do you think about CyberWire Daily? Please take a few minutes to share your thoughts with us by completing our brief listener survey. Thank you for helping us continue to improve our show. Want to hear your company in the show? N2K CyberWire helps you reach the industry's most influential leaders and operators, while building visibility, authority, and connectivity across the cybersecurity community. Learn more at sponsor.thecyberwire.com. The CyberWire is a production of N2K Networks, your source for strategic workforce intelligence. © N2K Networks, Inc. Learn more about your ad choices. Visit megaphone.fm/adchoices

David Stifter has spent more than two decades at the intersection of real estate, technology, and finance. As Managing Director and functional CTO at Digital Bridge (formerly Colony Capital), he led data architecture, process improvement, and finance transformation initiatives across major acquisitions. In 2020, he co-founded PredictAP with a team of B2B SaaS and AI veterans from Blizzard, Apple, and HubSpot. Since then, the company has grown to serve more than 130 real estate firms and process more than 7 million invoices a year. David is based in Lincoln, MA.(02:20) From Colony Capital to founding PredictAP(05:11) When a $60B firm couldn't pay bills(07:23) Why invoice coding is harder than it looks(09:29) CAM pools, triple net leases, and the permutation problem(13:43) When institutional knowledge walks out the door(16:38) How to tell a real solution from a pretty demo(20:48) PredictAP: From inbox to Yardi in 30 seconds(24:47) Related Group, Cushman & Wakefield, and what 130 clients taught PredictAP(27:59) Why integration makes or breaks AP automation(30:17) Build vs. buy: what CFOs & COOs get wrong(33:51) Sheriff vs. shepherd: how to manage AI adoption internally(35:50) AP coding, audit risk, and financial misstatement(38:44) Collaboration superpower: Richard Feynman

The federal government is modernizing electronic health records to give clinicians more seamless access to patient data across agencies and improve the delivery of care. Lance Scott, CTO for the Federal Electronic Health Record Modernization Office (FEHRM), discussed how the office is coordinating technology efforts across agencies to improve interoperability and standardize health information sharing. Scott explored the state of EHR modernization across the government, including ongoing efforts with the Departments of Veterans Affairs and War. He also discusses how FEHRM is working with additional federal organizations to expand the shared EHR ecosystem and improve the exchange of health information across the government.

Илья Прахт, CTO-консультант, тренер и ментор руководителей, в гостях у Андрея Смирнова из Weekend Talk. Avito.Tech.Conf, успейте поймать билет на офлайн – https://clc.to/6aoG_A Телеграм-канал Андрея Смирнова – https://t.me/itsmirnov 00:00 Начало 00:29 Чем можешь быть известен моей аудитории? 01:03 Рекламная пауза 02:18 Путь в IT от джуна до начальника своего начальника 10:13 Как после сокращения превратить преподавательское хобби в работу? 17:42 Чему учить руководителей, когда уже рассказал всё, что знал? 26:45 Как продавать свой опыт, если не любишь продажи? 29:39 Кто учит тренера и как самому не отстать от профессии? 38:34 Где заканчивается личный канал и начинается бесконечный прогрев? 43:27 Зачем запустил «Путь CTO» и как развиваешь, когда закончился энтузиазм? 56:04 Кем бы стал, если бы не было IT-сферы? 58:59 Почему стоит переехать в Питер и (не) стоит в Саратов? 1:01:26 В чём сейчас главная проблема современного IT? Ссылки по теме: 1) Телеграм-канал Ильи – https://t.me/sedoydirector 2) Сообщество «Путь CTO» – https://github.com/cto-community/ 3) Профиль Ильи на GetMentor – https://getmentor.dev/mentor/ilia-praht-1210 4) Интервью Стратоплана с Ильёй о проблемах CTO – https://youtu.be/Y-tID4Rr3dQ

Send us Fan MailConfigure-to-order manufacturing has unique ERP requirements because product configuration sits directly between the sales and production processes. Unlike make-to-stock, make-to-order, or engineer-to-order environments, CTO manufacturers rely on configurators to translate customer requirements into products that can actually be manufactured. These configurators may be used internally by sales teams or exposed directly to customers, but the resulting configurations must ultimately flow into executable ERP processes, including bills of material, routings, pricing, costing, procurement, and production. Depending on the product complexity, configurations can range from relatively simple options and attributes to sophisticated, multi-layered product structures, making the integration between front-end configuration and back-end ERP execution critical.In this episode, our host Sam Gupta discusses the top configure-to-order manufacturing ERP systems in 2026. He also discusses several variables that influence the rankings of these ERP systems. Finally, he shares the pros and cons of each ERP system.Video: https://www.youtube.com/watch?v=Ql3IkCx7QDoQuestions for Panelists?

Tickets for AIE NYC now open, and apply for the invite-only AIE CODE. Join us!We have an unusual relationship with today's guest: for years since coauthoring the InstructGPT paper, Diogo Almeida had been saying that API-available frontier models have been going down the wrong path, everything from the alignment to refusals to reliability perspectives, that we have dropped every mode other than autoregressive chat-tuned LLMs because of the overwhelming success of ChatGPT.In a launch video now viewed ~40M times (by comparison, GPT4o was 22M, Fable 5 was 15M, Navier Stokes was 74M, and 6 Astra was 137M), Diogo introduced Jev and it immediately took over the AI timeline — we'll skip full Jev explainers because your favorite AI influencer/educator has probably already done one. We also collected:* the official patterns and cookbooks you should see first, from Allie* Jev usecases* speed based - games and computer use* the voice + computer use example we discuss at 1h34 mins* voice + browser control* The must not miss Doom demo* Driving cars in games* Excalidraw* virtual try-ons* “Smart Games”/smart NPCs* guided responses in text messages* Jev for coding agents has an official guide * jev for linting* compacting tool calls* reasonable pushback from Theo - Diogo has published a note on the Tyranny of the KV Cache that you should read as a followup after the pod for Jev + coding agents, because of his belief that Cache Rules Everything* Programming Languages built atop Jev (Diogo's fave)* Jev for analytics replay and user journey review* “dark data”* entity resolution* natural language search* “smart software”* a core goal of Jev is to “disappear into the background” - eg as unremarkable as regex* Jev as a judge* Jev memes* Jev vs LLM capabiltiies* blending transformers and classifiers* about the confidence api* Jev vs GLiNER (note difference/pushback, agreed, agreed, agreed)* Jev on trolley problem* Jev BushInstead we'll focus on what we can uniquely offer — a broader philosophical and mission-based understanding of how and why Jev was created, and what you should expect next in terms of future models from TypeSafe (ReasoningJev?) and what usecases and ideas you should work on vs the 55th low effort clone of Jev's API or doing a generic JevBench benchmark - something Diogo has rejected publicly.Why RLCD: Three kinds of RLHF, and why they are ALL the wrong north starDiogo knows a good deal about RLHF, given that he was on the team that pioneered post-training at OpenAI — and traces the three branches to Christiano et al 2017 (the robot backflip demo), Stiennon et al 2020 (learning to summarize) and his baby, Ouyang et al 2022 (InstructGPT). From there on, every innovation from Function Calling to Structured Outputs to Reasoning felt like a hack on top of the string based, sequence to sequence prediction paradigm. As he mentions on the pod, from 2023-2024 he struggled unsuccessfully, due to both personal and organization underestimation, to train a model that accurately addressed what he saw as the core problem with making LLMs the heart of software: reliability.Jev's core innovation is "Reinforcement Learning for Calibrated Decisions”, a novel, unpublished technique that optimizes for “answers with epistemically honest probabilities on System One tasks” rather than human rated feedback (RLHF) — which causes hallucinations, sycophancy, and permanent reliance on humans — or programmatically verifiable outputs with rubrics (RLVR) — which solves Navier Stokes but exacerbates jagged intelligence and doesn't integrate well with other software.We've talked about the calibration problem before on the pod, but probably the single best place to understand why RLCD became necessary is Diogo's AIE talk, which discusses why a generation of training helpful AI assistants for