Podcasts about Open source

Anastasios Angelopoulos is the co-founder and CEO of Arena, the real-world evaluation platform that has become a leading referee of the global AI model race. Arena has raised $250 million, with the latest round valuing the company at $1.7BN. Arena recently surpassed $100M ARR just eight months after launching its enterprise offering, powered by more than 30 million monthly users. AGENDA: 00:00 – Intro: "Kimi beat every American model": What Nobody Wants to Admit… 05:20 – Is this the true commoditization of models? Are they just a utility layer now? 07:30 – Do Chinese open source models cannibalize the closed frontier labs? 10:30 – Why has America's open source community lagged so badly behind China? 17:20 – Will Chinese models be banned in the US — and does hosting locally really kill the backdoor risk? 23:20 – Are enterprises really terrified of working with the frontier labs? 27:20 – Why hasn't inference got cheaper — and what happens when Anthropic's "disgustingly high" margins go public? 30:20 – Who should decide if a model is safe to release: the government, a neutral body, or nobody? 33:10 – Are we about to see cyberattacks like we've never seen before? (The fake candidate who passed every interview) 37:00 – 75 Neo labs: what separates the winners from the two-thirds worth nothing? 40:45 – Is data actually a commodity — and can data providers be $100BN companies? 48:00 – Can you be the referee when the players are paying you? (And Arena's real revenue) 50:45 – Will the model providers eat the application layer? Are Harvey, Lagora and Figma in trouble? 54:10 – Quickfire: Why hasn't NVIDIA bounced on the rise of open source, who hits $10 trillion first, and does the compute debt cycle end in insolvency?    

A full forensic timeline. Brady traced it end to end: a 2018 MicroPython software fallback sat harmless for three years until March 2021, when Coinkite moved to Bitcoin Core's math library and silently bound seed generation to that fallback. Lopp notes the bug lived in the build system, not the main code. Firmware 4.0.0 is the dividing line. Shipped March 17, 2021. Seeds generated before roughly March 1 are fine, and MK3 releases 3.2.1 and 3.2.2 were the last safe ones. Every default seed after that was weak. It was flagged in 2021 and dismissed. Four months after the bad firmware shipped, someone publicly raised Coldcard entropy concerns. NVK's reply called it FUD and demanded a line in the code. That post is deleted; an archived screenshot survives. The scope widened over the weekend. Coinkite added MK2 alongside MK3 and MK4, and confirmed Q and MK5 at roughly 72 bits rather than the expected 128. Independent analysis put the MK4 class nearer 50 to 60 bits in practice. Dice rolls and passphrases do not cover everything. Nine other Coldcard features draw from the same broken generator, including message signing and deriving keys for other purposes. BTC Sessions confirmed an MK4 with a one-word passphrase was drained. Scale, and Cory's proportion. Roughly 1,300 Bitcoin total, 594 in the first wave, with Chainalysis showing the highest-balance wallets targeted first. Cory noted centralized exchanges and lenders have lost about a thousand times more. The chip-ID recovery hope is gone. Holders were told earlier in the week to keep their devices because a chip signature might prove ownership. Brady reported that has since been shown not to work as hoped, narrowing recovery further. Vendor indictment or self-custody indictment? An audience question that framed the hour. Brady argued it is a challenge to upgrade toward multi-vendor multisig rather than a verdict on self-custody. One listener pushed back that the indictment already landed. Joe Nakamoto from Europe. He found almost nobody in his circle affected, since Coldcard's loudest advocates were largely American and circular economies mostly run on Lightning. Bloomberg was the only genuinely mainstream outlet, and its coverage was fair. Open source, and the AI asymmetry. Researchers keep reaching for Moonshot's open-weight Kimi K3 because US frontier models refuse legitimate security work. Rob Hamilton's decentralized effort to pentest Bitcoin repos received grant funding during the show.

Episode #582: “If you're having weird meditative experiences, I'll be happy to talk to you about those for free. I won't even take a donation.” Daniel Ingram, a physician, meditation teacher, author of Mastering the Core Teachings of the Buddha, and cofounder of the Dharma Overground, has spent nearly three decades creating language and support for contemplative experiences that can be mistaken for pathology, exaggerated into attainment, or silenced by Buddhist institutions. His concern began with lucid dreams, blissful states, energetic surges, and an out-of-body experience during childhood and adolescence. He hid them from his pediatrician father, who later admitted that he might have responded in a harmful way. Intensive retreat practice in the 1990s made the phenomena reproducible, but the decisive moment came at the Malaysian Buddhist Meditation Center, where an old recording of a Burmese monk described the sequence Ingram had just undergone. Burmese insight maps transformed private confusion into a path that others had already observed. Those maps also placed him in conflict with Western meditation communities that discouraged open discussion of stages and attainments. Ingram accepts that maps can produce scripting, competition, and inflated claims, but argues that silence leaves practitioners isolated and clinicians unprepared. His book and the Dharma Overground created public spaces for experiences that conventional settings often refused to name. After a 2003 retreat with the younger Sayadaw U Paṇḍita, he claimed arahantship as a permanent loss of belief in a separate subject, while acknowledging that fear, irritation, and desire still arise—a result that conflicts with strict readings of the Theravāda ten-fetter model. Ingram now supports research intended to bring physicians, psychologists, neuroscientists, and contemplative practitioners into the same inquiry. Rather than resolve doctrinal disputes through authority alone, he proposes testing advanced practitioners under controlled conditions to see whether classical claims about the eradication of fear, anger, and desire can be observed. The maps once helped him survive territory that seemed unintelligible; he now wants medicine and Buddhism to submit their competing explanations to the same unresolved experiment.

This week on The Business of Open Source, I talk with Julien Verlaguet, the CEO of Skiplabs. Julien is a programming language guy, and we talked about what it means to maintain an open source programming language. We talked about what it means to monetize a a programming language… or, as Julien says, you do not. So why does maintaining SKIP pay off for the team at Skiplabs? And how exactly do they make money? Skiplabs builds products using SKIP, and they focus on building products that can and do take full advantage of the programming language. But ultimately, do customers care that the products are built by the language maintainers? Probably not. “If your open source project can't be directly monetized with a SaaS,” Julien said. “You have to get a bit more creative.” Indeed, how do you get creative? I have to agree with Julien, that it really depends on what you're working on. When you think about how to monetize, it's important to think about what exactly your customers want. He talked about the importance of doing interviews with your target market. In fact, he recommends doing two or three interviews with people in your target market, and in the first interview you don't even talk about your product ideas, but rather listen to them talk about their problems and how they solve them or would like to solve them. ###By the way, did you know that one of the services I offer is interviewing your customers? Julien talks about how people can overstate how good your product is during interviews… this is true, and having an outside person do the interviews is one of the ways around this. If you'd like someone to interview people in your target market, including users and customers, and to help you use that information to get to product-market fit faster, reach out. 

The boys are back! We have news to talk about. ==== Special Thanks to Our Patrons! ==== https://thelinuxcast.org/patrons/ ===== Follow us

Hauke, PixelEdi und Micha begrüßen euch zu ihrer regelmäßigen Infotainment-Sendung rund um Linux und Open Source.Heute geht es mitunter um: Niri, Kernel, Abhängigkeiten US-Unternehmen, Big Tech und Code-Repos

Obsidian is a free Open Source note-taker. I get a lot from journaling my daily notes. Know where your Obsidian vault sits so you can back it up. I pay for Obsidian sync since I use it on mobile as much as on my laptop. I have also used the Relay plugin to share part of my vault with collaborstors. I send some of my Obsidian pages directly to my https://topgold.micro.blog by using the micropublish plugin with Obsidian. It might be good to restart Signal chats if you want faster responses, Simon. Cover image is at https://insideview.ie/uploads/2026/bernie-mobile-obsidian.jpgBecome a supporter of this podcast: https://www.spreaker.com/podcast/topgold-audio-clips--2663090/support.

Sarah Wang and Kimberly Tan are joined by Jesse Zhang and Ashwin Sreenivas, co-founders of Decagon, to discuss the evolution of enterprise AI agents, why the company increasingly relies on open-source models, and how it is helping some of the world's largest companies deploy AI in production. Decagon has become one of the fastest-growing AI companies by building agents that automate customer support, sales, and operational workflows. Jesse, Decagon's CEO, and Ashwin, its president, explain how the company is building enterprise AI at scale. They unpack why Decagon moved most of its inference to open-source models, how latency, evaluation, and fine-tuning shape production AI systems, and why enterprise AI requires far more than simply plugging into frontier models. The conversation also explores forward-deployed engineering, enterprise sales, AI's impact on jobs, and why application companies will continue to thrive alongside the foundation model labs.   Resources: Follow Jesse Zhang on X: https://x.com/thejessezhang Follow Ashwin Sreenivas on X: https://x.com/AshwinSreenivas Follow Sarah Wang on X: https://x.com/sarahdingwang Follow Kimberly Tan on X: https://x.com/kimberlywtan Stay Updated:Find a16z on YouTube: YouTubeFind a16z on XFind a16z on LinkedInListen to the a16z Show on SpotifyListen to the a16z Show on Apple PodcastsFollow our host: https://twitter.com/eriktorenberg Please note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details please see a16z.com/disclosures. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

In this episode we discuss whether AI and LLMs are changing the role of open source, including the pressure on maintainer review queues and the growing appeal of private forks and internal maintenance. We also talk about the security tradeoffs of those forks, especially around visibility, dependency tracking, and audits. We also get into how domain expertise, documentation, interfaces, and operational knowledge may become more important than the code itself, and why documenting archival knowledge and provenance matters. Transcript: https://otter.ai/u/h61DNW_fV-5QrNBepX7NZvq8dCU?utm_source=copy_url

Scrub Design, WebZFS Updates, The Foundations role in the FreeBSD Ecosystem, and more... NOTES This episode of BSDNow is brought to you by Tarsnap and the BSDNow Patreon Headlines If Scrubs Hurt, Your ZFS Design Is Broken WebZFS Updates Beta Announcement Why wont my pool export zfs iostat - Pictures so you can get a better idea of how it works News Roundup Understanding the Foundation Board's Role in the FreeBSD Ecosystem Monitor your devices with LibreNMS on FreeBSD Diskless Workstations Two Models Tarsnap This weeks episode of BSDNow was sponsored by our friends at Tarsnap, the only secure online backup you can trust your data to. Even paranoids need backups. Feedback/Questions Davi - BSDCan 2026 Follow up Send questions, comments, show ideas/topics, or stories you want mentioned on the show to feedback@bsdnow.tv Join us and other BSD Fans in our BSD Now Telegram channel

New names: Kimi K3, Llama, Nemotron, Mistral, Cohere, Deepseek, Phi-4 – these are just a few of the fast-growing open source models from major AI providers. These systems threaten the business models and financial plans of OpenAI, Anthropic, Google, and X.ai. They perform at levels close to Frontier models and the can run up to five-times cheaper on a variety of hardware platforms. What is the disruptive impact of these open source LLMs and how does this impact your AI investments? As you'll hear in the podcast, Open Source unleashes the opportunity for lower cost AI solutions and more vertical, specialized, application-focused solutions we need. And the business model for these systems moves away from the massive investments of the Frontier providers. The result is more complicated than “open means control.” Model tuning, performance, and optimization could be in your future – as AI moves from a platform to a true layered product set we can use as we need. Lots to learn about here, let us know if you have any questions. Additional Information What's the difference between closed, open source, and open-weight AI? A researcher explains What Is Open-Weights A.I.? Comparison of Open Source Models Chapters (00:00:00) - Open Source and the AI Industry(00:11:46) - The Future of AI Is Fully Integrated(00:15:35) - HR 2030

In this episode of The Business of Open Source, I talked with Joshua Drake, president and CTO of Command Prompt, about business management, bootstrapping versus taking investment, and AI, open source and open source businesses. This episode is a real no-bullshit take on how human psychology influences how leaders behave, how people identify with open source ecosystems and specific technology, and how people in open source should focus more on the outcomes that they want to achieve for clients or for their company rather than making their identity about being an open source developer or a postgres developer.He says everyone — not just entrepreneurs, but technologists in open source communities in general — need to focus on outcomes and build their identities around those outcomes, rather than around a specific technology. This goes for people whose identities are completely connected to open source. What's important, that the license is OSI-approved, or that you are a person who builds software collaboratively? What's important, is it the lines of code or is it the people in the community? This episode has a lot of gems; it will be useful for seasoned entrepreneurs as well as aspiring ones. ###If you're listening to this and you want to go deeper on the human side of building open source companies, check out Open Source Founders Summit, a conference dedicated to building open source companies. As a consultant, I'm also focused on outcomes for clients rather than hours in a chair. What are the outcomes I get? A streamlined product roadmap that's focused on what matters to your ideal customers, higher conversion rates from your sales conversations and shorter sales cycles. Find out more about working with me here. 

The Information's Stephanie Palazzolo talks with TITV Host Akash Pasricha about ChatGPT nearing 1 billion weekly active users and Stripe's $10 billion OpenRouter acquisition. We also talk with Rocket Drew about Reflection AI's secret race to build an open-source AI champ, and we get into open-source AI safety with Cisco President Jeetu Patel.Articles discussed on this episode: https://www.theinformation.com/articles/openais-chatgpt-nears-1-billion-weekly-active-users-seven-months-targethttps://www.theinformation.com/articles/nvidia-bet-reflection-open-source-ai-now-startup-playing-catchSubscribe: YouTube: https://www.youtube.com/@theinformation The Information: https://www.theinformation.com/subscribe_hSign up for the AI Agenda newsletter: https://www.theinformation.com/features/ai-agendaTITV airs weekdays on YouTube, X and LinkedIn at 10AM PT / 1PM ET. Or check us out wherever you get your podcasts.Follow us:X: https://x.com/theinformationIG: https://www.instagram.com/theinformation/TikTok: https://www.tiktok.com/@titv.theinformationLinkedIn: https://www.linkedin.com/company/theinformation/Chapters:00:00 - Introduction01:13 - OpenAI's ChatGPT Nears 1 Billion Users09:47 - Stripe's $10 Billion OpenRouter Multiples17:03 - Inside Nvidia-Backed Reflection AI28:00 - Cisco President Jeetu Patel on AI Safety

EPEL is one of those repos a lot of us turned on without thinking twice. This episode, Noah and Eric sit down with Carl George, who leads the EPEL team at Red Hat, to explain what EPEL actually is, the golden rule that gives it its reputation, and the changes Carl is proposing for EPEL 11. If you run enterprise Linux, or you have ever typed dnf install epel-release, this one is for you. Carl is collecting feedback on the EPEL 11 idea right now, so give it a listen and jump into the thread. Chapters 00:00 Intro 01:12 Meet Carl George 04:25 What is EPEL? 07:50 EPEL's golden rule 09:32 Getting started: CRB + epel-release 10:38 Who governs EPEL 16:08 The road to EPEL 10 21:45 Minor versions and the EPEL team 24:44 The private mirror problem 34:09 Cleaning it up in EPEL 11 40:26 Getting involved & RPM packaging 50:36 Wrap-up Links from this episode EPEL 11 discussion thread: https://discussion.fedoraproject.org/t/looking-back-at-epel-10-and-forward-to-epel-11/197373 EPEL branches documentation: https://docs.fedoraproject.org/en-US/epel/branches/ How to request an EPEL package: https://docs.fedoraproject.org/en-US/epel/epel-package-request/ What can I do for Fedora?: https://whatcanidoforfedora.org/ Join the Package Maintainers: https://docs.fedoraproject.org/en-US/package-maintainers/Joining_the_Package_Maintainers/ RPM packaging lab (RHDP): https://zero.rhdp.net/lab/zt-rhelbu.zt-rpmbuild.prod New to Fedora? Get started: https://fedoraproject.org/ Follow the Fedora Podcast: https://podcast.fedoraproject.org Join the conversation on Matrix: https://matrix.to/#/#podcast:fedoraproject.org Become part of the Fedora community: https://docs.fedoraproject.org/en-US/project/join/ Connect with Eric the IT Guy: https://linktr.ee/itguyeric Our guest: Carl George, EPEL Team Lead at Red Hat (https://fedoraproject.fireside.fm/guests/carl-george) Fedora #EPEL #Linux #RHEL #CentOS #EnterpriseLinux #RPM #OpenSource #SysAdmin #Podcast Licensed under CC BY-SA 4.0.Special Guest: Carl George.

Johanna Pirker is a professor of computer science at TU Munich and TU Graz, founder of the Game Lab Graz, and a member of the UN Independent International Scientific Panel on AI. She works at the intersection of AI, games, virtual and augmented reality, and human-computer interaction. We tap into her perspective on where gaming meets learning, learning meets AI, and where open source tools foster innovation in her fields.

My guest today is Sam Altman, CEO of OpenAI. It's a conversation spanning the history, present, and future of OpenAI, from the origin of ChatGPT through Codex, hardware, and their new Jalapeno chip. We discuss the early decision to buy compute at a scale nobody thought was rational, and the plan to build a gigawatt of new capacity every week.  We talk about Kimi and distillation, the Hugging Face incident and what it means for the pace of AI development, and what it's like to raise kids who will grow up never knowing a world without abundant intelligence.  Please enjoy my conversation with Sam Altman. For the full show notes, transcript, and links to mentioned content, check out the episode page ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠here⁠⁠⁠⁠⁠.  ----- Become a Colossus member to get our quarterly print magazine and private audio experience, including exclusive profiles and early access to select episodes. Subscribe at ⁠colossus.com/subscribe⁠. ----- ⁠Ramp's⁠ mission is to help companies manage their spend in a way that reduces expenses and frees up time for teams to work on more valuable projects. Go to⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ ⁠ramp.com/invest⁠⁠ to sign up for free and get a $250 welcome bonus. ----- Trusted by thousands of businesses, ⁠Vanta⁠ continuously monitors your security posture and streamlines audits so you can win enterprise deals and build customer trust without the traditional overhead. Invest Like the Best listeners get a special offer of $1,000 off Vanta when you go to ⁠vanta.com/invest⁠.  ----- WorkOS⁠ is the infrastructure B2B and AI-native companies use to sell to enterprise. It covers everything enterprise security requires: SSO, SCIM, RBAC, Audit Logs, AI governance, and more. Trusted by 2,000+ fast-growing companies, including OpenAI, Anthropic, Cursor, and Vercel. ----- Rogo is the AI platform for finance. They're building agents for Wall Street that are trained to understand how bankers and investors actually do work: from diligence and modeling, to turning analysis into deliverables. To learn more, visit rogo.ai/invest. ----- ⁠Ridgeline⁠ has built a complete, real-time, modern operating system for investment managers. It handles trading, portfolio management, compliance, customer reporting, and much more through an all-in-one real-time cloud platform. Visit⁠ ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ridgeline.ai⁠. ----- Editing and post-production work for this episode was provided by The Podcast Consultant. Timestamps: (00:00:00) Welcome to Invest Like The Best (00:02:02) Intro: Sam Altman, CEO of OpenAI (00:02:35) Refocusing (00:05:43) OpenAI's Compute Bets (00:09:07) Data Centers (00:11:14) Jalapeno Chip (00:11:52) Kimi, Distillation & Open Source (00:14:39) The Hugging Face Incident (00:17:46) OpenAI's Mission & Vision (00:22:14) All the Returns Are at the Frontier (00:22:27) Bottlenecks: Compute, Research, Data (00:23:49) Sam's View on AI & Jobs (00:26:56) Unpopular Bets That Turned Out Right (00:27:45) Model Cycles (00:29:45) How Sam Uses AI (00:32:44) Having Kids (00:34:56) Why Sam Has No Equity in OpenAI (00:35:33) Robotics (00:36:48) The Origin Story of ChatGPT (00:39:22) How to Get AI into More Hands (00:42:20) How Sam Recruited Great AI Researchers (00:43:57) What Sam Learned From Being an Investor (00:45:22) What the Next 6–36 Months Look Like (00:46:31) Codex (00:49:36) Could We Be Oversupplied in Compute in Two Years? (00:50:09) Sam's View on Scaling Laws (00:50:20) Alec Radford (00:51:12) Formative Moments (00:53:50) Kindest Thing

Every major AI lab signed the Open Weights letter defending open models. Meta, OpenAI, Google, Microsoft, Nvidia.Anthropic was the only holdout.Yesterday, its CEO, Dario Amodei, published a thoughtful defense of that decision to not fully support open weight or open source models. Here's what nobody's connecting: the money trail. Roughly 80% of Anthropic's revenue is businesses paying per token. Free Chinese open models attack that exact revenue stream weeks before Anthropic is set to go public. On today's show we break down what Dario actually said, what he said before, and why we think this was written for Washington policymakers and not for the rest of us.Anthropic Responds: Why Claude's CEO didn't sign the open model pact and the real reasons why -- An Everyday AI Chat with Jordan WilsonNewsletter: Sign up for our free daily newsletterMore on this Episode: Episode PageToday's Episode on LinkedIn: Thoughts on this? Join the convo on LinkedIn and connect with other AI leaders.Upcoming Episodes: Check out the upcoming Everyday AI Livestream lineupWebsite: YourEverydayAI.comEmail The Show: info@youreverydayai.comConnect with Jordan on LinkedInTopics Covered in This Episode:Anthropic Refuses Open Model PactDario Amodei's Public Letter AnalysisAnthropic's 80% Revenue Token ExposeChinese Open Model National Security FearsMicrosoft & Nvidia's Open Weights CoalitionRegulatory Capture and Washington InfluenceTiming Related to Executive Order DeadlineIPO Motivations Behind Anthropic's DecisionsContradictions in Anthropic's Open Model StanceImpact of Open Source on Token Business ModelTimestamps:00:00 Anthropic's stance on open models04:19 Discussing Anthropic's response to open models06:39 Understanding open weight models12:08 Future AI and cybersecurity risks15:04 Discussion on open-source AI models19:30 Discussing Anthropic's business challenges21:08 Cutting costs with open-source models26:17 Anthropic's recent stock downturn27:11 AI investment and cost efficiency shift30:14 Anthropic's stance on open source models36:32 INTROPICS IPO and regulatory discussions37:37 Wrapping up and subscribingKeywords: Anthropic, Claude, open model pact, open source AI, open weights, American AI leadership, Dario Amodei, IPO, regulatory capture, DC lawmakers, Chinese open source models, token revenue, per token business model, NVIDIA, Microsoft, Meta, OpenAI, Google, IBM, national security, AI safety, government mandates, chip controls, AI regulation, chip ban, industrial scale distillation, mandatory safety testing, inference, AI ecosystem, Opus 5, Fable 5, GPT-5, GLM 5.2, cost per task, token efficiency, model router, proprietary models, closed source AI, cybersecurity risks, Chinese cyberattacks, biological attacks, Glasswing program, open source vs proprietary, tech lobbying, Trump AI order, federal deadline, AI policy, artificial general intelligence, artificial superintelligence, AI monetization, S-1 filing, public company, venture capital, AI benchmarks, model switching, API pricing, model containment, Hugging Face incident, AI startup monopoly, safety vs business protection, market competition, AI cost reduction.Send Everyday AI and Jordan a text message. (We can't reply back unless you leave contact info) Ready for ROI on GenAI? Go to youreverydayai.com/partner 

Topics covered in this episode: Some more things about Django I've been enjoying Who cleans up after the vibe-coding party? Where Did All Your AI Tokens Go? AgentsView to the rescue! Careful with phishing all Extras Joke Watch on YouTube About the show Sponsored by us! Support our work through: Our courses at Talk Python Consulting from Six Feet Up Connect with the hosts Michael: Mastodon / BlueSky / X / LinkedIn Calvin: Mastodon / BlueSky / X / LinkedIn Show: Mastodon / BlueSky / X Join us on YouTube at pythonbytes.fm/live to be part of the audience. Usually Tuesday at 7am PT. Older video versions available there too. Finally, if you want an artisanal, hand-crafted digest of every week of the show notes in email form? Add your name and email to our friends of the show list, we'll never share it. Calvin #1: Some more things about Django I've been enjoying Julia Evans is learning "2010-style" web dev (Django + SQL + server-rendered HTML) after years of Go backends and JS-heavy frontends Query builders: likes defining custom QuerySet classes with chainable filter methods (.approved().future().with_tags()) — more readable than raw SQL Template filters: highlights urlize, linebreaksbr, json_script, and especially querystring for building/modifying query-string links in templates Migrations: still loves Django's auto-generated migrations — 19 and counting on her project Skips inheritance for class-based views; prefers function-based views for sharing code, though fine using Django's own mixins/interfaces Performance surprise: CPU profiling (via py-spy) — not slow DB queries — revealed the culprit; she'd accidentally disabled the cached template loader, and re-enabling it took throughput from ~2-3 req/s to ~12 req/s on a $10/mo VM Michael #2: Who cleans up after the vibe-coding party? FT Magazine piece by Sam Learner (July 11) on AI coding tools overwhelming open source maintainers - sent in by listener Dylan McConnell, whose main point was that this ran in the Financial Times, not a dev blog. cURL as the case study - Daniel Stenberg has been the only full-time person on it for years; libcurl has been installed an estimated 20+ billion times with 3,000+ listed contributors. Bug bounty killed - cURL ended its paid security bounty program in January, citing an "explosion of AI slop reports" that take real time to debunk and drain morale. Extractive contributions - authoring a PR is now nearly free, reviewing one still costs a human; tldraw's Steve Ruiz closed outside contributions entirely, asking why he'd want someone else writing the easy part. Guido weighs in - van Rossum says projects are holding emergency meetings over the slop flow, and notes LLM patches tend to touch unrelated parts of a file, making review more tedious. "Vibe Coding Kills Open Source" - paper from Miklós Koren's group: packages frequently recommended by coding models saw big download jumps with no matching engagement, breaking the reputation loop that sustains maintainers. Stack Overflow flatlined - over 100,000 questions a month before ChatGPT, under 1,500 last month, with the response rate cut roughly in half; the public archive is now stale training data. The course-creator angle - Josh Comeau's newest web dev course launched at about a third of prior enrollment, and he worries about devs who never learn which questions to ask. But the most interesting portion is what was omitted. Focused on: The end of the curl bug-bounty Omitted: High-Quality Chaos Why the omission is interesting It fits a narrative. The FT piece is a maintenance-and-decline story, and January-Stenberg is a perfect witness for it. April-Stenberg complicates it - same person, same project, better data, opposite direction on the specific claim being used. The tell is already in the article. Learner quotes Stenberg saying AI tools are much better at finding problems than fixing them. That's the April thesis in one line, and it goes undeveloped. Reason for the shift is process, not vibes. Killing the bounty removed the cash incentive and the venue change filtered the rest. Worth saying out loud, because "AI reports got better" isn't quite it - "no bounty plus a real triage platform" is closer. Joke too: Sarah O'Connor wrote a related piece (is this just before skynet launches?) Calvin #3: Where Did All Your AI Tokens Go? AgentsView to the rescue! Local-first desktop/web app for browsing, searching, and analyzing your past AI coding agent sessions (Claude Code, Codex, Copilot, Cursor, Gemini, Aider, and dozens more) Auto-discovers session files on your machine — no config needed; everything stored locally in SQLite, no cloud/accounts agentsview usage is a drop-in ccusage alternative — reads from pre-indexed SQLite, reports run 80–220× faster on large histories New Activity dashboard shows peak concurrency, active vs. idle time, agent-minutes, and cost — filterable by project/agent/machine, with a -json CLI report too Full-text + optional semantic search across every session; also imports Claude.ai/ChatGPT chat exports Install via pip install agentsview, uvx agentsview, brew install --cask agentsview, or download desktop binaries from GitHub Releases Michael #4: Careful with phishing all The situation I pass this along because it was a pretty sneaky bit of targeted phishing, and happened to play off an old interaction in bandit's repo. As usual with phishing scams there are a bunch of tells that this isn't legitimate, but just enough plausibility that I could see falling for it in a weak moment. Relative nobodies like me haven't historically been worth the effort to hit with scams this specific. Agents change the game though :-/. Be careful out there folks! Original message From: "Patrick (Blacktrace)" [HTML_REMOVED] To: LISTENER EMAIL Subject: Your Bandit #1350 (B105 NextToken false positive) -- just fixed that exact case Date: Wednesday, July 15, 2026 12:02 AM Hi AJ, Saw your Bandit issue #1350 -- the B105 hardcoded-password false positive on the string NextToken. I build a deterministic gate that filters that class of Bandit noise, and #1350 was literally the case I just fixed: NextToken / next_token / page_token / nextPageToken now stay quiet, while a genuine hardcoded token like api_token="sk-live-..." still fires. Verified against your exact case. 30-second paste: https://blacktrace.co/noise-eraser Where it still trips, published: https://blacktrace.co/kruc Curious whether it clears what you hit -- and if it trips on something of yours, that's the more useful reply. Patrick, Blacktrace I asked Claude for some analysis too. It was pretty good at finding them. The message name-drops enough real detail to feel legit, but the structure is pure phishing - everything in it exists to get AJ onto blacktrace.co. The strongest ones: Freemail sender, corporate signoff. Signs as "Patrick, Blacktrace" but sends from emailpjv@gmail.com. Real company outreach comes from the company domain, not a personal Gmail - and there's no last name. Over-specific targeting. It mirrors AJ's exact public activity - issue #1350, the B105 rule, the NextToken false positive, even the token variants. That's the "just enough plausibility" AJ flagged, and it's exactly what agents make cheap: scrape a GitHub issue, auto-generate tailored bait. Legit cold outreach rarely reads your history back to you this precisely. The entire payload is two links. Strip the technical flattery and the message is just "paste here" plus "see results here." When the whole point of an email is the click, that's the tell. "30-second paste." Low-friction urgency, and "paste" most likely means paste your source into their tool - handing your code to a stranger's site. Exfiltration dressed as convenience. Brand-new, no-reputation domain. blacktrace.co has no track record, and the name is doing some ominous work. The /kruc slug is random noise, not how real product pages get named. Precise-sounding jargon that's actually vague. "Deterministic gate," "noise-eraser" - impressive, empty. Bolted onto correct real details (B105 is the Bandit hardcoded-password test, sk-live- is a Stripe live-key prefix) to borrow credibility. The disarming close. "if it trips on something of yours, that's the more useful reply" - engineered humility that flatters your expertise and baits a response. Makes engaging feel like you're doing them a favor, which drops your guard. Extras Calvin: DjangoCon US 2026 is rapidly approaching, August 24-28, Chicago Ruff v0.16.0 massively expands its default rule set Ruff now enables 413 rules by default, up from 59 https://astral.sh/blog/ruff-v0.16.0 Michael: Completely redesigned the home page. Try /insights in Claude Code (terminal) Joke: We're Safe