humans has impaired them for training models for composable, programmable AI for automation.At the end he also teases his contrarian opinion on scaling laws - which teases how to build a modern neolab without the billions of dollars the major labs have…The Bitterest Lesson: Tasks and Data beats ComputeWe spend a good amount of time discussing Diogo's essay on the Bitterest Lesson:His point is that “You get what you optimize for and the bitterest lesson in ML is that the most important part of it isn't ML at all.” - and picking the right north star, eg upvoting for user preference vs being integrated into tool calls - makes everything else fall in line.We're excited to catch up with a freshly dyed Diogo to discuss:* Why AI can solve extraordinarily hard problems but still fail to automate basic work* What System One Models are and why Jev is built for software rather than chat* RLHF, mode collapse, calibration, and the hidden costs of optimizing for human preferences* Why refusals become a problem when AI is buried inside software dependencies* Why TypeSafe rejects public benchmarks and optimizes for intelligence per dollar* The “bitterest lesson”: why the right task and the right data can matter more than compute* Why TypeSafe thinks of itself as a data lab rather than a model lab* RLCD vs. RLHF and RLVR as fundamentally different North Stars for AI* Why reliability and robustness matter more than simple determinism* Jev's programming primitives and how intelligence maps into software control flow* Why developers should decompose AI workflows into small, measurable decisions* How structured state replaces giant prompts and system messages* Why Diogo thinks AI should eventually disappear into the background of software* The “inverse SaaS-pocalypse” and how AI could supercharge existing software* System One vs. System Two intelligence and the limits of reasoning models* Dark data, computer use, real-time intelligence, and Jev's biggest early use cases* Why Jev could reshape coding agents built around a single-model architecture* Why Diogo says he wouldn't pre-train with $1 billion* The OpenAI journey that led to TypeSafe and why he thinks many neo-labs are approaching AI incorrectly* Coding agents beyond the KV cache, shared state, sub-agents, and the multi-agent futureDiogo Almeida* LinkedIn: https://www.linkedin.com/in/diogomda* X: https://x.com/CompleteSkeptic* TypeSafe AI: https://typesafe.ai/Timestamps00:00:00 Jev Launch Week and the AI Economic Revolution00:02:50 What Is Jev? System One Models and Programmable AI00:05:54 RLHF, Mode Collapse, Calibration, and Yann LeCun00:10:29 Programmatic AI, Refusals, and Safety Alignment00:17:21 Why TypeSafe Rejects Public Benchmarks00:20:43 The Bitterest Lesson: Data, Compute, and the Right Task00:24:59 RLCD vs. RLHF and RLVR00:28:42 Why Powerful AI Still Hasn't Automated the Economy00:39:55 Reliability, Robustness, and Determinism00:48:11 Model Versioning, LTS, Speed, and Intelligence per Dollar00:54:04 Inside Jev's API and Programming Primitives00:58:28 How to Build with Jev: Structure, Decomposition, and Small Decisions01:18:28 The Inverse SaaS-pocalypse and AI Disappearing into Software01:33:21 Computer Use, Dark Data, and Jev's Biggest Use Cases01:38:48 How Jev Could Reshape Coding Agents01:41:00 AI Safety, Frontier Pacing, and the Limits of RLVR01:48:03 Why Diogo Wouldn't Pre-Train with $1 Billion01:55:19 The OpenAI Story Behind TypeSafe02:01:41 Why Diogo Thinks Most Neo-Labs Are Getting AI Wrong02:08:00 Coding Agents Beyond the KV Cache and the Multi-Agent FutureTranscriptIntroduction: Jev Launch Week and Developer MomentumSwyx [00:00:00]: Okay, we're in the studio. A special occasion because this week, Diogo, my good buddy, launched Jev, and it's been taking over the complete timeline. How do you feel? What's it like to be you right now?Diogo Almeida [00:00:16]: Emotionally?Swyx [00:00:17]: Yeah.Diogo Almeida [00:00:17]: Never been worse. Like, I'm a ragged corpse of a person right now because there's so much going on, and I'm like a technical CEO, so I have, like, a lot of fires to fight.Swyx [00:00:29]: Yeah.Diogo Almeida [00:00:29]: But mentally, I feel—I say this all the time, and I've been saying this kind of for years in my over-under events. Like, I feel like the entire AI field is like one of those, like, carnival house of mirrors, and everyone is just insane and saying the weirdest stuff that doesn't make sense. And it feels like for just this week, like, I'm on a better in sync with reality and like, oh, people see it now. AI can be so much more than what was once thought.Diogo Almeida [00:01:06]: And like, yes, we are going to make. Like, an AI-based economic revolution is back on the table, and this is f*****g awesome.Diogo Almeida [00:01:17]: I'm so jazzed the developers get it. It's, it's, Yeah, and I want to show my eternal gratitude to the developers andSwyx [00:01:25]: Yeah.Diogo Almeida [00:01:26]: I'm so jazzed about the community and everything. It's so great.Swyx [00:01:28]: Yeah, you were saying yesterday that you decided to prioritize the town hall and not a bunch of, like, VIP, investor-type people because you wanted to make sure that they are the people that you get your most, attention, right? The engineers, the developers.Diogo Almeida [00:01:43]: Yeah, it felt a little like, oh man, I'm talking to, like, really important people right now.Swyx [00:01:47]: Yeah.Diogo Almeida [00:01:47]: I probably shouldn't reveal who.Swyx [00:01:48]: Yeah.Diogo Almeida [00:01:48]: But it feels a little bit dirty for me to, I'm, like, perhaps overly genuine in things. Like, it feels, like, dirty if, like, in my gigantic calendar event of people to talk to, the community isn't one of those.Swyx [00:02:04]: Yeah.Diogo Almeida [00:02:04]: And actually, in my ideal world, it would be, like, community all the time. I was thinking, “Should I host a town hall while walking to your studio?” And I'm like, “No, that's too crazy.”Swyx [00:02:12]: Sure. Yeah. Well, you guys have been hosting town halls on Discord. Discord is now 100,000 people. Your Twitter'sDiogo Almeida [00:02:19]: I don't follow these stats.Swyx [00:02:20]: Yeah.Diogo Almeida [00:02:20]: So holy s**t.Swyx [00:02:21]: Your Twitter's blown up. It was, it was really funny ‘cause, like, at AIE, you were like, “Yeah, follow me please,” and then you didn't, like, provide even your handle.Diogo Almeida [00:02:29]: I'm a noob. I'm a noob.Swyx [00:02:29]: You're such a noob.Diogo Almeida [00:02:30]: I'm a noob.Swyx [00:02:31]: But no, but that, like, that's, like, positive aura that, likeDiogo Almeida [00:02:33]: CoolSwyx [00:02:33]: You don't know how to promote yourself.Diogo Almeida [00:02:35]: Yeah. Someone, like, called me out when I posted, like, “Holy s**t, we're all three twending-- trending topics.” And then they're like, “That's a personal feed.”Swyx [00:02:42]: That's a personal, yeah.Diogo Almeida [00:02:43]: And I'm like, “Oh, no.”Swyx [00:02:44]: Of course, of course it'll trend to you.Diogo Almeida [00:02:45]: Cringe. Yeah.Swyx [00:02:45]: Yes, ‘cause it's what you clicked on.Diogo Almeida [00:02:47]: Yeah.Swyx [00:02:47]: So okay. Let's, Yeah, so congrats on everything.What Is Jev? System 1 Models and Intelligence per DollarDiogo Almeida [00:02:50]: Thank you.Swyx [00:02:50]: We'll talk about more, details as you have them. But let's, for people who are, like, living under a rock or just want, like, the definitive thing, what is Jev?Diogo Almeida [00:03:02]: Whew. Let me think about. That's a hard one.Swyx [00:03:07]: Okay. And I'm happy to, like, re-ask if you wanna kind ofDiogo Almeida [00:03:09]: No. I'm happy toSwyx [00:03:10]: OkayDiogo Almeida [00:03:10]: I'm happy to, like, just jam on it.Swyx [00:03:12]: Yeah.Diogo Almeida [00:03:13]: I will say, like, the first thing that I'm relieved about with this question is now I don't have to answer that question to my parents anymore ‘cause ChatGPT can just explain it.Swyx [00:03:20]: Nice.Diogo Almeida [00:03:21]: So the way I see it is we new-- need a new class of models. We're not attached to naming that class of models. Our-- the most accurate name we've come up with is System 1 models.Swyx [00:03:33]: Yeah.Diogo Almeida [00:03:33]: There will be reasons, but it's-- there's a reason why we don't call them decision models, because, like, they will be. Like, System 1 is beyond that. That's all I can say. We didn't expect this to be our big launch, so we have stuff in the tank.Swyx [00:03:48]: You should have said low-key research preview.Diogo Almeida [00:03:52]: It kind of was, right? It kind of was. But we. So there's a class of models that we describe them as, like, machine-native, System 1, large programmable. I think these are-- is the class of models where the goal is for code to be the consumer. So as opposed to, lar-- pre-trained large language