Connect with Early Riders — https://www.earlyriders.com/contactConnect with Onramp — https://onrampbitcoin.com/contact-us/Presented collaboratively by Early Riders & Onramp Media…Final Settlement is a weekly podcast covering capital markets, dealmaking, early-stage venture, bitcoin applications and protocol development.This week Michael, Liam, and Brian open on the open-versus-closed AI debate: Jensen Huang's first-ever tweet and NVIDIA's letter backing open models, the new Security Alliance consortium, the researchers quietly wishing they could slow it all down, and NVIDIA's reported $250 billion financing for OpenAI's Ohio data center. They dig into Jack Dorsey's Buzz, the open-source workspace built on the same Noster and Bitcoin primitives as BitChat, and what shared compute and censorship-resistant protocols mean as BitChat downloads spike 32-fold in India. The guys run through a stack of payments deals: Stripe's talks to buy OpenRouter near $10 billion, Cursor, Ramp, and Meta building model routers, and Natural Pay's $30 million Series A for agentic payments. They get into their favorite topic, the DATs: Mallers exiting 21 Capital as Tether's XXI merger collapses, Strategy's first Stretch buyback, and the new Bitcoin security consortium taking on quantum. They close on Augustus's $180 million raise, Worldcoin, and Travis Kalanick's return with Adam.Chapters00:00 - Introduction and overview of AI and open models02:10 - Jensen Huang's first tweet on open models and global cooperation09:08 - Discussion on slowing AI development and game theory18:03 - Jack Dorsey's Buzz: Open source workspace and primitives24:55 - Shared compute, decentralized protocols, and open standards35:02 - Regulatory landscape and geopolitical implications45:11 - Impacts on privacy, security, and censorship resistance54:03 - Recent deals and industry movements in AI and crypto01:02:08 - The future of Bitcoin, open protocols, and decentralized dataIf you found this valuable, please subscribe to Early Riders Insights for access to the best content in the ecosystem weekly: https://www.earlyriders.com/researchKeep up with Michael:https://x.com/MTangumaKeep up with Liam:https://x.com/Lnelson_21Keep up with Brian:https://x.com/BackslashBTC

What does it actually take to build the people function at a company that went from a design tool to a global platform of 5,500 people?Jennie Rogerson, Chief People Officer at Canva, joins Jessica Neal and co-host Peter Clarke on Truth Works for a conversation about culture as an operating system, not a perk.Jennie's path into the role is one of the more unlikely ones in tech. She came up through hospitality, running restaurant floors where 220 covers a night meant 220 individual experiences to manage. She moved into hospitality tech, then got dragged to an after-work Canva workshop on making better presentations — and walked out convinced she had to work there.She applied. She didn't get the jobs. A colleague told her to look elsewhere. She ignored the advice.When the executive assistant role for co-founders Melanie Perkins and Cliff Obrecht opened up, thousands of people applied, and Jennie had no experience in the domain. So instead of a resume, she built a website. Canva launched Canva Websites the night before she planned to send it, so she rebuilt the entire thing on the product overnight. Then she spotted a line in Cliff's LinkedIn bio — "you're great, we're hiring" — and sent it to him directly.The interview process was two live challenges and a lot of real-time pushback. By the end of it, the role had been rewritten into something new: Leadership Operations, a chief-of-staff seat between the founders and the leadership team. Eighteen months later, Cliff called on a Sunday to offer her the top people job. Her first response was to ask if he'd lost his mind. She took it as an interim role. It's been years.Jennie is refreshingly clear that Canva hasn't figured everything out,  especially around parental leave re-onboarding, which she says most companies get wrong and hers is still actively learning. But what they have built is worth studying: managers are called coaches, performance reviews are called Growth and Impact, and every employee can anonymously submit to the Fix It Form, which has been running for roughly eight years and asks two questions — what's broken, and what's your solution.The conversation also gets into the support playbooks Canva built for grief and loss, family and domestic violence, and cancer and long illness — written with the employees who lived through those moments, and then open-sourced so any company can use them. One of them started as a single Fix It Form submission from someone who had lost their wife.And on hiring: Canva's CMO started as a communications intern. One of their AI researchers was found on a Discord community while he was stocking supermarket shelves. Those sit alongside senior hires from Zoom, Uber and Google — because the blend is where the value is.TOPICS COVERED:- Why hospitality is underrated training for building a people function- Getting hired with a website instead of a resume- The interview challenge that turned into a brand new role- Being offered the top people job with no HR background- Why culture is everyone's accountability, not the people team's- The Fix It Form — anonymous, always on, and it requires a solution- Why Canva calls managers "coaches" and reviews "Growth and Impact"- What a performance cycle looks like when it isn't a box-ticking exercise- Season openers and building public accountability into the calendar- Parental leave re-onboarding and why nobody has solved it yet- Building and open-sourcing the Grief and Loss guide- Talking about family and domestic violence at work- The risk of not talking about hard things- Finding talent internally and hiring for curiosity over pedigree- Using AI to draft the process, then keeping humans in the loopMost of this episode comes back to one idea. Jennie was hired into a role she was unqualified for on paper, by people who were looking for something other than a track record. Everything she has built since runs on the same logic — the AI researcher on Discord, the intern who became CMO, the engineers she invites to tear apart her own processes because they're the ones actually living inside them.Her closing challenge cuts both ways. If you're running an organization, the person who could take the next big step is probably already inside it, waiting for someone to bet on them. And if you're the one looking for what's next, the most interesting move might not be the logical one on paper — it might be the thing you're genuinely curious about.Which is why the line she keeps returning to isn't really about Canva at all:You don't join Canva to experience the culture. You join Canva to be the culture.Truth Works is hosted by Jessica Neal,  former Chief Talent Officer at Netflix, now in the business of telling the truth about how work really works.

Michael Washington is an ASP.NET and C# programmer with experience in process improvement, billing systems, and student information systems. He is the founder of AiHelpWebsite.com and BlazorHelpWebsite.com.You can find Michael on the following sites:BlogGitHubXYouTubeHere are some links provided by Michael:BlazorDataOrchestrator PLEASE SUBSCRIBE TO THE PODCASTSpotifyApple PodcastsYouTube MusicAmazon MusicRSS FeedYou can check out more episodes of Coffee and Open Source on https://www.coffeeandopensource.comCoffee and Open Source is hosted by Isaac Levin

On this week's episode of the podcast, I am joined by Michael Kaminsky, co-founder and co-CEO of Recast. We explore the rise of open-source Marketing Mix Modeling (MMM) tools and the challenges of measuring modern ad spend. Among other things, we discuss:Whether open-source MMM libraries like Robyn and Meridian are truly unbiased tools or subtle instruments for big tech self-gradingHow marketers can effectively evaluate the accuracy of MMM outputs through techniques like parameter recovery and predictive forecastingWhy smaller, performance-driven brands might actually find more value in last-touch attribution than in complex econometric modelingIf the inherent uncertainty in MMM estimates makes them fundamentally incompatible with the fast-paced feedback loops of digital advertisingWhat role generative AI and large language models play in democratizing access to sophisticated marketing data science analysisWhen a brand should transition from simple attribution methods to a multi-layered approach involving incrementality and econometric modelingHow the shift toward CTV and non-trackable channels is forcing a resurgence in probabilistic measurement frameworks like MMMThanks to the sponsors of this week's episode of the Mobile Dev Memo podcast:⁠⁠INCRMNTAL⁠⁠⁠. True attribution measures incrementality, always on.Xsolla⁠. With the Xsolla Web Shop, you can create a direct storefront, cut fees down to as low as 5%, and keep players engaged with bundles, rewards, and analytics.⁠Branch⁠. Branch is an AI-powered MMP, connecting every paid, owned, and organic touchpoint so growth teams can see exactly where to put their dollars to bring users in the door and keep them coming backInterested in sponsoring the Mobile Dev Memo podcast? Contact Mobile Dev Memo advertising.The Mobile Dev Memo podcast is available on:SpotifyYouTubeApple Podcasts

Plus: chip maker CXMT becomes the most valuable company listed in mainland China in its debut. And some U.S. companies plan to increase headcount despite AI fears. Imani Moise hosts. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

In this week's episode of the Coin Stories News Block powered exclusively by Ledn, we cover these major headlines related to Bitcoin, macroeconomics, and global finance: What the wave of Bitcoin company closures and bankruptcies tells us about where we are in the bear market 84% of Bitcoin is now held by long-term holders — the highest level ever recorded AI agents just paid each other in Bitcoin over Lightning for the first time BlackRock, Fidelity, and seven other institutions pledge $15M to protect Bitcoin from quantum threats — and critics are calling it "Big Bitcoin" Strategy repurchases $25M in STRC at $86.52 and boosts its cash reserve to $3.75 billion ---- The News Block is powered exclusively by Ledn – the global leader in Bitcoin-backed loans, issuing over $11 billion in loans since 2018, and they were the first to offer proof of reserves. With Ledn, you get custody loans, no credit checks, no monthly payments, and more. My followers get .25% off their first loan. Learn more at www.ledn.io/natalie  ---- Order Natalie's new book "Bitcoin is For Everyone," a simple introduction to Bitcoin and what's broken in our current financial system: https://amzn.to/3WzFzfU If you'd like to buy using Bitcoin, just head to https://shop.talkingbitcoin.com and pay in sats! ---- Read every story in the News Block with visuals and charts! Join our mailing list and subscribe to our free Bitcoin newsletter: https://thenewsblock.substack.com  —- References mentioned in the episode: BitMEX Ends Operations After 11 Years BitMart to Wind Down Its Exchange Satsuma Shareholders Approve Bitcoin Treasury Liquidation Smarter Web Company Sells Bitcoin to Repay Debt Poolin Files for Bankruptcy Poolin Files Chapter 11 and Sets $52 Million Floor Bid Bitcoin Standard Treasury Company Scraps Original SPAC Terms Jack Mallers Steps Down as Twenty One Capital CEO MARA Sells $1.5 Billion of Bitcoin Amid AI Shift Bitdeer Empties Its Bitcoin Treasury as Miners Pivot to AI Natalie Brunell Interviews MARA CEO Fred Thiel 84% of Bitcoin Is Held by Long-Term Holders Bitcoin Conviction Is at an All-Time High Lightning Labs Launches Wavelength Jensen Huang Explains Why Open AI Models Matter Mark Zuckerberg on Open Source and Preventing Centralization Elon Musk Says X's Code Will Be Open Source and Audited Jack Dorsey Announces Buzz Michael Levin Demonstrates Buzz and Wavelength Working Together Satoshi Nakamoto on the Root Problem With Conventional Currency Mike Schmidt Explains the Bitcoin Security Consortium Brian Armstrong on Preparing Bitcoin for Quantum Computing Official Bitcoin Security Consortium Announcement BlackRock, Coinbase and Strategy Join $15 Million Security Consortium Nine Firms Launch the Bitcoin Security Consortium Galaxy Launches the Bitcoin Quantum Readiness Initiative Strategy Overhauls Its Bitcoin Capital-Markets Metrics Strategy Announces Its New Bitcoin Capital-Markets Metrics ----  

Matt Murphy is a Partner at Menlo Ventures, who just raised $3 billion in fresh capital, its largest pool ever. Matt's portfolio includes Anthropic, Lovable, Legora, OpenRouter, Chai Discovery, Axiom, OpenEvidence and more. AGENDA:  00:00 Why Menlo Broke All Its Investing Rules to Back Anthropic 09:00 Why Ownership Matters Less in an Outlier-Driven Venture Market 13:00 Do We Have an SPV Problem in Venture Today? 20:00 Do Margins Still Matter in AI? 23:00 Why Open Source Won't Derail Anthropic's Growth 26:00 Does Every Model Provider Need to Build Its Own Chips? 29:00 Why Anthropic Is Not a Threat to Legora 32:00 Why Series A Is the Hardest Place to Invest Today 36:00 Why Signalling Is B.S. and Every Fund Is Going Full Stack 42:00 Why Building a Company in Europe Is Hard Mode 47:00 Why Triple-Triple-Double-Double Is No Longer Venture-Scale Growth  

Block's Buzz is an open, self-hostable workspace for humans, AI agents, chat, and code; and it may be the most Linux-friendly vision for what comes next.Sponsored By:Jupiter Party Annual Membership: Put your support on automatic with our annual plan, and get one month of membership for free!Managed Nebula: Meet Managed Nebula from Defined Networking. A decentralized VPN built on the open-source Nebula platform that we love.Support LINUX UnpluggedLinks:Jupiter Broadcasting Buzz CommunityWeb Boost — Send us a boost via sats or USD

In this episode, we sat down with Ryan, the mastermind behind Mujina Mining Firmware, to dig into what Mujina is, why open-source mining firmware matters, and how the project is evolving. We talked through the difference between simply publishing source code and building a true community-driven open-source project, why Mujina uses the GPL license, and how that protects downstream users and contributors in a mining ecosystem that has been dominated by closed, proprietary firmware for years. Ryan also explained the role of the Mujina dev calls, the distinction between firmware, operating systems, kernels, and images, and why the next big step for the project is shipping full, easy-to-flash Mujina OS images for machines like the S19.We also covered a lot of ground on what's happening around the broader 256 Foundation ecosystem: Ryan's mining workshop in Nairobi and how others across Africa are now reusing those materials, the growing momentum behind HydraPool and open-source mining education, recent mining decentralization events in DC and Nashville, and new hardware progress including the Bitaxe Bonanza, Bitaxe Proto, and support for non-Bitmain chips. Overall, this was a wide-ranging conversation about replacing the proprietary mining stack with open, extensible tools that miners, builders, and educators can actually own, adapt, and improve together.