models, which are meant for, like, autocomplete of the internet, or RLHF models, like chatbot instruction-following models, which are meant to, like, reply to text, or RLVR. It's in a weird gray area with RLHF. Like, these are meant to have things that directly are consumed by code, hence the name type safe. So the thing we really want is to have, like, AI, like, be as powerful as possible, and we think the way to do that is to integrate it with software. And we are designing everything, beyond just the outside, the deep internals of the model to be optimized for software. So number one, Jev is our first large programmable model, or a System 1 model, whatever you want to call it. Jev is meant to be optimized for intelligence per dollar, hence the name Jev.Swyx [00:05:03]: Jevons Paradox.Diogo Almeida [00:05:03]: Jevons Paradox, yeah. And it's optimized for intelligence per dollar. I love this debate with people about what is the most important between reliability, cost, calibration, and speed. And Jev is meant to be. Jev will be the name of models that will be on the frontier of intelligence per dollar. There's other ways to optimize it, like, ML, or at least if you're good at ML, it's all about trade-offs. And we are just going all out on that.Calibration, Mode Collapse, and the Limits of RLHFSwyx [00:05:31]: Yeah. And to me, like, calibration is one of the new things that people weren't talking about as much. We've done an episode In the past, with Clementine Foreia of Hugging Face, where they were like, “Yeah, actually, y- they're just.” Or, and this is your whole argument about RLHF, is they're more collapsing towards what you want to hear the mostDiogo Almeida [00:05:50]: OohSwyx [00:05:50]: Or what is most likely, instead of, like, their own internal confidence about a thing.Diogo Almeida [00:05:54]: Can I soapbox on that for a second?Swyx [00:05:56]: Go ahead. Yeah.Diogo Almeida [00:05:57]: Cool. Like, I've been heard that your audience is the most technical, so I actually want to get into that.Swyx [00:06:02]: Yeah.Diogo Almeida [00:06:03]: And if- I went through extreme precision to make sure everything in our launch video is accurate and real. Apparently, that's very unusual. One of the things that no one paid attention to was the downsides of RLHF, in particular mode dropping.Swyx [00:06:17]: Mode dropping or mode collapse?Diogo Almeida [00:06:19]: It's the same thing.Swyx [00:06:19]: Is that what you call it?Diogo Almeida [00:06:20]: It's the same thing.Swyx [00:06:20]: All right.Diogo Almeida [00:06:21]: And I wanna have a blog on this eventually, but I, like, want to tell as many people this as possible ‘cause I think it's a very interesting thing. So the spicy take, I believe in Yann LeCun a lot. I think Yann LeCun's takes are actually among the closest toSwyx [00:06:36]: What about this?Diogo Almeida [00:06:37]: Well, should I address this now or should I wait and go into mode collapse?Swyx [00:06:40]: No, later. Go mode, go mode collapse. I don't know.Diogo Almeida [00:06:42]: So I actually think that among takes, Yann LeCun's is among the most accurate. But he has this very famous/infamous slide about,Swyx [00:06:52]: The cake?Diogo Almeida [00:06:53]: LLMs are doomed.Swyx [00:06:54]: Okay.Diogo Almeida [00:06:54]: Like that one where he, like, has, like, a pie chart with, like, a tiny par-- tiny little thing- and says that as you increase sequence length, the probability of it making an error goes in. Yes, this one. This one. I love this one, because it's one of these things that seems mathematically obvious, but is obviously wrong, right? Like, it's mathematically obvious, but it doesn't empirically hold. And this is my favorite thing to teach people about, like, where youSwyx [00:07:21]: What's the disconnect, right?Diogo Almeida [00:07:22]: Exactly. And may I or you want to tell me?Swyx [00:07:27]: About mode collapse?Diogo Almeida [00:07:28]: Oh, no. Oh, so mode clop-- collapse is related to this.Swyx [00:07:31]: Yeah.Diogo Almeida [00:07:31]: The disconnect happens because if you are in a mode covering or a calibrated distribution, you are, like, not. You are not overly punished about having outliers. You'd expect, like, something. Some amount of the time you'd be out of distribution, some amount of time you'd be in distribution. That's what happens when you cover the distribution. This was like models before GANs. They made blurry images, right?Diogo Almeida [00:07:54]: Instead, GANs mode drop. They, like, drop the minority classes and just do the really common ones. And this is why this effect doesn't happen, right? Like, instead of be-- in order to generate really long strings, without making errors, they need to, like, be extremely conservative because it's e- really easy to see when an error happens. It's very hard to see when, like, a subtle thing that looks correct happens. And that calibration is, like, total poison into, like, the probability distributions of strings.Swyx [00:08:22]: Yeah.Diogo Almeida [00:08:23]: And it's, it's a nuanced take and like, I think that This is why this doesn't happen, and this is why strings are so bad at, decision-making or, overloading the string models are for decision-making is, like, a bad time.Yann LeCun, JEPA, Scaling Laws, and Practical ResearchSwyx [00:08:38]: And while we're on the topic of Yann, do you agree that his fix i- with-- which is like a world model, like a JEPA-type, embedding thing is the right solve? So basically, like, the. One of the reasons that it could fail is because you're trying to reason over token outputs and then, and then just looping back again and going. Keep, continuing going until you reach, like, a end of sentence. Like, is that, And his solve is JEPA, right?Diogo Almeida [00:09:02]: Yes.Swyx [00:09:02]: Which is, like, joint ambition,Diogo Almeida [00:09:04]: YeahSwyx [00:09:04]: Joint embedding prediction. So like, is that the solve or, like, do you have a. Do you have a take on that?Diogo Almeida [00:09:10]: Oh, man. I probably shouldn't talk too much about the insides of ML, but I will say that my brand, other than unhinged, is practical.Diogo Almeida [00:09:20]: Like, even my take here is practical. And like, I'm. Am I a scaling law fan? Depends. It dep-- it's, it's, it's, like, it's. Scaling laws tell you how much better you get at a thing for amount in.Diogo Almeida [00:09:33]: A scaling law does mean exponentially more resources for normally sublinear gains, which looks to be a bad investment unless those, like, linear gains are, like, really valuable. But it's all. To me, it's all about, like, what can we do with what we have to make the biggest possible f*****g difference? I can curse.Swyx [00:09:51]: Yeah.Diogo Almeida [00:09:51]: Yeah.Swyx [00:09:52]: Yeah.Diogo Almeida [00:09:52]: Yeah.Swyx [00:09:53]: We're, we're, we're approved for adults.Diogo Almeida [00:09:54]: Hell yeah.Swyx [00:09:55]: And also we have a scaling law thing if you wanna go into that later.Diogo Almeida [00:09:58]: Oh, I could if we. See, that part is not super relevant right now.Swyx [00:10:02]: Yeah.Diogo Almeida [00:10:03]: I actually. If you wanna go into my bitterest lesson, I think that's more relevant.Swyx [00:10:06]: Okay.Diogo Almeida [00:10:06]: But like, to me, I'm all about, like, pragmatics. And I think that the JEPA stuff is really cool early research. I really love awesome research. Is it practical yet?Diogo Almeida [00:10:21]: Probably shouldn't say. But like, there's just a lot of.Diogo Almeida [00:10:29]: I just think there's just, like, so many diamonds in the rough let all over the research world right now that haven't been polished because people don't know how to, like, do the right task. And I think that what our launch did, it. Does it kickstart us as a company? Like, yes. Will it be great for us as a company? Yes. I think it's gonna be, like, even greater for this direction of, like, programmatic AI. There was going to be, like, a gold rush on top of us for. ‘cause, like, software is super f*****g charged. But I think there's gonna be a gold rush parallel to us as well on, like, all the different ways we can expose things to make software more powerful so people can make even cooler stuff. And then we are back to, like, early internet energy?Swyx [00:11:12]: Yeah.Diogo Almeida [00:11:12]: And I think that's why, like, the Twitter is just like, “Jev.”