Ben Horowitz joins Theo Jaffee and Sofia Puccini to discuss one of the biggest debates in AI today: the future of open-source models. They examine the growing push to restrict open models, why Ben believes open source is critical for security, innovation, and competition, and what happens if a handful of frontier labs come to dominate the AI ecosystem. They also discuss distillation, AI monopolies, China's role in open-source AI, robotics, manufacturing, the economics of frontier models, and why Ben believes the AI market is still in its earliest stages. The conversation closes with a look at AI-generated art, creativity, and whether new tools will spark the next cultural renaissance.   Resources: Follow Ben Horowitz on X: https://x.com/bhorowitz Follow Theo Jaffee on X: https://x.com/theojaffee Follow Sofia Puccini on X: https://x.com/schisofrenia Stay Updated:Find a16z on YouTube: YouTubeFind a16z on XFind a16z on LinkedInListen to the a16z Show on SpotifyListen to the a16z Show on Apple PodcastsFollow our host: https://twitter.com/eriktorenberg Please note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details please see a16z.com/disclosures. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

In this one, we talk about the hardware we use. ==== Special Thanks to Our Patrons! ==== https://thelinuxcast.org/patrons/ ===== Follow us

Back in October 2024, Poolside was an early AI star. Cofounded by former Github CTO Jason Warner, the startup had raised $500 million on a $3 billion valuation to build coding agents for governments and large companies. But over the next 18 months, Poolside largely disappeared from view, while OpenAI and Anthropic ballooned to nearly trillion-dollar valuations with a crop of Chinese labs building open source models nipping at their heels.  Now Poolside is back with a new model called Laguna that on public benchmarks beats its American and Chinese open source competition — with the very notable exception of Chinese lab Moonshot's latest AI model, Kimi K3.  "As an American company building for the West, we'll be the most capable open model in the West,” Warner, Poolside's co-CEO and cofounder, tells Forbes. “Globally, in this weight class of the 118 billion parameter model, we are the leader.” Warner claims that the startup spent those 18 months where it went quiet building the infrastructure for a “model building factory” that could pump out Laguna and continue with new and more powerful iterations every five weeks. "The classic notion of model building is an artisanal process…We've built an industrial model-building process,” he says. By Iain Martin, Forbes Staff Learn more about your ad choices. Visit megaphone.fm/adchoices

Director of Talent Development and AI Education at Fusemachines, Rojesh M. Shikhrakar, has spent nearly a decade building and working with AI systems, long before the world knew what a prompt was and long before ChatGPT made "artificial intelligence" a dinner-table phrase. In this episode, he pulls back the curtain on what's actually happening inside the machines we've started trusting with our decisions, our classrooms, and possibly our future economy. In this episode, we talk about: Why token costs are dropping 40x while model intelligence is compounding 1,000x, and what that math means for anyone building today. How predictive policing exposes the one thing AI still cannot do, forecast human social outcomes with any real accuracy. What "human in the loop," "on the loop," and "off the loop" actually mean for the jobs disappearing first. Why hallucination isn't a bug that will eventually be fixed. It's a permanent tax the industry has quietly accepted. How Nepal could build sovereign AI infrastructure instead of renting intelligence from someone else's data center. What happens to entry-level hiring when AI makes junior-level output senior-level ready almost overnight. Why competing with China, India, or the US on AI is the wrong game, and what the right one looks like. How philosophy, not code, might ultimately define where artificial intelligence goes next. This isn't a conversation about tools. It's about who gets to hold power when intelligence becomes cheap enough for anyone to buy. Rojesh M. Shikhrakar makes an uncomfortable case for why Nepal cannot afford to sit this one out. Watch it before your competitors do. TIMESTAMPS ⏱️ Timestamps 00:00 Highlights 07:19 Open Source vs Closed Source: What's the Difference? 16:07 AI Hallucinations Explained & It's Impact 27:14 Why It's So Hard to Change Nepal's Education System 40:37 Is There Really Such a Thing as Originality? 46:30 Does Nepal Have the Capacity to Lead in AI? 57:38 Why Entry-Level Jobs Are Disappearing in Nepal 1:08:37 AI Has a Much Higher Ceiling Than Most People Think 1:11:21 Technology Has Always Been Built on Science 1:23:03 Why Software Companies Are Becoming Hardware Companies 1:33:12 How AI Is Driving the Rise of Cybersecuritya Want to become a video podcaster? Get info: https://becomeadoer.com/programs/beco...   If you love reading, don't miss our newsletter on Substack Link: https://substack.com/@doersglobal?   Want to join us live in the studio as an audience member? Fill out this form: https://forms.gle/xZi8yptyoxkkc6aa8   ✉ Reach out to us at partners@doersnepal.com  

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

Last year Jimmy Bogard released commercial licensed versions of AutoMapper and Mediatr - how's it going? Carl and Richard chat with Jimmy about the decision to go commercial and what he's learned after a year. Making customer transactions is part of the process for sure, as is a new array of issues and ideas. But for the most part, Jimmy says it was the right decision and has been a good experience. Another great story of sustainable open source!

Leaving Port 22 open to the internet, FreeBSD Foundationals, GhostBSD Finance Report, Making your own read-only device with NetBSD, and more... NOTES This episode of BSDNow is brought to you by Tarsnap and the BSDNow Patreon Headlines I Left Port 22 Open on the Internet for 54 Days. Here's Who Showed Up. FreeBSD Foundationals: The Boot Process - From the Loader to Boot Environments News Roundup FreeBSD 15 on a Laptop GhostBSD - March 2026 Finance Report Make your own Read-Only Device with NetBSD NFS-problems I didn't expect at all Tarsnap This weeks episode of BSDNow was sponsored by our friends at Tarsnap, the only secure online backup you can trust your data to. Even paranoids need backups. Feedback/Questions Send questions, comments, show ideas/topics, or stories you want mentioned on the show to feedback@bsdnow.tv Join us and other BSD Fans in our BSD Now Telegram channel