? It's, it's like. It is a partySwyx [00:11:18]: It's inspiring because it's, it's, like, so different than what we're used to, which is, “I'm sorry you can't do this, but we do scaling laws and only the big labs can do it,” right?Diogo Almeida [00:11:28]: That. Actually, if I. I'll, I'll make a tangent if that's okay.Swyx [00:11:32]: Yeah.Diogo Almeida [00:11:32]: I think you might enjoy this.Swyx [00:11:33]: Really? Our five tangents in. It's good. It's fun. Yeah.Diogo Almeida [00:11:35]: Oh, yeah. I get lost at all my tangents.Swyx [00:11:37]: This is gonna be horrible for the listeners to figure it out, but they're gonna figure it out. It's fine.Safety Alignment, Refusals, and API PhilosophyDiogo Almeida [00:11:40]: Yeah, we can edit it in post.Swyx [00:11:40]: This is my response. Yeah.Diogo Almeida [00:11:41]: So popular thing on Discord, that people keep asking me, I haven't had the time to explain it yet, is why am I opposed to safety alignment and why do we not refuse? I'm not opposed to safety as a principle, but I think that safety alignment is generally misaligned with users. And refusal is just, like, obviously a type error. Like, if you're a human being and you're chatting with, like, a bot or whatever, you're cloud coding, and a refusal happens, like, “I'm sorry, I can't read DNA.py.” that's an annoying time. It's anno- it's, it's annoyingDiogo Almeida [00:12:18]: Right? But you can work with it, right? And you're forced to work with it ‘cause of Stockholm syndrome.Diogo Almeida [00:12:23]: I have stories about that too. I need another tangent deep in here. But like, if you ever want this in a dependency running in the background, what happens if that refuses? What if someone else is using that dependency? They don't know what that system is. Like, you want the software to just stochastically break because a user sent, like, a weird message in there?Diogo Almeida [00:12:42]: Like, that is, like, straight-up insanity. It's coming from a place of, like, people who do not understand software, do not understand programming, and like, they are obsessed with, like, I believe this, horseless carriage of, like, AI coworker instead of unearthing, like, the full power of AI.Swyx [00:13:01]: Fair enough.Diogo Almeida [00:13:01]: Yeah.Swyx [00:13:01]: You want something that is the core kernel that is usable everywhere.Diogo Almeida [00:13:05]: Yes. Exactly. Like, the cognitive core, right?Swyx [00:13:07]: Yeah.Diogo Almeida [00:13:08]: And you need this thing to be s- like, so general, so optimized for its use cases. You want it to be, like, you want it to work on all the future use cases, all the weird s**t that people are doing.Swyx [00:13:19]: Yeah.Diogo Almeida [00:13:19]: We obviously didn't train on any of that stuff. Is it surprising that it works? No, ‘cause we trained on weirder stuff, my friend.Diogo Almeida [00:13:28]: So. But one tangent up about, like, safety alignment.Swyx [00:13:32]: Okay.Diogo Almeida [00:13:32]: Safety alignment makes sense for a product, in my opinion, for, like, ChatGPT and Claude. Like, it, What safety, what makes safety and capability alignment different is capability alignment is, like, about doing what the user wants. That is sick for software engineers. They want their thing to do the thing, and the more predictable it is, the less they have to test it and play around with it. Jeb is not anywhere close to that yet. It could be, but like, there's so many more nines of reliability that we want in order to make it so good, like a database query, that you don't even have to think about it. It is just there when you need intelligence. But safety alignment is, like, the opposite of instruction following. It's when you want to follow someone else's instructions, like OpenAI and AnthropicSwyx [00:14:13]: The RAGs value stack.Diogo Almeida [00:14:14]: Exactly. And this makes a lot of sense for a product. Again, like, ChatGPT should do. Y- you sh- like, if they don't want to, like, do, like, some, not-safe-for-work role play with ChatGPT, that's on them because, like, maybe that's, what their users who have, like, parents and kids want. Like, n- that's fine. But in an API, that's nuts, right? Like, that's completely unacceptable because, like, people need to, like, program around this, and that is, that's so anti-user that it's. It. I'm. Huh. I can be an angry person, so I should try to calm down.Swyx [00:14:52]: It's, People get your passion, and I think that's really good. The one pushback I'll give you is, like, what if we use it to kill people, right? Like, that is the actual. Like, n- the not-safe-for-work thing, it's private, personal, whatever. But like, yes, like, we will use it in war. And like, that is, something that companies can reasonably prefer their APIs not be used for.Diogo Almeida [00:15:14]: I get that. I think that there's, like, pragmatic places where that opinion can be held. I don't think the foundation of, like, a general-purpose technology is that place, personally.Diogo Almeida [00:15:27]: Like, would I prefer that our stuff is not used to kill people? Obviously. Would I prefer it's used for, like, all sorts of, like, great stuff in the world? Obviously. Will I put my thumb in the scale for that? Yes. Will I do it at the technological layer? Absolutely not, because that will fracture the intelligence. Every single time you mean it to overfit to some weird stuff, you're fracturing its intelligence more and more. And like, these things are fractured to the, like. They're so darn fractured right now.Swyx [00:15:54]: Yeah.Diogo Almeida [00:15:54]: So and as a furthermore thing, to me, it's like I think intelligence will be more like a database than a coworker. Like, I don't think it's up to databases to add checks on whether or not they're used for, like, what's something that's not great? Like, CIA. Actually, I don't know what the CIA does, really. You can imagine. You can imagine, killing people who are not even bad or whatever.Diogo Almeida [00:16:21]: And like, I don't think it's the database's responsibility for that. And furthermore, like, a thing that has been weird to me is when people, like, sign up for our thing on Slack and they're like, “Hey, we're gonna deploy this. Can we deploy this thing?” I am just like, “My brother, we are an API. You are a developer. It's none of my business,” right? Like, you shouldn't know what the whole task even isSwyx [00:16:46]: YeahDiogo Almeida [00:16:46]: Because it should be decomposed into small things. We shouldn't be able to know what the downstream users are doing, and that is, like, a good boundary to give software engineers maximum power. Ideally, they use it for the good stuff, and ideally, we can, like, help them and like, we've talked about, like, doing open source and charity and all of that. We have absolutely no time for anything else right now. But like, they will get any of that bias out of the technological layer as long as I'm in charge.Privacy, Benchmarking, and Trusting IntelligenceSwyx [00:17:11]: Yeah, that's great. While we're on the topic, let's also briefly talk about your privacy stuff, terms of ser- terms of use, which, got a little bit ofDiogo Almeida [00:17:18]: OohSwyx [00:17:18]: Misunderstanding. I just wanna clarify that upfront.Diogo Almeida [00:17:21]: Hell yeah.Swyx [00:17:21]: I think this probably takes two sentences from you about, like, you will not. You're not being that restrictive about your API. Like, clearlyDiogo Almeida [00:17:27]: Oh, yeah. Oh, yeah, so yeahSwyx [00:17:27]: Ideologically, you articulate your role as a platform very seriously.Diogo Almeida [00:17:30]: Yes. Yes. I don't know what you're referring to, but like, this was. I've seen a couple of things about, like, benchmarking.Swyx [00:17:38]: Yes.Diogo Almeida [00:17:38]: Like, obviously we're not stopping people from do. Oh, man, I should be careful about what I say. I'm realizingSwyx [00:17:43]: No, you said, you said it publicly thatDiogo Almeida [00:17:44]: YeahSwyx [00:17:44]: That was in the preview period. You didn't take it out for the launch.Diogo Almeida [00:17:47]: Yeah. Okay.Swyx [00:17:47]: And now you're gonna take it out.Diogo Almeida [00:17:48]: So the team is doing stuff thatSwyx [00:17:49]: YesDiogo Almeida [00:17:49]: I'm not even aware of, so it's great to know the team communicated that. I asked them to check in with the lawyers about that.Swyx [00:17:54]: Yeah.Diogo Almeida [00:17:54]: Like, we are obviously not stopping people from doing that type of thing. I'm extremely in favor. So I'm extremely anti-public benchmarks. I'm extremely in fa- I'm medium about private benchmarks that are proxies. ISwyx [00:18:09]: So are you worried about, saturation or, like, training on public benchmarks? So it's, like, easy to cheat.Diogo Almeida [00:18:15]: Not only is it easy to cheat, there's a lot of ins. So I think that we are. Or anyone who's, like, competition with us that, vaguely there is. Like, you could say, likeSwyx [00:18:28]: There's like 50 Jev clones, yeah.Diogo Almeida [00:18:30]: Well, sure.Swyx [00:18:31]: Yeah.Diogo Almeida [00:18:32]: Well, the, these. Let's say that there is competition.Swyx [00:18:34]: And we'll talk about those. Yeah.Diogo Almeida [00:18:34]: Or let's just say that there's. Let's just assume that there's an industry two years from now