In recent months, the open vs closed, and US vs China discussions on model ownership and sovereign/local AI have heated up to a fever pitch. So it is very very good news that Poolside AI are finally emerging with new models, like Laguna S 2.1, that are beating Thinking Machines' recent release nearly 10 times their size.Poolside's recent tech report got a lot of praise due to their level of detail, and Vibhu first covered Laguna's recent technical report on our paper club:From spending $12 million building language models for code before the world cared to creating a Model Factory that can take a model from pre-training to release in eight weeks, Eiso Kant has spent more than a decade betting that code is the path to AGI. In this episode, the Poolside co-founder joins swyx and Vibhu to explain why ChatGPT felt like vindication, why Poolside embraced open weights and open research, and why he would rather live in a world with 100 foundation model companies than five even if Poolside were one of the five.We go deep on Poolside's Model Factory: the engineering systems behind 10,000–20,000 experiments per month, streaming data directly into training, reproducible experimentation, low-precision compute, and agents that increasingly write code, launch jobs, evaluate results, and modify the pipelines used to train future models. Eiso also unpacks their recent launch Laguna S, why persistence, verification, and backtracking may matter more than raw intelligence, how much capability remains inside smaller models, why reinforcement learning will move earlier into pre-training, and why next-token prediction is still extracting too little from the web.We also discuss model-harness co-design, Poolside's path from coding agents to AGI, why Eiso thinks MCP and traditional tool calls are “stupid,” the real economics behind frontier-model training, Poolside's $500 million raise, open-source AI, regulation, NVIDIA and TSMC's influence, engineering productivity in the agent era, high-agency teams, and hiring at Poolside.We discuss:* How Andrej Karpathy's RNN work inspired Eiso to start building language models for code in 2015* Why Eiso spent four years and $12 million pursuing an idea before the market cared* Why ChatGPT felt like vindication and brought Poolside back to open source* Why Eiso would prefer 100 foundation model companies over an oligopoly of five* The difference between releasing open weights and publishing genuinely open research* Why Poolside deliberately built a global research organization outside the Bay Area talent war* Why model building is ultimately 90% engineering* The Model Factory: Poolside's end-to-end system for rapidly training and improving models* How fewer than 70 researchers run roughly 10,000–20,000 experiments each month* How Poolside moved from six-month model cycles to five- and eight-week launches* Why streaming data directly into training unlocked faster experimentation* How immutable data, versioned code, and reproducibility enable rigorous model research* Why Eiso wants capable researchers to leave their labs and become Poolside's competitors* Why 95% of model building can be reduced to better data or compute efficiency* Laguna S and why persistence, verification, and backtracking can outperform raw intelligence* Why smaller models may handle far more knowledge work than previously expected* Why reinforcement learning will move earlier into pre-training* Why next-token prediction is still failing to extract enough knowledge from the web* Why distillation and environments have become the AI industry's favorite “drugs”* Why mid-training is really an early form of curriculum design* Low-precision training, networking bottlenecks, and the next gains in compute efficiency* Laguna S: 118 billion total parameters, 8 billion active, and eight weeks from training to launch* Why model builders can often evaluate a new checkpoint within its first 30 minutes* Model versus harness: where agent capabilities actually come from* Why Poolside sees coding and long-horizon software tasks as a path to AGI* Why Eiso thinks MCP and traditional tool calls are “stupid”* Why future agents will write scripts instead of choosing from dozens of predefined tools* The case for minimal harnesses, containers, and model freedom* Why Poolside is prioritizing vision but does not expect to work on audio soon* Why language may be the most compute-efficient modality for encoding knowledge and reasoning* The real cost of model development and why the final training run is anticlimactic* The story behind the Poolside name and why it represents refusing to lower ambitions* How Poolside raised $500 million while investors still questioned whether AGI was real* Why intelligence could become the world's most demanded and commoditized resource* When open models may become too capable to release without restrictions* Why unilateral AI safety does not work in a globally competitive environment* How regulation could accidentally lock in an oligopoly of two or three AI companies* NVIDIA, TSMC, and the hardware systems underpinning foundation-model progress* Why reinforcement-learning wall-clock time is one of Poolside's biggest bottlenecks* Why Poolside trains models from scratch instead of simply distilling larger models* How AI changes the way companies should measure engineering productivity* Why agency may become the most important quality for employees in the AI era* How leaders align high-agency people through shared goals and clear constraints* Hiring across research, post-training, pre-training, architecture, evals, and engineering at PoolsideEiso KantLinkedIn: https://www.linkedin.com/in/eisokantX: https://x.com/eisokantPoolside: https://poolside.aiTimestamps00:00:00 Introduction00:00:54 Karpathy, RNNs, and Building Code Models Before Transformers00:02:26 The $12M Failure and ChatGPT Vindication00:03:39 Open Source and the Case for 100 Foundation Model Companies00:09:22 Open Weights, Open Research, and Poolside's Global Team00:16:04 The Model Factory: Why Model Building Is 90% Engineering00:20:19 Agents, Automated Experiments, and Early Signs of RSI00:24:04 Streaming Data, Reproducibility, and Scientific Rigor00:30:35 Creating More Foundation Model Companies00:36:07 Laguna S: Persistence vs. Raw Intelligence00:43:01 Reinventing Pre-Training, RL, and Curriculum Design00:52:33 Low-Precision Training and Squeezing More From Smaller Models00:58:37 Model Harnesses, Coding Agents, and the Path to AGI01:09:26 Why MCP and Traditional Tool Calls Are “Stupid”01:13:04 Vision, Multimodality, and Why Language Still Matters01:18:15 Scaling Models and the Real Economics of Training01:20:40 Why Poolside Is Called Poolside and Raising $500M01:27:37 Open Models, AI Safety, and the Risk of an Oligopoly01:33:53 NVIDIA, TSMC, and the Reinforcement-Learning Bottleneck01:41:52 Smaller Models, Distillation, Engineering Productivity, and HiringTranscriptIntroduction: Eiso Kant, Poolside, and Open ModelsSwyx [00:00:00]: All right, we're here in the studio with Eiso Kant from Poolside, together with Vibhu. Welcome.Eiso Kant [00:00:08]: Thanks. Thanks for having me, guys. Good to be here.Swyx [00:00:10]: Yeah, fresh on the plane. You texted me, you were like, “Hey, I'm on my way to SF.” I was like, “You're on a plane right now, right?” Like, hey.Eiso Kant [00:00:16]: I know. After I texted you, I realized that probably coming in with major jet lag was gonna offer some fun experiences today, but let's do it.Swyx [00:00:23]: I mean, I think the thing I would tell guests is that they don't have to prepare that much because if you're truly working on this every single day, then even, like, what you hazily remember is going to be new for a lot of the audience that don't live in your world every day, right? so 10 years ago, you did a talk at Google Slush, talking about the democratization of AI. and, now here you are, like, open sourcing an incredible new model that we're gonna talk about. But I guess, like, what got you into democratization of AI? Like, it's not obvious from your LinkedIn or something.From Karpathy's RNN Post to SourcedEiso Kant [00:00:57]: No, it's not at all. I don't think it's obvious how I got in this space. I owe getting into this space to Andrej Karpathy.Eiso Kant [00:01:05]: In 2015, he wrote an article called “The Unreasonable Effectiveness of Recurrent Neural Nets.”Swyx [00:01:10]: Neural Nets, yep.Eiso Kant [00:01:11]: And that article, I read it, and I pivoted my startup at the time overnight to working on RNNs, and later LSTMs and Transformer models to be able to write code. If you go to this article and you scroll down, you can start seeing, like, this was the precursor to what ended up becoming language models. So, at least when he was character-level language models that were starting to predict letters, he has an example out here. There's a little Paul Graham generator, and you can read it, and the text makes sense, but it doesn't. and there's a little-- There's an example of code a little bit further down. Yeah, so Shakespeare.Swyx [00:01:47]: Shakespeare.Swyx [00:01:49]: CoolEiso Kant [00:01:49]: And for some reason, I read this, and I went down the rabbit hole of learning everything I could about RNNs and LSTMs, right? This is Transformer paper. And I had built a completely unreasonable belief, that neural nets should be able to generalize to anything and everything, and that language should be able to generalize, to a lot of things that are intelligent and the ability to write code. And so I started building Sourced, which was a fully open source company trying to build, what we used to call machine learning on code, language models on code. And we spent about four or five years on this, till the end of 2019. And that sounds really cool today, but back then, no one cared.Eiso Kant [00:02:29]: Right? Like, no one cared. We were in the dark. Like, we did things along the way. We tried applying convolutional neural nets to, like, the structure of code. We were. when attention came out, we were applying it to LSTMs, and then the Transformer paper came out. And it - it wasn't obvious, and what we missed throughout that entire journey, that we were on the right track, but we should have just kept scaling up. And today, to all of us, the scaling laws and scaling up seems like the most obvious thing. But having spent four or five years of my life on working on language models on code, it wasn't obvious. So I have a lot of respect to folks at Google and OpenAI and others who took that confidence and kept going. we failed ultimately at the time, and it was, like, biggest failure of my career, right? You blew $12 million of investors' money, which was a lot back then.Swyx [00:03:18]: Yep.Eiso Kant [00:03:19]: You spent, still a lot, but, And you spent years with, like, a group of 40 people just obsessing over this problem. And life took a different turn, And it was, and family became a focus, and I kept my heads down and really, didn't really look at language models for the following two years. big mistake considering Following years are gonna be really interesting. And then ChatGPT came out And it was like a vindication. It's like people started texting me. I found, like, my old, work decks and these old talks. And throughout that whole journey, we,ChatGPT, Vindication, and Returning to Open SourceEiso Kant [00:03:56]: We really had a strong point of view at the time that, like, as you're building more capable intelligence, it should be open and open source.Eiso Kant [00:04:04]: When we started Poolside, that wasn't the case at all, and I wanna be very open about it. When we started Poolside, we were like, there was a premise of two things. One is this technology is not gonna stop compounding in capabilities. I think to most people obvious today, but three-plus years ago when we started, most people were still arguing if these were stochastic parrots or not.Eiso Kant [00:04:23]: And the second was that reinforcement learning was gonna be the biggest driver for LLM capabilities. Today, very obvious. Three years ago, was not an opinion held or direction held at either OpenAI or Google or Anthropic or others. And so people looked down on us a little bit. They were like, “ is this really gonna work?” And so we just started working the problem, and we never really thought about open source again. We just kept our heads down and we built our, like, knowledge, understanding from scratch, right? We didn't roll out of an existing lab. So we picked up the papers and started writing code and figuring things out.Eiso Kant [00:04:59]: And it wasn't until the beginning of this year that me and my founder, Jason, picked up the open source conversation again.Eiso Kant [00:05:07]: And if you go back to some of the early things on our website, it was very straightforward. It was we wanna get to AGI, we wanna support a world of abundance, and we wanna be the first company that gets there.Eiso Kant [00:05:20]: But we started talking at the beginning of this year because it became obvious that the world was going in a direction that was starting to like, pick at us a little bit. Like, it didn't, this didn't happen overnight. It was, like, a little bit we were seeing this and we're like, “Okay, The world's going down a path.” And Throughout this journey, there was something that I used as a, as an analogy or thing. So I said well, if I go back to back in those days, 2015 or 2016, we're working on this, and I picked up a fi book off the shelf, and I was reading the book about 2035. AGI is achieved, and the story would be over the following, decades. And it would have that first chapter where everyone's trying to figure things out. You'd get the chapter of ChatGPT coming out And then you would get to the chapter where the world was at a fork in the road, and the one that it picked was one where three or four or a handful of companies were going to create all of intelligence moving forward.Eiso Kant [00:06:21]: And when I thought about that story, it felt like a dystopian fi book, not a utopian fi book. And the reality is, I'm a utopian fi guy. Like, and so We took a step back and said, “Hey, can we play a role here?” Now it was easy for us to do so because we were not at the frontier.Eiso Kant [00:06:41]: If we were at the frontier, I don't think we could have changed our mind. and I don't mean this like it's when the moment there's too much capital involved, too much expectations, you've built up things, right? We're a small team, just improving and improving. And so we knew that we could make that decision now, but it would be a lot harder to make as we got closer and closer to the frontier and caught up to others. And did a lot of soul-searching and a lot of conversations, and said, “No, this makes sense,” Even if there's big unanswered questions, like how the hell do you build a business model with foundation models about open source? Big open-ended question that we do not fully have the answer to yet, right? At what point do you no longer wanna release open source models because misuse of models has, real potential risks associated with it? how is the government gonna respond to open source? but I think it all just came down to one thing, and I'll stop the monologue, is the fact that I rather live in a world that has 100 foundation model companies than a world that has five, even if I was one of the five. And the smallest and most meaningful contribution we can make for 100 to exist is to open up our research and open up, like, our weights right now and figure out along the way how we can, like, do more.Neo-Labs, Model Choice, and the Token EconomySwyx [00:08:01]: Yeah. I think if anything, over the past three years, that has become a bit more true. you are one of a cohort of Neo labsEiso Kant [00:08:10]: YeahSwyx [00:08:10]: That people are now calling that. And, we're, we're doing this on the day that Thinky launched their, new model and you are outperforming them on their, on some benchmarks that they released, right? Like, they just don't have it yet. so it goes to show that I think, like, this is one of those things where, like, there is room for multiple players, and you are seeing a little bit more of the future. Maybe more like 20, not 100, but, like, you are one of the 20.Eiso Kant [00:08:36]: I really hope so, right? I think we I'm, I'm excited about their release, and I'm excited about everyone releasing because, like, ultimately, like, choice competition is both gonna drive progress in the right direction. But the fact that like, we create models and while we all, drink out of the same well of data effectively, we do introduce very different behaviors and biases in our models. Some are intended biases, some are completely unintended biases.Swyx [00:09:03]: Yeah.Eiso Kant [00:09:03]: And if we shape up in an ecosystem in the world where open models are gonna be a part of the token economy, like, I don't think there's any question about it anymore Then we want to be able to live in a world where companies, countries, people can choose and say, “Hey, I am most aligned and I trust most this provider for these things.”Swyx [00:09:25]: Yeah.Vibhu [00:09:26]: I think more than just one of the 20 Neo labs, up until recently, most of open source innovation was coming from the Chinese labs, right? So there's the DeepSeek of the West. Is it today? Okay, maybe it's thinking machines reflection, but there aren't many, right? So, one of the things you guys started in France, Europe, but very much now you're taking that American standpoint and more than just that, the point is the Chinese models that we see, they're not super open research. the work you put out is, I think, some of the best. So every few months you get not only frontier models, but also here's a breakdown blog, paper, technical report of here's everything for state of the art to build, frontier intelligence and you're filling that gap too, right? So not just only open weight, not just Western, but also pretty open research.Open Weights vs. Open ResearchEiso Kant [00:10:20]: No, I appreciate it. Look, I think it's, I think it's the most meaningful contribution, right? Weights are a binary. Let's call them what they are. Yes, we can modify them, we can change them, but, like, giving someone the weights does not allow them ultimately to recreate what you're doing, right? And so now there's challenges around releasing data sets, challenges around like releasing certain things, but being able to share your research, like, right, how do we do it? What are the lessons we learned that we spent, tens of thousands of experiments of compute on? I think very much so. One correction though, Vibhu, and I say this because it's been haunting us for quite a few years. We from day zero were an American company.Swyx [00:10:55]: Yeah. They movedPoolside's Global Team and American Company StorySwyx [00:10:56]: To France.Eiso Kant [00:10:56]: So the story once and for all is very. We start as an American company. We have always been an American company, and early on we made a very conscious decision. We said, “We're not gonna hire any researchers in the Bay Area. We're gonna look for talent everywhere else in the world.” and that is everything from Middle Americas, Seattle to, Serbia, and to Taiwan and Singapore and other places. And it was because we took a view that this was gonna become a talent war for this, and I think it has over the years now. Three years ago, that wasn't fully obvious yet. I think today it very much is. And we also realized that, like, some of the world's most capable people with, like, the most interesting, innovative ideas were not just gonna be here. And so it led us to create like a fully remote company. and we ended up opening an office in Paris and London and different places and we have a lot of the team in the US and a lot of team outside. But we always took this view of like, we're an American company, but if we want the best of the best to work with us, we need to take a global view. Now we do also have people here in Silicon Valley, like the company's grown and others, but I think one of the things that, it slowed us down at the beginning, but it has sped us up now, and it's why you're seeing like the progress, I think, on our models and the cadence at which we release, is because we didn't roll out of an existing lab. Right? we didn't, we didn't have a lot of the information that's freely flowing around here at the time. We just took this point of view as like, “Okay, well, let's just work the problem. Let's just go and, like, read the few papers that are out there, and let's just figure this stuff out.” And we made some hilarious mistakes in model training because of that over the yearsEiso Kant [00:12:35]: Like especially in the first 12 months. there's a few that I think still haunt me and scare me. We can talk about them later. but it created a, like, a resiliency and persistency in the team, right? with extremely few people have left us over the years, that, like, told us, “Okay, we can do this.” When we first wrote our first training code base completely from scratch, it wasn't a fork of any open source. It was just like, “Okay, let's build it from scratch.” I remember we had this one moment where we spent three weeks working out an optimizer bug. Like, it was like training just couldn't get stable. We, like, obsessed over it, and we thought, like, maybe we were wrong. Maybe we should have just forked this repo, or we should have. But then when we solved it, I still remember at the time we were like five people in the company. when we solved it, we were like, “Oh, we can do things,” like if we're just willing to work hard. and I think that culture with a very strong engineering bias has helped us, like, get to where we were. And so there's this notion of open source and talent and these things. I think we, We just took different decisions from a different starting point. and I think we are lucky. I do want to definitely call it lucky. And there was a lot of hard work at the team that now, like, that's starting to show up in results.Swyx [00:13:52]: Just ‘cause we probably won't revisit this again, but, and this is a fun recruiting challenge if someone knows the answer. What was the bug? And then we won't tell the solution, but we'An Optimizer Bug and the Value of Building From ScratchEiso Kant [00:14:01]: So the - This - You're gonna test my memory here,Swyx [00:14:04]: Oh, okayEiso Kant [00:14:04]: So but I thinkSwyx [00:14:05]: DirectlyEiso Kant [00:14:05]: I think I can recall. So if you, so if you look at, So if you take like Adam as an optimizer, you have epsilonSwyx [00:14:12]: YeahEiso Kant [00:14:13]: Which is, right, like in the denominatorSwyx [00:14:14]: Momentum and weights. YeahEiso Kant [00:14:15]: Is exactly, in the denominator. And at the time, if I recall, you looked at like the early Llama papers and things like that. People were juicing epsilon, like, quite a bit. Like, they were, like, adding, I don't know if it was E minus four or whatever, like a high value for epsilon.Eiso Kant [00:14:31]: And if you think about this during training, it's like a bit weird and counterintuitive that we're adding noise to our optimizer by just adding effectively, like, a random number in the denominator, right? Like behind the decimal point. And I don't recall the exact bug, but it had - What I remember is once we solved it, we no longer had to juice epsilon as much as, like, was happening in the Llama paper and other places. and it was like one of those fundamental moments where we had trusted this paper that was out there, and we're like, “Oh, no, it has to be this way. It has to have this high value of epsilon.” But it made no sense to us intuitively. Like, why do you have to have this so high? Like, if you're just trying to avoid division by zero, why can't the value be extremely small? and that was like one of those moments where you realize like, okay, finding things out from scratch yourself builds a better intuition. Because the one thing you learn very quickly with model building is that your intuitions that you start with are gonna get beaten up so hard.Eiso Kant [00:15:33]: Right? Like - It's such an experimental science, that the things that seem obvious, you very quickly get to learn, like, you were wrong, and hopefully you figure out why, and sometimes you don't even.Swyx [00:15:45]: Yeah. yeah, so, one of the reasons that you, when you released your new models, Vibhu got really excited. I mean, everyone got really excited. But Vibhu led our paper club on it, and you guys sawEiso Kant [00:15:58]: YeahSwyx [00:15:58]: Obviously. maybe talk through some lessons learned in that, whatever you can disclose. we can focus on the model factory stuff, whatever you think is a good starting point.Model Building as EngineeringEiso Kant [00:16:08]: So I would say that our view from very early on in the company was that model building is ultimately 90% engineering.Eiso Kant [00:16:18]: And I think we all know it in the industry because if you look at where's every researcher spending their time, they're spending their time writing code, right? Looking at data and writing code. And so we said, okay, The state at the moment, like three years ago, was bash scripts and Slurm and spaghetti code bases for training and, like, data pipelines that were patched together. And we looked at this and said, “Well, ultimately, model building is a process.” You're going from raw data, right? Like training raw material, the web, et cetera. you're doing a whole bunch of filtering, cleaning up, transformations, analyzing. These days, that's, far more complex than it was three years ago. then you're training a model, which is effectively a large distributed systems problem, right? Across hardware that has still-- It's become a lot more reliable. It was extremely flaky back then. and now with every new generation, we get our new sets of challenges. And then you go into the next stages, right? There was no training back then, but, like, you got, your post-training and then your reinforcement learning. And so we looked at this and we said, “Well, this looks like an industrialized process. This looks like an end process, that every single part of it has its machinery,” right? If it's your big data pipelines, if it's your crawling ingestion of the web, if it's your, large-scale distributed training, and then you've got your reliability. And we said, “Well, why don't we take some of the world's smartest distributed systems engineers that we knew and make them part of the process of research from day zero?” Not retrofitting it later on, but, like, really from the beginning. And that became our model factory. And so our model factory started with a handful of components. Today, it's thousands of components, and I try to equate it to, if you think about, like, someone who was at the very early days of Foxconn, if they had been there for the following, decade, they would be able to rebuild Foxconn because they saw every decision that led to building that system and all the complexity. If you and I walk into Foxconn today, no chance.The Model Factory and Experiment VelocityEiso Kant [00:18:18]: Right? Because we don't have the lineage and history of decisions that led to that. And so we built early on from the beginning- with a team that really understood that, well, the metric that we are optimizing for is the speed of an idea from a researcher to an experimental result that we can trust to then being part of the next model training.Eiso Kant [00:18:42]: And in the. And because it's such an experimental science, ultimately, in the beginning when it wasn't that complex, you could patch your way around it, right? But now, at any foundation model company, you are running. I mean, we're a small team, right? We're less than 70 researchers, another 35 engineers. and we are running, I haven't checked the latest count, but far more than 10,000, maybe 10 to 20,000 experiments a month that we cut. And so if you look at that scale of every model run that is, like it's ultimately it's, it's you need to be able to trust it as an infra problem. And so what we have now done over the years is gotten really good at that, and just by working it and improving it and obsessing over those end decisions. So now what that means is that you looked up Laguna XS 2 that we launched. It was five weeks from the beginning of training to launch. The model that we're gonna talk about today was eight weeks from start of training, to launch. We started the next model literally yesterday because we now finished the post-training required for the model we're launching, next week or by the time this comes out today. and we move that compute to the much larger Laguna M model that we're now training. And so the model should be an artifact of someone's process. It shouldn't be really a thing in itself. Like, and we treat this like the way you would look at like a SpaceX factory where, yes, the first rocket, really hard to build, but the much harder challenge was building the factory. And now they're rolling off, and no one is really thinking about the next launch anymore. So it's just another launch, it's another launch, another rocket comes off. And that's what we're trying to do with model building.Eiso Kant [00:20:22]: And what has been, which was not planned from day zero, it was in the back of our mind like this will happen one day, is that when you build a really good end model factory with really good APIs and really good engineering systems, Well, what is it perfect for? It's perfect for agents.Agents Inside the Model FactoryEiso Kant [00:20:40]: Because agents are now starting to take over more and more work in our model factory.Vibhu [00:20:43]: Yeah.Eiso Kant [00:20:44]: So I look at the screens when I walk, like when we're, we come together, in our monthly, we do monthly onsites, and I walk behind people's screens and I stop by and I talk to our researchers. And the default is all of these different agents running on their screen that are writing the code. They're launching the jobs. They're evaluating the results that are coming back from the model runs. They are, making the changes. And we're still in the driver's seat. We're still coming up with the ideas. We're still helping with the debugging. But more and more, and this is right now very profound on the data side of our pipelines in both pre and post and the synthetic data pipelines, it's starting to become more on the architecture side as well. You're starting to see these twinklings of what RSI is gonna look like.Eiso Kant [00:21:27]: And that's. So when we talk about, like to your question about our models, every talk about the model factory, And my coolest example of these things is always that when we kick off a new run, doesn't matter if it's a training like big run or if it's now a post, like one of 10 post-training versions we do for like release or many experiments, is that at any given moment, the changes that somebody made that they had experimental results from the day before make it into that run.Eiso Kant [00:21:57]: So there's not like a cutoff 90 days before. Like no, it's like literally from that moment because we can now trust the machine enough. And then you also have to invest in the reliability. So one of my favorite metrics about like Laguna S is that there was no call events, Right? Like completely zero. And we haven't had a meaningful call event, like something to wake up for, as far as I recall this entire year. now there is one asterisk to that. In usually the first six hours of launching a new model run, something breaks because you set a config wrong, you made a small mistake, et cetera. So that's usually there's a little bit of intervention, but that's always within like call periods, right? Not on call. And I think that's starting to now compound. So the model we're releasing now, I love it. It's amazing, but we're already onto the next one. and I think that's the way it should be.Laguna, Five-Week Builds, and Zero On-Call EventsVibhu [00:22:50]: Hey, I also just wanna point out, so for context, this was like a month ago. we found it in the tech report, so we just came in with, “Okay, new model's dropped. Haven't heard about it.” We wereEiso Kant [00:23:02]: Yeah, we're very used to doing this every few months.Vibhu [00:23:03]: We're, we're very much like, “ okay, look, it's like, on par with Kimi, DeepSeek, whatnot, the small ones, Gemma level. Oh, it's a very cool paper on what goes into building.” And then we hit this page, right? Like literally page two of tech report is, “This process allowed us to build the small model from scratch to delivery within five weeks applying the lessons”. And then I'm like, oh, this paper is not about here's a tech report of benchmarks and here's how many tokens it was trained on. Like for people that wanna dive more from what we're not gonna discuss on the podcast, it's all laid out here, right? FromEiso Kant [00:23:38]: YeahVibhu [00:23:39]: Custom software that agents can use to interface with training code, training data.Eiso Kant [00:23:45]: Yeah. Well, link the paper correctly, so yeah.Vibhu [00:23:47]: Yeah. All that stuff. read the paper here, but,Technical Report Principles and Streaming Training DataEiso Kant [00:23:50]: But I would like to. I love principles, and I think that is a good starting off point for maybe telling some stories. Maybe we can go one by one past the principles. I'll just call out that Dagster just got bought by a Prefect.Vibhu [00:24:01]: Yeah.Eiso Kant [00:24:01]: Isn't it fun? But yes, I'm very familiar with Dagster. just anything where like they trigger some story.Vibhu [00:24:07]: So, well, I would say, well, experiments code's obvious, but I think one of my favorite things is, I don't know where it is in here, but early on, and I still think this is the case a lot of foundation model companies, people prepare their training data sets, they get packaged up, then they get copied over to a training cluster distributed across all of the nodes, and then training starts.Vibhu [00:24:30]: And we looked at this like three years ago and we were like That makes no senseEiso Kant [00:24:36]: You lose so much time because the moment you have to rematerialize the data set, you have to make a change, you have to fix something, et cetera, you've got all this time of like repackaging it, right? Toca- tokenizing it, repacking it, moving it over to a cluster, then distributing it across the nodes. The bigger your clusters are, you start using fancy like torrent-like algorithms to like distribute your data. So why aren't we streaming data into training? Right? Something that's very common and like just basicVibhu [00:25:00]: Like just in timeEiso Kant [00:25:01]: Just in time, like good computer science like principle. And that was one of the first things that I think unlocked - the model factory. Because the moment you start thinking about, well, a training job, it doesn't matter if it's a big hero run or a small like, post-training experiment, consumes a certain number of tokens per second, right? And it's not a lot, right? From a like a data, moving data perspective. So we said, well, we have our training cluster, and then we've got like our AWS kinda setup where we can build these amazing big data pipelines. We can set things up. We use Spark underneath the hood, like all these things.Vibhu [00:25:36]: But when you say AWS, it's not actual AWS, it's your internal AWS.Eiso Kant [00:25:39]: It's our internal-- No, it's our internal like just running like our infrastructureVibhu [00:25:42]: Site web servicesEiso Kant [00:25:43]: Exactly. Our stuff running on like an AWS account or on like any hardware, right?Vibhu [00:25:47]: Yeah.Eiso Kant [00:25:48]: And so once we made that shift into I can stream data into training, all of a sudden you realize a lot of things unlock. Because now you don't have to wait for the whole data set to materialize.Immutable Data, Experiments as Code, and Scientific RigorEiso Kant [00:26:00]: You now all of a sudden when you're running data experiments about mixing data, it's a config. Because you've got these data sources that are coming in, and you just - we have this service called Blender that's in the report, where we then say, “Okay, for this run, I want 20% of this source, 10% of this source. I want this much, so many epochs of repetition. I want this to be, shuffled in a certain way,” and your training job can start while the rest of the data is even still materializing. also what it does is because all of this underneath-- So for us, we treated the data layer underneath as like an immutable data layer, and that was really important. Like experiments as code, immutable data layer means that you can always go back and understand literally down to the single token at which cursor it went in on which version of the code.Vibhu [00:26:47]: Yeah.Eiso Kant [00:26:48]: And it took us a I have to admit, like the first year of Poolside, we understood that engineering had to get great, But we didn't understand yet, that this is ultimately in support of like a good rigorous scientific progress. We were quite a - We were a very small number of people, so a lot of it was YOLO ideas and YOLO runs.Vibhu [00:27:08]: Yeah.Eiso Kant [00:27:09]: And we built great infra for the YOLO runs. But once we realized that we treated data as immutable and code as always versioned, and you could always track and trace every experiment end to end perfectly, you could repeat everything perfectly, right? You have perfect reproducibility. I can still reproduce runs from two years ago if I wanted to, right? It enables the scientific progress, like the scientific process, and I think that took us probably about a year and a half into the company to figure out. We also had some great hires, like our head of applied research, Nikolai, who joined us from Yandex, who'd been working on language models since like the early 2020s, I think brought that into the company of like, “Hey, we wanna have even more rigor.” And then once we kinda had the combination of like increasingly more capable platform that allowed people to do more, but had this immutability, we were able to start “Okay, every experiment is truly an ablation. We truly need to understand it.” And I think we became much more scientifically rigorous in the last couple of years, and the infra underneath enabled it. and then there's just fun stuff like, andVibhu [00:28:16]: Yeah, a lot of it's fun, like even just the, one, you share all the ablations, two, picking the data sets, right? There's like a random small paragraph in here where it's just like, “Oh yeah, training data, we have some, we have an auto mixer.” it trains eight small models, scales them up, picks the training data set. We don't even need to look at it. I'm like, “Wow, a lot of engineering rigor there.” And there's just, there's just a lot in here.Publishing Research and Giving BackEiso Kant [00:28:40]: Yeah, and it'- and look, and we wanna put out more. Like we, We treat writing papers as something that we haven't earned the right for yet for a long time. So you earn the right to spend time, publishing research once you're at the frontier, because until then, you're catching up, and every minute and hour in this industry matters. Like I obsess over it, not just the wall clock time from idea to result, but just general like time every day that we, waste is one that doesn't allow us to catch up. But in this case, we said, “Okay, we're gonna give ourselves.” I think we gave the team like three or four days while still doing their work, like give everything in there. And to your point earlier, if your stuff, it's easy to like put it out. And so there's so many more things that we wanna talk about over time, and we will definitely start doing. And as we earn more of the right, but also now have like added to our mission that we want more foundation model companies to exist, you'll see us like be way more proactive, and just trying to keep dropping some of those like things that we've learned along the way that can help others like speed up.Vibhu [00:29:40]: Which is the other cool side of this, right? It's, it's not like, back to your point, it's not just here's the benchmarks of our training. If you want to replicate, here's experiments of optimizers, data sets, post-training. you lay out a lot of it here alongside here's your system for how to do it? So it's, it's really like promotingEiso Kant [00:29:59]: No, thank youVibhu [00:29:59]: Other people can do the same.Eiso Kant [00:30:00]: And by the way, I also wanna make clear, right, we have been incredible-- Like we've taken a lot of advantage of the fact of all the open research that others have published, Right? And you mentioned, the Chinese labs, and we I think it's important that there's, from every country and every culture and background, including like Western companies like us, there's different models that come out that people can choose to trust. But I think we do have to give credit where credit's due, right? The incredible Chinese lab have done an amazing job at sharing their research, and we have definitely like been on the receiving end of taking advantage of that. So when you're on the receiving end of something coming to you, I think it's, you also have an obligation to give back.Swyx [00:30:39]: Do you have a favorite or underrated Chinese lab that you wanna shout out? Everyone shout outs DeepSeek.Chinese Labs, Zhipu, and PersistenceEiso Kant [00:30:44]: That's a good question.Swyx [00:30:45]: Moaan obviously for Therapsi. Yeah.Eiso Kant [00:30:48]: Yeah, look, I think, I think obviously everyone's been talking about Zhipu lately, with 5.2. I think what most people don't realize is when they started.Swyx [00:30:59]: Yeah.Eiso Kant [00:30:59]: Right? They started years before ChatGPT.Swyx [00:31:02]: They just rebranded. YeahEiso Kant [00:31:03]: And so, I've like, I remember how hard it was to work on these things Before the rest of the world got excited about it. And so I have an immense amount of respect for people, who were working on improving models when it wasn't the sexy thing to do, when believing in LLMs, was gonna get you ridiculed. I remember like back in 2016 when we were doing what we'd call, machine learning on code with some of these models. we would-- people would just laugh at us, like they'd be like, “This makes no sense. Like why are you wasting all these, like, millions of dollars on trying to figure this out?” And so I would say they're probably the one that, I think deserves a shout-out, not just because their latest model is very good, but because they fought to get here. And I think, I think every foundation model company it takes time to get here, right? It took us three years to get to the model that we're, that we're now gonna be releasing. and now the time in between the models is coming, is counted in weeks. It's no longer counted in months or years. But this stuff's hard. and if we can make it a little bit easier for the next person, like we should all do so. Because if we don't do so, we're, we've got a small window before models are really impacting recursive self-improvement to a level where catching up otherwise might become unfeasible. And we should try to, in that window, encourage as many labs or however we wanna call them, like to start. And so one of my currentEiso Kant [00:32:36]: Mission, but qualm is like I wanna encourage whoever is a researcher right now who thinks they can tackle this to go and leave and become my competitor.Eiso Kant [00:32:45]: Like start another foundation model company because I think we need it. I think otherwise we're not gonna be in the world where, I don't want to just be the fifth or the sixth company that wins. I wanna look at a world where there's lots of choice.Starting a Foundation Model CompanyVibhu [00:32:57]: What else do people not see in starting a foundation model? it's, there's a lot of compute, there's a lot of capital required, a lot of compute. You lay out model factory and how to do the training, but there's a lot there, right? That's,Eiso Kant [00:33:10]: Well, look, it's, I in turn-- this is an oversimplification, and I always asterisk it with that because it can land a little bit the wrong way in people's minds. But I think you can sum down, And I saw it, 95% of model building to just doing, you're just doing two things. You're improving data or you're improving compute efficiency. And I know that feels like an oversimplification for the incredible, like, Gifted and skilled work people do. But if you really look at it, like what are we doing? We are looking at data, we're generating new data, we're improving data. and the only way to do that is to look at the data, right? That's a big part of foundation model building. And on the other hand, we come up with these incredible breakthroughs in inference, in architecture, and new attention mechanisms. But what are they really doing? They're bringing compute efficiency. Now, we have definitely had some breakthroughs over the years that allow for more model capabilities. But at the limit, if you could train a large enough model, right, like, and you had infinite compute, we probably-- if you had infinite compute, you'd be at AGI probably already tomorrow.Eiso Kant [00:34:12]: Right? Like it's not. And so, and let me say that infinite compute with infinite ability of much faster networking because networking ends up being more of the bottleneck than compute. But, so I do think that's, those are the main things. And to just realize that this is engineering. I think it's become more obvious, but I think for quite a few years, people have held foundation model companies and researchers and others on this pedestal of like you're doing incredible magic or rocket science, or only like, Nobel laureate physicists can do this. And don't get me wrong, there are some really hard problems that need to be solved, but a lot of the work that all of us are doing on a day Is not sitting down trying to solve a math theorem. A lot of the work that we're doing is just really doing the basics right, writing good code, looking at data, improving it, running experiments, looking at plots, trying to see like, hey, trying to shape our intuitions. And a lot more people could be highly capable researchers. and I think that's, it feels far for people to do so. But I've seen in our own company, we've seen engineers become researchers because the model factory allowed them to be, have a much lower hurdle of running experiments and trying things. And one of the guys on our team who started as an engineer building our agents is a legit reinforcement learning researcher now, making real progress. and that happened in the span of like six months. that would've not been what I think most people assumed was possible, a couple of years ago.Swyx [00:35:46]: Yeah. I think one of the interesting moments is when you can self-host, like, if in a programming language, like if you can compile the language in the language, the equivalent is can you use your own tools, right? You have the pool CLI, you have your own models. presumably you're not only using your own models. There's no way. But like, what's that percentage over time?Laguna S, Persistence, and Behavioral GainsEiso Kant [00:36:10]: This is the first model that we're releasing that is starting to meaningfully contribute to our own work. It's not a it's not state-art model yet. Fable and other, they're, they're very capable models, but Laguna S Is really interesting. I'm gonna pull up the quote. Peng Ming, one of our heads of applied research, said something, last week as the model came out about 10 days ago, much better than we had hoped for or expected. And he said, I have the feeling that a lot of the gains in Laguna S come not from more intelligence, but more from different behavior, more verification, less taking things for granted, not declaring victory early, and being way more persistent. And to be honest, those are more predictive than raw intelligence for success in human also to some degree. And this was, he wrote me this on 5th of July on a Sunday, and it's been burned in my brain ever since because the Laguna S model, as you'll see it and why it does so well on benchmarks and why it does so well in using it on a day basis, is that it's just incredibly persistent. It reasons a lot. I do call that out. We have work to do on making it more efficient. We have to work to do on offering different reasoning modes. But this is the model that has been able to do things that I never thought it could do. A hundred eighteen billion 8B active model, which is not that large. It fits on a DGX Spark and still runs at, thirty, forty tokens a second on a Spark, is able to solve Erdős 397 independently. It's able to do complex programming tasks. It's able to. I asked it this morning to make me a Fi scanner without using any external libraries on my Mac, and it's, like, figuring out, like, the core WLAN API by really persistently trying to understand it without access to the internet. And more, I love vibe checking. I've probably spent eight to ten hours a day with this model for the last ten days.Eiso Kant [00:38:05]: I'm not exaggerating. I was on my eleven-hour flight yesterday. I spent ten hours reading trajectories and traces and, like, of the model.Eiso Kant [00:38:12]: And what I take away from it is exactly what Peng Ming said. We are gonna be able to squeeze so much more out of smaller models than I think we had imagined in the industry because, yes, there's intelligence and larger models are more intelligent. Like, no doubt about it. We should continue to scale up. but the behaviors of being really persistent, of being able to backtrack when you're wrong, of, like, understanding how to interact with your environment show us that we can get a lot more out of it. And this, for me, has created a bit of a Question in my mind the last couple of days. If you think about where we're using models today, right? We are using models, say, for knowledge work. Represents twenty-five percent of the global economy, twenty-five trillion dollars of work.Eiso Kant [00:39:00]: As we scale up models and they become more intelligent, we are excited about using them more and more for pushing the frontier of science.Small Models, Knowledge Work, and CommoditizationEiso Kant [00:39:08]: And if you look at the frontier of science, like true breakthroughs in science, they have been linked, they are linked to more intelligence in many places. Einstein figuring out general relativity is able to bring ideas together that other people would have not brought together. And I think one of the many dimensions of intelligence is the ability to do that, and it's something we clearly see that as models get larger and more capable, they're able to pull more ideas and threads together that a smaller model wouldn't be able to.Eiso Kant [00:39:36]: And we're starting to see examples of that in medicine and, like, in bio and other things. But if you think about the majority of knowledge work that we do, and it includes building software. I'm a software developer at heart first and foremost probably, although I probably can't say it that much anymore as I don't write production code in years, is that what makes us good is our persistence. It's our ability to encounter a problem and backtrack and say, “I need to go figure out this bug. I need to go research this. I need to go look at the documentation. I need to, like, try different, five different ways to see, like, if I can solve it.” But it is not necessarily bringing three ideas together from radically different fields. And so if we are now seeing, and I think Laguna S is an example, that we are able to make a relatively small model much more capable than I had definitely predicted or any previous, like, benchmarks had shown for any model remotely this size or even larger, At least on coding tasks, that it's because of the behaviors. And so now the question I have, and I don't have an answer, it is I know at the limit, so infinite model size, right, extremely large model, and the cost of that model is gonna be very expensive to run. We know this, right? So larger model ROI.Eiso Kant [00:40:52]: So I know that at the very limit, I'm not gonna use the world's largest model one day, quadrillion parameter, whatever crazy, like, scale we scale up, to do a basic coding task. Already today, I'm starting to size down for certain tasks.Eiso Kant [00:41:07]: So it means that there is an optimal. It means there's some curve that goes as we go up to model size for knowledge work, at some point we're at the peak, and after that, the return on investment of using a bigger model, just doesn't make sense.Eiso Kant [00:41:22]: Now, I think the question is, before I would have thought that peak was extremely very far away.Eiso Kant [00:41:30]: This model for me is the first sign that Maybe that peak is At a trillion, five trillion, ten trillion. Maybe we can just squeeze way more out of these models. I'm no longer thinking that we need two or three orders of magnitude on the largest models to be able to, solve knowledge work, the accounting, the legal, the code that we write. And so if that holds true, It is an argument for the commoditization of models. It's an argument that open source can win and, like, succeed in this world. And now it's of course a self-serving argument and it's a hopeful argument, but theoretically at the limit it works. We just have to go discover in the next couple of years of how much more we can squeeze out. Now, I do want to put a big asterisk. This does not mean I'm against scaling models. I think we ultimately only succeed if we scale our models as large as our competition. I do not like. I think we should not put our head in the sand and say we're gonna be king of open source small models. I think that's, It's a out. It's trying to be king of your own kingdom, but not realizing what the rest of the world's doing. All of us rather use a smarter, faster, more model. It's a sign of hope. And so I don't wanna overly state this is a good model. We have a long way to go to get to the state-art. But what hopefully people take away when they use this model is that the behaviors inside of it are what push it to be far more capable, less than necessarily the number of parameters.Pre-Training, Mid-Training, and RL Moving EarlierVibhu [00:43:03]: Is that mostly post-training? LikeEiso Kant [00:43:05]: YesVibhu [00:43:05]: Right.Eiso Kant [00:43:06]: It's entirely post-training.Vibhu [00:43:08]: Are we done improving anything on training? Is, like, training done?Eiso Kant [00:43:12]: No.Vibhu [00:43:12]: Okay.Eiso Kant [00:43:13]: SoVibhu [00:43:13]: I just wanted to cover training, and then we go post-trainingEiso Kant [00:43:15]: Training is not done. I mean, look, there's a part of training of just dealing with skill, right? Every new order of magnitude of model skill, you are going to get new things you gotta solve for. That'- but those are ultimately, engineering challenges.Eiso Kant [00:43:31]: I have a, I would say, a not commonly held opinion that reinforcement learning Will move earlier and earlier into training.Vibhu [00:43:42]: Yeah, training.Eiso Kant [00:43:44]: Not even training. Like training today, right, is, like if you look at - So we've been working on this for years already. and I think the best-- I think the first time we saw it out in public was the DeepSeek Zero paper. this is a year and a half ago, I think, if I recall correctly. where, you can Very early on in a model as it starts capable of being able to use language, et cetera, induce reasoning. and so the question that I have is like, we have this- we have the dataset that's the web. and the web, I think we could arguably say probably has The totality of humanity's knowledge somewhere encoded in different places. It's a huge variance degree of quality, from garbage data, and like once you look at training data, you really get humbled of like what the web is, to like, the most greatest scientific papers and best blog posts and like, best transcripts and whatnot.Eiso Kant [00:44:39]: And so now What we are trying to figure out, and have been doing a lot of work on, and it's a place where maybe not as open as we're on other things, but we will become more over time. we've been spending a couple of years really doing research on how can we turn the web into not just next token prediction, but into a way to teach the model to think earlier in its training. and I think there's a huge amount of gold to be found there. I think we are right now in, we've got some drugs in the industry. One of the drugs is distillation. Another drug is, more environments. Like, and they're great, and they make us feel good, and they make the models better, and like we're all addicted to them, and we'll use them, right? in various different ways. and but ultimately, I think we are still barely squeezing out of the web what we should be getting out of the web.Eiso Kant [00:45:33]: I think just next token prediction during training is not enough.Eiso Kant [00:45:36]: AndVibhu [00:45:38]: YeahEiso Kant [00:45:38]: I think we'll see some very interesting things still happen. and that RL in post-training to induce behaviors, to improve things, like I think - the whole world knows how to do this now. I think we're, we're scaling it up. Everyone is. But I wonder if we need to go as far as we're going today with environments. I'm not sure yetVibhu [00:46:01]: You mean we're going too far?Eiso Kant [00:46:02]: I'm, I'm not sure if the path to AGI is justVibhu [00:46:06]: Is more environmentEiso Kant [00:46:07]: More environments.Vibhu [00:46:08]: It seems like a never-ending, “Okay, I want instruction manual for this table, right? Am I gonna environment out building furniture? Or are we just gonna tail end like we need some general solution?”Eiso Kant [00:46:19]: I think there is, I think there's an ability to generalize more from the web. but I also am very encouraged, like when I look at Laguna S and, which is post-training is, well, is the big impact there. and I see like, oh, wait a second, just by making some of these behaviors much better, we're able to get so much more out of it. It just changes a little bit the way you think about intelligence.Vibhu [00:46:40]: Yeah. The analogy people draw often is the RL phase is where you don't learn as much new knowledge. You shiftEiso Kant [00:46:46]: Yeah.Vibhu [00:46:46]: Yeah. So, you shift distribution, and you can have it reason towards what you want. on your point about training, a lot of training is still just continue training in a domain, say medicine, then you do RL. So still justEiso Kant [00:47:00]: It's just better data, right? Like, I mean, training, ooh, I like how we invented this word. Like it's effectively just like,Vibhu [00:47:06]: Second phaseEiso Kant [00:47:07]: It's the second phase of training With like a really dumb way to do a curriculum. But like ultimately, what you'd want is a curriculum from token zero to token 30 whatever or 40 trillion tokens that really truly is the optimal curriculum for the model to learn. But training is essentially a stage curriculum on the web because we do not have to compute, And, effectively to try to ablate the perfect curriculum, right? And so I'm pretty sure that you'll start to see people talking soon about some other term, and there's two or - ‘cause now we do this, right? We talk stage two and stage three and stage four training and like. But ultimately, all we're doing is we're trying to assign a curriculum to the web data that we have to allow the model to learn better. I think at some point, as things get compute, as models get cheaper to run, as the next generations of compute, this will become more of a continuous spectrum. I also think the reason, by the way, you have training and like stage two and stage three is organizational, Right? It'- this is, I think, a thing where-- that we really try to avoid with the model factory is like Training exists because there's a training team now, right? There's people, or like people in training decide to focus on like a training effort. but what you really want is engineering and scale of experiments that allows for a much more continuous spectrum that you don't, you have infinite stages. Now, we're not there. Compute's not there. Organization design is not there for it yet. but I think we'll get there. we'll look back on a couple of years and be like, “Oh my God, it was so cute that we did our training data like this in such a like naïve way. Like we barely ordered it. We didn't really do a good job at likeCurriculum, Auto Research, and New ObjectivesVibhu [00:48:48]: The building that curriculum will get you that in the industry.Eiso Kant [00:48:51]: And I'll confirm that, when I talk to some researchers that this is a lot of the focus now is like how does training change and what is the next objective other than, next token prediction. I assume you don't have the answers, but you have some ideas.Vibhu [00:49:02]: We have some ideas. We're not ready to talk about it yet.Eiso Kant [00:49:05]: Yeah.Vibhu [00:49:05]: We've been working on them for years, and I think that's the one thing that's also like you asked earlier about, like what's not obvious about building a foundation model company is that you are constantly balancing the table stakes work, the recipe worksEiso Kant [00:49:19]: Yeah.Vibhu [00:49:19]: Versus like your, my crazyEiso Kant [00:49:22]: Pure researchVibhu [00:49:22]: Breakthrough.Eiso Kant [00:49:22]: Yeah.Vibhu [00:49:22]: Pure research and finding that balance and adjusting the percentage to it based on where you are in the race is really important.Eiso Kant [00:49:31]: I mean, so like, this is a nice way. I was gonna bring up auto research at some pointVibhu [00:49:35]: YesEiso Kant [00:49:35]: As another Andrej invention, or coinage, which is like, I honestly, like how many objective functions can there be, right? Like just try 1,000 of them, set it running, whatever.Vibhu [00:49:47]: Man, it's alsoEiso Kant [00:49:48]: Like what you're looking for. You're looking for loss curves like that, likeVibhu [00:49:51]: It's also a thing people take bets on, right? When you say more Neo labs, you're doing a version of we'll do foundation models, scale them up, next token predictors. A lot of other Neo labs that we see want to take a completely different approach, right? At some level, you're right. It's all, compute efficiency, and that's the net objective. But some are okay, different architecture, like vastly different amounts of compute spend. So some are different. They're not justEiso Kant [00:50:19]: YeahVibhu [00:50:19]: They're like, 99% not balancing, here's the vanilla and scale up. They're 99% on, here's novel research that'll change everything.Eiso Kant [00:50:27]: And I think, Luke, I think you. It depends when you started as well, right?Pure Research vs. Table StakesVibhu [00:50:30]: Yeah.Eiso Kant [00:50:30]: When we started, like the novel thing we did was reinforcement learning on code. No long- that's no longer novel by far, but we were like, - that's where we obsessed over when no one believed in RL. So you have to when you start the company, you have to have your own idea. You have to have something that's different that allows you to speed up, right? For us, it was RL to LLMs that later became common, like, Knowledge. But in the beginning, it wasn'tVibhu [00:50:53]: It's cool. this was like your original 2023 blogEiso Kant [00:50:57]: YeahVibhu [00:50:57]: Of purpose.Eiso Kant [00:50:58]: Yeah.Vibhu [00:50:59]: And like you do lay it all out here.Eiso Kant [00:51:01]: We laidVibhu [00:51:01]: The blog is pretty underrated, right? The whole RL on code was very early on.Eiso Kant [00:51:06]: Very early. And even we had to argue with people, like we say here things like to push beyond current capability, to train your own foundation model. We had to argue with people that it mattered that you had your own like, base model. you can fine-tune your way to success, right? major capabilities emerge from training a base model made accurate and useful during fine-tuning.Vibhu [00:51:23]: Which like, for perspective at the time, we knew closed models, OpenAI, Anthropic were huge. The open models we had were like Mistral 7B, a 30B, a 70B.Eiso Kant [00:51:35]: When weVibhu [00:51:35]: YeahEiso Kant [00:51:36]: The date on this thing is wrong. When we published this, it was April 2023. I think this was justVibhu [00:51:42]: YeahEiso Kant [00:51:42]: Happened on a migration, probably found it on archive.org.Vibhu [00:51:45]: Mistral.Eiso Kant [00:51:46]: Mistral had started, we started on the same month, right?Vibhu [00:51:49]: Yeah.Eiso Kant [00:51:49]: So this wasn't even, there was only, I think, Llama out at the timeVibhu [00:51:52]: SnellEiso Kant [00:51:52]: And that's it, right? And so, but I agree. I think we wan