of people who are doing similar things to us. The thing that we are selling is intelligence per something, per, like, dollar or per second. The. No one. Like, people obsess about the cost and the speed. I believe that is. It's cool, but like, the thing that matters is the intelligence. Like, the cost and the speed are, like, are bad things. You're paying them for something, and you need the thing back, and the intelligence is what truly matters. The problem with intelligence is that there's a je ne sais quoi to it, right? Like, the good model smell. Like, the thing that happened after we launched of, like, two hours later that actually went way bigger than the video, which was like, “Holy s**t.”Swyx [00:19:16]: This is actually usable.Diogo Almeida [00:19:17]: It. WellSwyx [00:19:17]: Yeah.Diogo Almeida [00:19:17]: It's, like, beyond that.Swyx [00:19:20]: Yeah.Diogo Almeida [00:19:20]: Like, the. Whew, the launch was crazy, and people could really sense how hard we care about that, and that's truly what I think the long term of this is. And I think public benchmarks are antithetical to this. Like, they are a way to get people trust in intelligence because intelligence has a je ne sais quoi, but the public benchmarks are extremely gameable. Even if they try not to, they still will. Like, back in the old days, every lab had a team to collect data that looks like MMLU to make it look better, which is just benchmarking with extra steps.Diogo Almeida [00:19:58]: So I believe that in the long run, it needs to be vibes and trust until you put it into a workflow and evaluate it for that workflow and measure it and have your own sense of, like, how it does on the exact workflow that matters. And our job is to keep moving the nines of reliability. This is like an ever-present part of o- of what we need to be doing as a company, and we need to do everything to have people know that this is something we care so much about. Like, if we wanted to, we could have released Jev, like, a year and a half ago if we wanted it to be dumb.The Bitterest Lesson: Tasks, Data, and North StarsSwyx [00:20:34]: Oh.Diogo Almeida [00:20:34]: It. Like, the. My bitterest lesson, right? Like, architecture and Yeah.Swyx [00:20:40]: I'll bring it upDiogo Almeida [00:20:40]: Hell yeahSwyx [00:20:41]: Since you, since you talked about it, here.Diogo Almeida [00:20:43]: Hell yeah. T- like, Sutton says that algorithms beats compute very roughly. Data matters way more than compute, obviously. And doing the right task, having the North Star is the hardest, most important thing. This has happened, in LLM land twice so far, right? Maybe 2.2 times. There's RLHF, which, like, shifted the task to instruction following. No one realized that was possible. RLVR did, like, a tiny little, like, edit to the, to the direction, and now us, right? RLCD. We have a new task, and the goal is, programs in the loop. And yeah, data matters soSwyx [00:21:28]: RightDiogo Almeida [00:21:28]: Unbelievably much.Swyx [00:21:29]: SoDiogo Almeida [00:21:29]: Like, I can't, I can't emphasize it less.Swyx [00:21:31]: Yeah, you consider yourself a data lab rather than, like, a model lab. Is thatDiogo Almeida [00:21:35]: AbsolutelySwyx [00:21:35]: Something. That's the wording you guys use?Diogo Almeida [00:21:37]: Yeah. We are. We will always, like, care so much about data. To me, model capabilities means data. Data is so unbelievably complicated, and that is what gets nines. Like, you have no idea how much data can shift everything. Data is so important.TypeSafe as a Data Lab and Synthetic Data StrategySwyx [00:21:57]: Yeah.Diogo Almeida [00:21:57]: Holy crap. So if people are looking for a job, we are hiring infinite data people, actually infinite.Swyx [00:22:04]: What is a good data person? Like, clearly somebody who cares about reading through the transcripts of, whatever. You've said, for example, that y- all your data is synthetic.Diogo Almeida [00:22:15]: Yep.Swyx [00:22:15]: But that's only, like, the scratching the surface, right?Diogo Almeida [00:22:18]: Yeah.Swyx [00:22:19]: Like, it's not. Like, synthetic, so what, right? Synthetic, but we have people with a lot of taste and a lot of care looking at, looking at these, articulating what's wrong, going back, regenerating. Is that what a good data person is these days?Diogo Almeida [00:22:31]: Let me try to figure out how to. Like, it's, it's super complicated, and like, I literally onboard the data people with a Talk that I assume is longer than this podcast will end up being. So I will try to say, like, the high level of it. So number one, we don't do the kind of synthetic data that people ki. Well, I'll do. Actually, number is zero. Data and synthetic data depends on your task. Like, the shape of your data. The shape of your task changes the data. Like, RLVR's data is kind of environments, right?Swyx [00:23:03]: Yes.Diogo Almeida [00:23:04]: RLHF's is the human feedback? Each task has its own unique kind of data, and we, of course, have our own unique kind of data, right? So number one, we have that. Number two, the thing I. The reason why we don't want to train on our users' data, even if we could, right? Like, we could probably ask for any terms right now, and it will. We. I don't know if it would make a difference. We truly don't want that, because no matter what, the real-world data has so much bias. There's, like, a power law of, like, people, like, asking the same things where you'll end up, like, overfitting to it and like, fracturing to it and all of that. And number two, we are, like, aiming for, like, a complete sci-fi future years from now where, like, these models are going to be, like, the general infrastructure, layers and layers and layers and deep down the stack to, like, things people can't even imagine. Like, I would like to think of our model, like, kind of like, UDP as LLMs and TCP as our models. All sorts of stuff can be built on top of that, and we need to be able to nail those futuristic use cases such that software developers can actually build that futuristic stuff. And the way to do that is even if we had all of the data of the present, we would just overfit to the present, and then it wouldn't work. What we need is to, like.Diogo Almeida [00:24:21]: It almost feels like a. Like, they're the artists? They study this cognitive core. Our cognitive core is, like, way less jagged than anyone else's. And then they find the jaggednesses, and then they address them surgically in a way that. And you can never perfectly do this, right? But they do it in such a way that it addresses it in every single possible, like, dimension, past, present, future.Swyx [00:24:45]: The general case rather than the specific case.Diogo Almeida [00:24:47]: Exactly. And like, that requires a lot of intelligence every time.RLCD vs. RLHF: Defining a New TaskSwyx [00:24:50]: Okay, so we mentioned a little bit. You sort of criticized my thinking as r-- like, very RLVR influence, which is, like, very fair. Let us actually mention RLCDDiogo Almeida [00:24:59]: OohSwyx [00:24:59]: Which obviously you have some secret sauces to our knowledge. You've never actually published a paper or anything like that on it. No, right?Diogo Almeida [00:25:05]: No, not yet.Swyx [00:25:06]: But like, what should people get from this? Like, what. Can you give people some confidence that you're just not just making up jargon for the sake of sounding cool, right? Like, one thing for me is, like, calibration I do think is a. To me, like, well understood because we've covered it in. On the podcast.Diogo Almeida [00:25:22]: Yeah.Swyx [00:25:22]: But I don't know what you mean when you say RLCD versus what people are familiar with.Diogo Almeida [00:25:26]: It's a great question.Swyx [00:25:27]: Yes.Diogo Almeida [00:25:27]: And actually, I will give a related question.Swyx [00:25:29]: Okay.Diogo Almeida [00:25:29]: What is RLHF?Swyx [00:25:31]: Okay.Diogo Almeida [00:25:31]: Right? And actually, RLHF means multiple different things, right?Swyx [00:25:34]: Okay.Diogo Almeida [00:25:34]: Like, there's the RLHF of the original. I think it was, like, Paul Christiano teaching a robot to backflip or something like that. Wasn't there somethingSwyx [00:25:42]: Was that it?Diogo Almeida [00:25:43]: That was the originalSwyx [00:25:44]: I referenced the PPO paper, but I don't know.Diogo Almeida [00:25:46]: And so PPO was not necessarily from human feedback, if I recall.Swyx [00:25:51]: Okay. That's trueDiogo Almeida [00:25:52]: But I b- I believe it was, like, an OpenAI alignment work that could teach hard to specify outputs, like a backflip. I'm not 100% sure. And then there was actually learning to summarize. This was work, by a bunch of the team that helped with, instruct-- and co-authored, the instruction following paper, which was teaching, doing PPO on language models.Swyx [00:26:15]: This is the, sorry. I'm trying to, tryingDiogo Almeida [00:26:19]: YeahSwyx [00:26:19]: Trying to manipulate this thing. This is 2017.Diogo Almeida [00:26:23]: Yeah.Swyx [00:26:23]: Right.Diogo Almeida [00:26:23]: I'm not 100% sure, but like, that looks quite right.Swyx [00:26:26]: Yeah.Diogo Almeida [00:26:26]: If it has, like, a robot doing backflips or something like that might be it. Yes. Okay, cool. I guess I got it right. Hell yeah.Swyx [00:26:35]: There you go.Diogo Almeida [00:26:36]: Yeah.Swyx [00:26:36]: That's the one.Diogo Almeida [00:26:36]: So the idea was can, like, can you do, like, ill-specified things with it? So that's, like, version one. Version two was, the learning to summarize work, that, like, OpenAI did, which is actually, like, PPO on language models to do something somewhat ill-specified. This is, like, another thing that people refer to as RLHF Which I did not co-author.Diogo Almeida [00:26:57]: Oh, Dario's there. Cool. Hell yeah.Swyx [00:27:01]: And Radford.Diogo Almeida [00:27:02]: Yeah. Shout-outs to Alec and Ryan. Love them.Swyx [00:27:04]: Yeah.Diogo Almeida [00:27:05]: But the thing that I refer to RLHF is the, Oh, man.Diogo Almeida [00:27:13]: I'll get toSwyx [00:27:14]: You have comments on that, yeah.Diogo Almeida [00:27:15]: I have comments on that paper, but like, we're so many, tangents deep.Swyx [00:27:18]: Yeah.Diogo Almeida [00:27:18]: So the thing that really got. To me, the thing that I'm calling to RLHF is the task of instruction following. It's not about the PPO. That part doesn't matter. It's about, like, setting a North Star of this is a valuable direction. It's kind of like the Bitris lesson North Star.Diogo Almeida [00:27:34]: And for us, RLCD is this new task. And it is not. I don't see it as jargon. Like, I try to communicate with precision. It's just that, “Hey, here's another North Star.” Just like DPO and all of its, like, descendants also do RLHF, despite not using the algorithm in that paper.Swyx [00:27:55]: And so clear- clearly stating the North Star is, being program- programmable AI is one, word that I really catch onto, removing the human in the loop,Diogo Almeida [00:28:06]: YesSwyx [00:28:06]: From. Because RLHF is tuningDiogo Almeida [00:28:09]: YesSwyx [00:28:09]: For this so that you can automate everything.Diogo Almeida [00:28:11]: Yes. Everything that makesSwyx [00:28:13]: Did I miss anything else in the, in the thesis of, like, what the North Star is?Diogo Almeida [00:28:17]: There is. That is. That is right. I'm overly nuanced in my communication. The one nuance is that we need to be practical. We need to be aware of what language models can do really well. Like what AI can do.Diogo Almeida [00:28:30]: Right? Like, there could be programmatic types that are, like, sick AF, but if you. If the technology is not ready for it to. It's not a tragedy if that's not out in the world.Why Programmable AI MattersSwyx [00:28:41]: Yeah.Diogo Almeida [00:28:42]: But to me, like, the pre-Jev world was a tragedy becau-- it sounds arrogant. Hear me out.Swyx [00:28:49]: No. I strongly believe you.Diogo Almeida [00:28:50]: Cool. It sounds arrogant, but like, I felt this way since long before I even had a company.Swyx [00:28:54]: Yeah. I can, I can vouch that,Diogo Almeida [00:28:56]: Yes, I've been talking about this for so longSwyx [00:28:57]: You said this at All Around Her for, like, three years.Diogo Almeida [00:28:58]: Yeah, I've been talking about this for so long. And I've been saying it because I thought it would have been easier. They say they do not do things because they. It. They're easy. They. It's ‘cause they thought it was easy, soSwyx [00:29:08]: Yeah, exactlyDiogo Almeida [00:29:09]: Something like that. I thought it. This whole project would take a week.Diogo Almeida [00:29:13]: And I was unbelievably wrong. So I am so sorry to everyone at OpenAI that I thought. I was like, “Man, I'm solving this right now.” but like, I think that the tragic thing is when. Well, I think overpromise, underdeliver is tragic too. And like, AI is super extreme on that axis. And I think RLVR is, like, the main. Well, both RLVR and RLHF are extreme perpetrators of this.Diogo Almeida [00:29:40]: But like, it. To me, it's like it's just there's just so much potential there. Like, AI is clearly so smart. I l- smart. I love this in my talks, when I ask people, like, “How can AI be so unbelievably smart? How can we, like, solve millennium prize problems in math, but still not automate even the most basics of works?” Like, really basic rote stuff that, like, the. It d- it doesn't take, like, extremely smart people to do this. It's not a satisfying job. Like, there's other things these people could be doing, but yet we need them to do, like, this ba- like, super basic- non- unsatisfying stuff because, like, we can't automate it yet, but we have this, like, supercharged engine of automation that just does not have, like, the right plugs and stuff to plug into all of this economically valuable work. And like, if the whole company of TypeSafe disappears, like, maybe it'll take, like, a year or two for people to, like, truly catch up. I actually don't know how long it'll take. If model quality matters, then we are gonna be in a very good position for a long time. But it, like, it's done, right? Like, there, like, this has changed the path of, like, technological history.Swyx [00:30:49]: Yeah.Diogo Almeida [00:30:49]: And like, we will be exploring that space as a field.Swyx [00:30:53]: Yeah. I think, I definitely agree with that. You've created possibilities. So I think, if I can paraphrase so that people can un- also understand, you should not take the success of TypeSafe and Jev as just like, “Well, that is a new model type. Now we're done. We go back to business.” Like, no. Like, actually, there's, there are, like, five other model types that you should be exploring and like, let a thousand flowers bloom.Diogo Almeida [00:31:15]: Absolutely.Swyx [00:31:16]: Right?Diogo Almeida [00:31:16]: Like, early internetSwyx [00:31:17]: And some of that, some of which you will probably also build.Diogo Almeida [00:31:18]: Of course, yes.Swyx [00:31:19]: Yes.Diogo Almeida [00:31:19]: Early internet energy. I think it's back to tech utopia. It's no longer like, “Oh, man, like, sometimes my coding agents work, but the, all of the best ones are hoarded internally.”Swyx [00:31:29]: Yeah.Diogo Almeida [00:31:30]: Right? It's like creation is back on the menu.Diogo Almeida [00:31:34]: ? Though it's gonna be a wild-ass world, and buckle up.Diogo Almeida [00:31:38]: It's. And I'm so jazzed about that.Manifesto, Launch Strategy, and Early Internet EnergySwyx [00:31:42]: Yeah. And now you have the funding and the momentum to do whatever you envision there, which I, which I think is, like, very gratifying to see you have after, so long of saying these thingsDiogo Almeida [00:31:53]: YeahSwyx [00:31:54]: But actually show the world.Diogo Almeida [00:31:55]: I know. I just. Such a, such an interesting thing to be a tease the whole time. Like, my talk, like, felt like it was a cliffhanger ‘cause I didn't say how the automation would occur.Swyx [00:32:05]: Yeah.Diogo Almeida [00:32:06]: Sean reviewed our manifesto And he's like, “It's a little bit vague in these parts.”Diogo Almeida [00:32:12]: And like, “What's step one? What is, what is the intelligence model?”Swyx [00:32:16]: Well, I asked you for model, and you were like, “Yeah, model coming.”Diogo Almeida [00:32:18]: Yeah.Swyx [00:32:18]: And like, Well, I just, I mainly objected to the word composable But build prod.god is fantastic.Diogo Almeida [00:32:24]: Thank you.Swyx [00:32:24]: Yeah.Diogo Almeida [00:32:25]: I. We've really rallied around that. I'd like to think we're not entirely a cult like some companies are.Diogo Almeida [00:32:32]: But like, we are, like, jazzed about what we're doing, and like, we are. Like, my brand is being practical, and like, we are all, like, so super-duper practical.Swyx [00:32:42]: Yeah.Diogo Almeida [00:32:42]: It's really great.Swyx [00:32:43]: Yeah. So here. And by the way, here is the step, the secret master plan, right?Diogo Almeida [00:32:47]: Yep.Swyx [00:32:47]: Shape, the shape of machine-native composable AI.Diogo Almeida [00:32:49]: It was your idea to make a secret master plan, soSwyx [00:32:51]: It's a, it's that Elon thing. When he started TeslaDiogo Almeida [00:32:53]: YeahSwyx [00:32:53]: He was like, “Here's what we'll do.”Diogo Almeida [00:32:54]: But I did. Yeah. I'm giving official credit to you.Swyx [00:32:56]: Oh, thank you. Thank you, thank you.Diogo Almeida [00:32:56]: Yeah.Swyx [00:32:56]: Thank you. But like, you should've told me your, you're also gonna do this model launch, ‘cause you, like, you told me, you told me half of the story, and then the other half, you didn't have the doom demo