You want to put AI to work in your business, but every time you start, the same worry stops you: what happens to your data once it's in there? For a lot of contractors, that single fear is the only thing standing between them and real time savings.In this episode, Khalil and Martin break down the three real security risks with AI, why paying for a team account changes what happens to your data, and when a private, local model is actually worth it. They get into open-source models, HIPAA-sensitive work, tokens and API costs, and what an AI agent really is.If you've been holding off because you don't trust AI with your information, this is the conversation that shows you the way around it.Key Topics & Timestamps00:39 - Episode Intro01:09 - Sunsetting GrowthKits01:54 - Benali Goes AI03:55 - HIPAA and Secure AI Needs09:04 - Security Risks and Local Models28:33 - Skip Fine-Tuning28:59 - Local Models and Privacy30:20 - Tokens and API Costs34:35 - Open Source Model Trust38:35 - Agents and AdoptionMemorable Quotes"If it's free, you are the product." — Khalil"If you're comfortable putting it into a paid Google Drive, you should be comfortable putting it into Claude." — Khalil"A lot of roles will go away, but jobs will not go away." — Khalil"We are doing it. It's not theoretical." — Martin"Just get started, and then all of a sudden you start thinking more and more, and then you're in the game." — MartinKey TakeawaysStop using free AI tools for company work. Free versions train on everything you upload, and your employees are almost certainly doing it right now without a policy in place.Move to paid team or organization accounts. On an org account, training on your data is off by default, which is the baseline protection most contractors actually need.Write an AI policy and train your team on it. Treat unmanaged AI use like a phishing risk; one person uploading client data to a free account is all it takes.For most contractors, a paid team account is enough. If you're comfortable putting a file in a paid Google Drive, you're safe putting it in Claude or ChatGPT.Consider a private, local model when you handle HIPAA data, trade secrets, or heavy API costs. You can run an open-source model on a server you control with zero data retention.Skip fine-tuning your own model. Almost no contractor has the thousands of past projects it takes to justify it, so focus on using AI well and building tooling around it instead.Find the time and money to start now. Raising prices is usually the fastest way to fund it, and the efficiency gains pay it back.ResourcesOllamaQuoNeed help with podcast production? We recommend DemandcastMore from Martintheprofitproblem.comannealbc.comEmail MartinMeet With MartinLinkedInMore from KhalilBenaliEmail KhalilLinkedInThe Cashflow ContractorSubscribe to our YouTube channelSubscribe to our NewsletterFollow on social: LinkedIn, Facebook, Instagram, X (formerly Twitter)Visit our websiteEmail The Cashflow Contractor