at the time.Diogo Almeida [00:33:08]: Yep.Swyx [00:33:08]: You didn't have any numbers to give me.Diogo Almeida [00:33:10]: Yep.Swyx [00:33:10]: I was like, “what?”Diogo Almeida [00:33:11]: Well, the problem is I don't believe in benchmarking.Swyx [00:33:13]: Exactly.Diogo Almeida [00:33:14]: Right?Swyx [00:33:14]: Exactly.Diogo Almeida [00:33:14]: So like, it is a thing that you need to feel, and like, I think that this is the way to build long-term trust, even though it, like, hurt, it hurt us a, us a lot? Like last year when we did fundraise, no one believed us.Diogo Almeida [00:33:27]: ? Like, and they wanted just benchmarks and stuff, and we're like, “We're not gonna do that. We are principled. We're gonna stand by our guns. That rewards bad actors. I don't give a s**t, like, what you want. Like, this is who we are, and we are standing by that.” So Sorry. It's notSwyx [00:33:43]: No, yeah. Well, and in some ways, I think, like, choosing the hard path, it. But you end up making the company that you wanna work in.Diogo Almeida [00:33:49]: Yep.Swyx [00:33:50]: Right? Otherwise, if you sell out, then you're just working in, like, OpenAI but with my people, right? Which is like.Diogo Almeida [00:33:56]: Yeah. Yeah. Like, I'm, I don't have too many regrets on that, obviously.Swyx [00:34:01]: Yeah.Diogo Almeida [00:34:01]: Like, it worked out so unbelievably well. And like, I, The. I was emotional last night when I was talking about, like, the reasons I left OpenAI, and because, like, it actually had to change my wording after the launch. My phrasing was, “If an AI winter did happen and I did not do every f*****g possible thing I could to, like, avert that, I would see myself as personally responsible both for, the RLHF direction, which I think really widened overpromise versus under-deliver, and also not going all in on this because I think this is, this is where value is going to just be, like, printed.” So. And it was really cool because I feel likeDiogo Almeida [00:34:47]: The AI winter I'm worrying about is averted. Like, AI will be useful. It'll be used for automation.Diogo Almeida [00:34:53]: It's been less than a week, and like, the numbers are already undeniableSwyx [00:34:57]: YeahDiogo Almeida [00:34:57]: That it's, like, being used for real work, and like, there's. It's, it's the Wild West. Yeah.Launch Traction, Tokens, Rate Limits, and Developer UsageSwyx [00:35:03]: Yeah. Can you sh- just if you have top of your head, what numbers are you seeing? Like, what's, what's, like, signups? Like, whatever you can share.Diogo Almeida [00:35:11]: I'm actually not super on top of everything. Like, the team is the ones who are telling me all of these things.Swyx [00:35:16]: Yeah, and I'm sure it's, like, changing every day, right?Diogo Almeida [00:35:17]: It's, it's,Swyx [00:35:18]: But likeDiogo Almeida [00:35:18]: It's kinda nutsSwyx [00:35:19]: If there's a milestone that you're like, “Well, yep, that's one thing we were hoping for. We reached it.”Diogo Almeida [00:35:23]: I will say a milestone that we've passed is tokens per day.Swyx [00:35:27]: Nice.Diogo Almeida [00:35:27]: And this is not, like, fleeting tokens per day.Swyx [00:35:32]: Yeah.Diogo Almeida [00:35:32]: This is, like, even at night, like, it's constantly training, so machines are calling it and not just people trying things out.Diogo Almeida [00:35:39]: So that is, That is so cool. A trillion tokens a day is a lot.Swyx [00:35:45]: Yeah.Diogo Almeida [00:35:45]: So surpassing that is awesome. Signups to me don't really matter. And actually, this was, like, a bit of a mistake we made, if I'm, like, totally honest. People on Twitter were calling us, like, marketing geniuses and all of that, and that was just us. We don't have a marketer. Also hiring. And we were just being our genuine, goofy, like, irreverent selves, and we were, we were just, like, offboarding people off the waitlist so hard. - Our platform team is so unbelievably cracked. I think we have more n- up nines of uptime than Anthropic while having the most Unprecedented launch ever. Like, that is kind of nuts, soSwyx [00:36:21]: YeahDiogo Almeida [00:36:21]: Like, props to them.Swyx [00:36:22]: Yeah.Diogo Almeida [00:36:23]: And the thing we didn't realize. So number one, waitlists, waitlist sign-ups don't matter for, like, a developer platform, in my opinion? I would guess that a large number of them are not even developers. So they go in, they try some queries, and a lot of people don't get it because they are not programming, right? Like, they're just like, “What? This is not a chatbot. Where's my ChatGPT 2?”Diogo Almeida [00:36:45]: Right? But if, like. I haven't exactly calculated this. My sense is that if every single human being in the world, like, just wrote a couple of queries, that would be a rounding error compared to, like, one power user's for loop that is just, like, creating value.Swyx [00:37:01]: Yeah.Diogo Almeida [00:37:01]: And the thing we are-- didn't realize with the waitlist is, like, we could just w- off-board anyone off the waitlist. It doesn't matter. The scary part is rate limits. And then once people start getting value from that, then they just want tons and tons of rate limits because this is what software is, right? Like, you spend effort upfront to specify your rote task, and then this rote task creates more value than it takes to put in. And then now that you have thatSwyx [00:37:25]: Set it and forget, yeah.Diogo Almeida [00:37:26]: Exactly, yeah. You run it in the background. You make it a dependency, to, like, other things. You can make, like, higher level stuff. And like, you just create so much value in the world. Early internet people probably did not imagine, like, the wonder of early 2000s internet, which is still not early internet. But like, it's, it's through, no offense, composabilitySwyx [00:37:47]: NoDiogo Almeida [00:37:47]: That all of the crazy stuff happens, and I just really wanted to emphasize that in our manifesto. We are going for emergence. We are going for, like, being the catalyst. We're wanting to empower people, and we are going to do whatever we can for that, be it, like, Discords in our town hall with me wearing a garbage bag or not.Swyx [00:38:05]: And podcasts and Diogo Almeida [00:38:08]: Hell yeahSwyx [00:38:09]: Getting all that.Diogo Almeida [00:38:09]: Absolutely.Swyx [00:38:09]: Like, ‘cause I want the long form, right?Diogo Almeida [00:38:11]: Yeah.Swyx [00:38:12]: It is like, yes, we'll get past the, some of the superficial things, and then we'll go deep andDiogo Almeida [00:38:15]: Hell yeahSwyx [00:38:15]: And people will really trust and understand your mission and like, the people that, will resonate that will end up joining you or, buying you. Or No, but sorry, as a, as a customer.Diogo Almeida [00:38:27]: Oh, as a customer.Swyx [00:38:28]: As a customer, as a customer.Diogo Almeida [00:38:28]: Okay, yeah. That was funny. I'm sorry.Swyx [00:38:30]: Sorry. I didn't, I didn't mean to say that. But no, any-- one version, one very flattering version of this, like, 36 million views of your launch video.Diogo Almeida [00:38:37]: Cool. Up to 38 now.Swyx [00:38:39]: Yeah, rounding error.Diogo Almeida [00:38:40]: Yeah.Swyx [00:38:40]: Navio still has got 74. Fable 5 got 57. So like, as far as, a- and I didn't, I didn't do the stats for, like, original ChatGPT, likeDiogo Almeida [00:38:48]: YepSwyx [00:38:49]: Which there was no video.Diogo Almeida [00:38:50]: Yep.Swyx [00:38:50]: So like, up there, right?Diogo Almeida [00:38:52]: Yep.Swyx [00:38:52]: Like, as far, as far as, like, if you were to launch a Neolab in 2026, I think you're, like, number one right now, which is, like, pretty crazy.Diogo Almeida [00:38:58]: Yeah. Well, I actually would rather. I do have the shirt, like, your favorites Neola-- favorite Neolab's favorite Neolab.Swyx [00:39:05]: Huh.Diogo Almeida [00:39:05]: I don't give a s**t about being a Neolab. I think being a Neolab. Actually, we have a lot of, like, swag that's being a parody of a Neolab. One of them, one of them I have is, like, Neolab with product, which actually is not a Neolab. Like, I don't care about that, really.Swyx [00:39:20]: Yeah.Diogo Almeida [00:39:20]: What I care about is being a reliable dev platform. So Swyx [00:39:23]: YesDiogo Almeida [00:39:24]: Appreciate the comparison, but likeSwyx [00:39:25]: YeahDiogo Almeida [00:39:25]: Hopefully we transcend past them and we go back into, like, a thing-- like, a revolutionary moment for developers and like, this stable thing that people can rely on and trust.Reliability, Robustness, and DeterminismSwyx [00:39:35]: Yes. To that end, I think that's one thing that really impressed me about you guys is that, yes, you do talk about reliability. I thought it was mostly about calibration, which, like, we talk about RLCD. But actually it's also about just, like, uptime and scalability and all those things, right? They're, they're all sort of the kind.Diogo Almeida [00:39:55]: And nines.Swyx [00:39:56]: And nines.Diogo Almeida [00:39:56]: It's, likeSwyx [00:39:57]: Which uptime is, in my opinion.Diogo Almeida [00:39:58]: Oh, but that's part of it. But like, there's reliability in, like, how intelligent the thing is. Like, how consistently does it do the thing that you want? And I think that, like, the big reasoning models are very smart. In my opinion, they still lack reliability. I think there's many use cases where you-- they look like they should be smart enough to automate their work. There is economic incentive to automate that work, yet still they're not reliable enough as, at an intern because they're optimized for different things. And so like, I think that there's the reliability of being able to, like, trust the outputs. And also we are. Like, there are dimensions of reliability that we are not yet at that I'm, like, so excited by.Swyx [00:40:38]: Yeah.Diogo Almeida [00:40:38]: Like, I want to automate the easy work before the hard work? Like, I think that's just a common sense thing to do. But to me, we will be sufficient. I don't know if there's such thing as sufficiently reliable, but I wanna get so good that people don't even need to try the model to know that it'll work. It's like, that's like what flow state is in programming, right? Like, I'm just, like, writing queries because I need intelligence in here. And like, when. For non-trivial branching, I can just write it in like a, like a type-safe System 1 query and then get the results out of it and it just branches accurately. Like, that would be so good. Like, that's the. That is the dream.Swyx [00:41:12]: Yeah.Diogo Almeida [00:41:12]: And that is, like, going to be, like, a long slog.Swyx [00:41:16]: Yeah. We're gonna go into your API design in a little bitDiogo Almeida [00:41:19]: OohSwyx [00:41:19]: Just to give people examples and like, maybe paths not taken, that kind of stuff.Swyx [00:41:23]: One thing up the front that I do wonder about in terms of reliability is I noticed that there's no seed. There's no, And so basically, same input, do I always get the same output?Diogo Almeida [00:41:34]: SoSwyx [00:41:36]: And if not, why not?Diogo Almeida [00:41:37]: Oh, great question. So this is actually, like, a common question we have between. So reliability is actually a catchall. Like, whenever AI can't automate something, it's due to some form of reliability. Could be, like, type safety. It could be determinism. It just could be, like, it's, it's jagged, right? So reliability is a catchall. I just think that it's also a catchall for, like, what the North Star is. Re- determinism is, like, same inputs, same outputs. I do believe that this is, like, slightly interesting for unit tests, but I believe that to be the wrong North Star. I believe robustness is what peopleDiogo Almeida [00:42:16]: I don't wanna tell people what they really want, ‘cause that would be a little arrogant of me.Diogo Almeida [00:42:19]: I believe that is, like, the more important property. You want, given similar inputs, get similar outputs. And it's kind of wild how unreliable LLMs are.Diogo Almeida [00:42:31]: Like, a way that we test this is you put, like, UUIDs in, like littleSwyx [00:42:36]: YeahDiogo Almeida [00:42:36]: I think they're called nonces In the prompt. And what you want is similar outputs from all of those, ‘cause it's truly semantically the same question, and that is the part where you really want. Th- like, that robustness is where, like, people get, like, burnt with AI making decisions. So I think that is the. A super-duper important property. We could also have determinism. That is, that is a thing that can be available. As far as I can, like, mentally model for programmers, like, it, I- it could be valuable for some use cases, so like, please educate me, in comments or view. But my. In general, it's easy. Determinism is something you can, like, trade off for better cost. Like, we are, we are constantly wanting to be on the intelligence per dollar frontier. We are doing, like, absolutely disgusting things to be there. Like, this is,Diogo Almeida [00:43:32]: I shouldn't say this, but no one's here to stop me.Swyx [00:43:37]: If you s- you sign off on your own PR.Diogo Almeida [00:43:40]: That is not how it works at this company. I believe for this week, my chief of staff, Kay, is the most powerful person in tech.Swyx [00:43:49]: Yeah. And shout-out to Kay for organizing this.Diogo Almeida [00:43:50]: Holy shSwyx [00:43:51]: Yeah.Diogo Almeida [00:43:51]: Holy s**t. She is so f*****g competent and powerful. She's incredible.Diogo Almeida [00:43:58]: She sucks. Don't poach her. But so I try to be a bit more filtered, but like, people are telling me, “Don't call it a Frankenstein's monster of models,” but because that has, like, negative implications. I think Frankenstein's monster was, like, the good guy in this whole. It was innocent, right? I didn't read it. Okay.Diogo Almeida [00:44:18]: I'll, I'll confess. Okay. That. Well, one facial expression, ISwyx [00:44:21]: This is aDiogo Almeida [00:44:21]: My cards on the tableSwyx [00:44:21]: Decent Jacob Elordi movie if you wanna seeDiogo Almeida [00:44:24]: ISwyx [00:44:25]: The adaptation. Anyway.Diogo Almeida [00:44:26]: The. You have no idea how little time I have right now.Swyx [00:44:28]: Yeah.Diogo Almeida [00:44:29]: My priorities are sleep?Swyx [00:44:31]: Developers.Diogo Almeida [00:44:32]: Developers, yes. Developers. But yes. It. We do, like, absolutely disgusting things to be on the Pareto curve of intelligence per dollar, and we are going to keep doing that.Swyx [00:44:47]: Yeah.Diogo Almeida [00:44:47]: We're gonna be doing crazy-ass stuff, and I think people really need to think outside of the box. Like, part of the reason we're surprising is, like, people Are thought inside the box, and we continue to do that. As of right now, we are obviously the best at this, and we want to continue being the best at that whole thing.Swyx [00:45:05]: Yeah.Diogo Almeida [00:45:05]: So Wait, where did, where did we tangent from?Swyx [00:45:07]: No. SoDiogo Almeida [00:45:08]: YeahSwyx [00:45:08]: I asked you about, will you have seeds and determinism?Diogo Almeida [00:45:11]: Oh, yes. SoSwyx [00:45:11]: And then you basically defined reliability and likeDiogo Almeida [00:45:14]: And robustnessSwyx [00:45:15]: How you see it. Yes.Diogo Almeida [00:45:16]: But like, determina- likeSwyx [00:45:17]: I have a robustness example that's, that's, real quick I can show you.Diogo Almeida [00:45:19]: I would love that. I will just say one thing.Swyx [00:45:21]: Yeah.Diogo Almeida [00:45:21]: We can make a deterministic model.Swyx [00:45:22]: Exactly.Diogo Almeida [00:45:23]: Like, we're hap- if people can convince us that is a valuable thing to doSwyx [00:45:27]: YeahDiogo Almeida [00:45:27]: And we don't have a gigantic GPU shortageSwyx [00:45:29]: YeahDiogo Almeida [00:45:29]: We can happily make all of these models. We live to please. And rev- and revolt, revolute,Swyx [00:45:38]: You will throw over everything, except you'll do it in a nice way.Diogo Almeida [00:45:41]: Yeah.Swyx [00:45:41]: And findDiogo Almeida [00:45:42]: So like, determinism could be on the cards.Swyx [00:45:44]: Yeah.Diogo Almeida [00:45:45]: It just gets you less intelligence per dollar.Swyx [00:45:46]: Yeah. Well, just having seen the trajectory of OpenAI and Anthropic, you will. Just trust me now that you will be peer pressured into doing it. So like, just people will want it even if they. If you tell them they don't need it. They'll still want it. So like, yeah, that's the TL;DR of that.Diogo Almeida [00:46:01]: Okay.Swyx [00:46:02]: Yeah.Diogo Almeida [00:46:02]: I will love to. Maybe one day we will see how that happens.Swyx [00:46:07]: Yeah.Diogo Almeida [00:46:07]: I've been told I'm, They say that part of our brand is being unshakeableSwyx [00:46:13]: HuhDiogo Almeida [00:46:13]: And they say that's just the nice way of saying stubborn.Swyx [00:46:15]:

P.M. Edition for Sept. 18. In a letter to investors today, legendary investor Warren Buffett said he would step down as chairman of Berkshire Hathaway. WSJ deputy markets editor Justin Baer discusses why it's happening now, and the details of Buffett's long-held succession plan. Plus, Disney has hired the head of artificial-intelligence company Character.AI to be its first chief technology officer. We hear from Journal reporter Ben Fritz about how this fits into CEO Josh D'Amaro's strategy. And President Trump says he's banning CNN, MS Now and Politico from the White House over their coverage of his administration. Alex Ossola hosts. Sign up for the WSJ's free What's News newsletter. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.