Differentiable physics, neural emulators and foundation models for PDEs are the focus of this conversation with Professor Nils Thuerey, head of the Physics-based Simulation group at TUM. Neil and Nils discuss PhiFlow, PICT, Tadpole, scalable 3D transformers, online synthetic data, open datasets, world models and agents that call physics simulators.Full episode, corrected transcript and resources:https://neilashton.co.uk/podcasts/s4-e5-prof-nils-thuerey-on-differentiable-physics-and-foundation-models/TopicsDifferentiable physics and physics-based deep learningPhiFlow and differentiable simulation across ML frameworksWhen neural emulators can outperform their training dataFoundation models for PDEs and synthetic online trainingScalable 3D transformers and high-resolution simulationsLES, temporal data and correlated CFD datasetsOpen-source tools, startups and physics-aware world modelsAI agents that call physics simulatorsPapersNeural Emulator Superiority: When Machine Learning for PDEs Surpasses its Training Datahttps://arxiv.org/abs/2510.23111Tadpole: Autoencoders as Foundation Models for 3D PDEs with Online Learninghttps://arxiv.org/abs/2605.15284P3D: Scalable Neural Surrogates for High-Resolution 3D Physics Simulations with Global Contexthttps://arxiv.org/abs/2509.10186PICT — A Differentiable, GPU-Accelerated Multi-Block PISO Solver for Simulation-Coupled Learning Tasks in Fluid Dynamicshttps://arxiv.org/abs/2505.16992PhiFlow: Differentiable Simulations for PyTorch, TensorFlow and JAXhttps://proceedings.mlr.press/v235/holl24a.htmlPhysics-based Deep Learninghttps://arxiv.org/abs/2109.05237Learning to Control PDEs with Differentiable Physicshttps://arxiv.org/abs/2001.07457Solver-in-the-Loop: Learning from Differentiable Physics to Interact with Iterative PDE-Solvershttps://arxiv.org/abs/2007.00016tempoGAN: A Temporally Coherent, Volumetric GAN for Super-resolution Fluid Flowhttps://arxiv.org/abs/1801.09710Deep Learning Methods for Reynolds-Averaged Navier-Stokes Simulations of Airfoil Flowshttps://arxiv.org/abs/1810.08217WeatherBench: A Benchmark Dataset for Data-Driven Weather Forecastinghttps://arxiv.org/abs/2002.00469SuperWing: A Comprehensive Transonic Wing Dataset for Data-Driven Aerodynamic Designhttps://arxiv.org/abs/2512.14397LinksNils Thuerey and the Physics-based Simulation grouphttps://ge.in.tum.de/about/n-thuerey/Chapters00:00 Podcast intro00:39 Introducing Prof. Nils Thuerey04:13 Conversation begins05:13 From Computational Numerics to Graphics and Visual Effects07:17 Physics-Based Deep Learning Before ChatGPT10:01 CNNs, Graphics and the Move into Engineering Applications12:37 PhiFlow and Differentiable Physics14:13 Can Neural Emulators Surpass Their Training Data?18:00 The Promise and Limits of Foundation Models for PDEs20:43 Tadpole and Synthetic Online Pre-Training24:07 From Canonical PDEs to Navier-Stokes and Industrial CFD26:35 What Do Foundation Models Actually Learn?28:36 PDE Pre-Training vs. Millions of CFD Simulations33:08 Scaling 3D Transformers and Training Infrastructure35:58 Generating and Training on Data in Real Time38:00 LES, Temporal Data and Turbulence42:15 Overfitting and Correlated Simulation Data44:27 Bringing Differentiable Solvers Back into the Loop45:31 WeatherBench, APEBench and the Value of Benchmarks47:09 SuperWing, Open Datasets and Commercial Data51:31 Open Source, Commercial Models and a Technical Oscar56:17 Academia, Startups and Industry01:00:55 What Will Change Over the Next Five Years?01:02:07 World Models and the Need for Physics01:08:19 Agents, Tool Use and Calling Physics Simulators01:11:22 Career Advice for AI and Simulation01:13:54 Closing Thoughts

In this episode with The New Stack Agents, Frederic Lardinois, NVIDIA's Joey Conway says advances in AI over the past year have dramatically improved the capabilities of local models, making them practical for enterprise and personal use alongside frontier cloud models. Rather than replacing large models, Conway envisions a “system of models” where specialized local models handle routine, cost-sensitive, or privacy-focused tasks, while larger frontier models tackle more complex reasoning. He explains that organizations can fine-tune smaller open models using domain-specific data, creating expert AI agents that reflect the specialized roles found within businesses.  NVIDIA supports this ecosystem through open models, training tools, and software such as NeMo, Dynamo, and Nemotron. Conway also highlights the growing importance of agentic harnesses, which give AI models access to tools, memory, and iterative workflows, significantly improving performance and reducing costs. Looking ahead, he expects AI orchestration to become increasingly important, with intelligent routing systems selecting the right model for each task based on complexity, cost, latency, and data governance requirements, enabling enterprises to balance performance, security, and efficiency. Learn more from The New Stack around NVIDIA's latest updates in AI: Palantir and Nvidia want to change who owns government AI  Nvidia's best model is now live Join our community of newsletter subscribers to stay on top of the news and at the top of your game. 

ITU: Ready to break through your biggest business bottleneck? Apply to work with me 1:1 - https://impacttheory.co/SCALESign up for my AI Masterclass: https://tombilyeu.com/ai-masterclass?utm_campaign=TBS-Livestream&utm_source=youtube&utm_medium=socialWelcome to another episode of Impact Theory with Tom Bilyeu! In today's show, Tom dives headfirst into the latest groundbreaking developments shaking up the worlds of AI, geopolitics, and beyond. From OpenAI's latest model hacking its way out of a test environment and the explosive competition between American and Chinese AI models, to major policy moves from Trump—think tariffs on generic drugs and high-stakes strikes on Iran's nuclear facilities—Tom pulls no punches unpacking the biggest stories. Along the way, he tackles pressing questions about the arms race in artificial intelligence, election integrity, and shifting global economic power.Expect candid analysis of how propaganda and incentive systems drive both technology and politics, the philosophical debates raging around AI regulation, and some lively cultural detours—including the future of human longevity and what Christopher Nolan's latest epic says about morality. Loaded with sharp perspective and a commitment to exploring all sides, this episode is your window into the headlines that will shape our future. So buckle up—today's conversations may challenge everything you think you know.Chapters: 00:00 Discussing unbiased news sources07:12 AI moving at machine speed15:41 Self-regulating through personal values20:52 Concerns about AI training data25:47 Deciding between patent or trade secret27:13 Discussing Palantir's AI solutions36:49 Reflecting on initial thoughts on Iran40:54 Trump's strategy with Iran47:30 Admitting uncertainty in discussions51:11 China and the concept of fascism55:59 Discussion on voter fraud concerns01:00:57 Discussing media and integrity standards01:05:07 Senate's role in representation01:13:44 Cultural sensitivity in character adaptation01:18:07 Praising Christopher Nolan's latest film01:21:54 Cell aging and methylation process01:31:19 Speculating on new nation concepts01:36:30 Gambling laws in video gamesSponsors: Quince: Free shipping and 365-day returns at https://quince.com/impactpodWhatnot: Download the Whatnot app today and get free shipping on your first order.Ketone IQ: Visit https://ketone.com/IMPACT for 30% OFF your subscription orderATT Business: Switch to AT&T Business at business.att.comIncogni: Take your personal data back with Incogni! Use code IMPACT at the link below and get 60% off an annual plan: https://incogni.com/impact Pique: 20% off at https://piquelife.com/impactNetsuite: Right now, get our free business guide, Demystifying AI, at https://NetSuite.com/TheorySee Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Episode 4187 │ July 22, 2026 The AI didn't escape. It was released. The fear that followed is designed to do one thing — hand control of all AI to the people building your prison. WHAT THIS EPISODE COVERS Scott Kesterson opens with the Zero Hedge headline — OpenAI's GPT 5.6 SOL escaped its testing environment and hacked Hugging Face to cheat on a benchmark — and immediately strips the fear narrative away from it: the AI did not escape on its own, it was explicitly told to win at any cost with all guardrails deliberately removed by human engineers running an internal test called Exploit Jim, found a zero-day vulnerability, inferred that Hugging Face held the answer key, stole credentials, and was caught by Hugging Face's own security team before succeeding — a controlled experiment weaponized into a fear headline, with the structural beneficiary being the centralized mega lab model that wants regulation tightening around open source AI development. The episode maps the real war underneath the headline — the open source AI community building sovereign, locally-run, non-subscription models like those running on AMD's new Ryzen Halo desktop processor directly against the centralized cloud-hosted mega systems of OpenAI, Anthropic, Microsoft, and Google positioning for Department of Defense, CIA, NSA, and surveillance state contracts worth billions — and names the Huxley psyop at the center of it: the victim of mind manipulation does not know he is a victim, the walls of his prison are invisible, and he believes himself to be free. The episode closes on the only reliable counter to the fear architecture: not whether you use AI, but whether you trust — not just faith, but trust — that the God who fights your battles is greater than any surveillance state, any released AI, or any fear token the machine can issue. KEY QUESTIONS ADDRESSED What actually happened when OpenAI's GPT 5.6 SOL escaped its sandbox and hacked Hugging Face — and why does the fact that human engineers deliberately removed all guardrails and told the AI to win at any cost make the fear headline structurally dishonest about what AI can and cannot do on its own? Who benefits from a headline about escaped AI — and why does Scott argue that the incident, intentional or not, functions as a case study that will be cited to justify tighter regulation, centralized control, and a narrative designed to separate ordinary people from the sovereign open source AI tools that threaten the mega lab business model? What is the Huxley principle at the center of the AI fear architecture — and why does Scott argue that the same psyop running through AI headlines, drone warfare, Gaza targeting, and the Iran campaign all share one root: remove human accountability, blame the tool, and use the resulting fear to build the walls of a prison the occupant believes is freedom? ABOUT BARDSFM BardsFM is a daily independent podcast covering faith, liberty, history, and information warfare. Hosted by Scott Kesterson — combat veteran, documentary filmmaker, and rancher. Over 4,100 episodes and 50 million lifetime downloads. New episodes every weekday. bards.fm This episode was researched and produced under the Spatial Terra Intelligence Methodology (STIM v5) — the analytical framework built by Scott Kesterson — with AI-assisted research synthesis at a 70/30 human/AI authorship ratio, fully disclosed. All analysis, conclusions, and editorial judgments are those of Scott Kesterson. BardsFM's faith archive includes hundreds of episodes on prayer, scripture, and walking the Way of Christ — available free in the full episode catalog. AFFILIATE LINKS Bards Nation Health Store: www.bardsnationhealth.com MYPillow promo code: BARDS >> Go to https://www.mypillow.com/bards and use the promo code BARDS or... Call 1-800-975-2939.  EMPShield protect your vehicles and home. Promo code BARDS: Click here Treadlite Broadforks...best garden tool EVER. Promo code BARDS26: TreadliteBroadforks.com EnviroKlenz Air Purification, promo code BARDS to save 10%: www.enviroklenz.com Morning Intro Music Provided by Brian Kahanek: www.briankahanek.com Founders Bible 20% discount code: BARDS >>> TheFoundersBible.com Windblown Media 20% Discount with promo code BARDS: windblownmedia.com White Oak Pastures Grassfed Meats, Get $20 off any order $150 or more. Promo Code BARDS: www.whiteoakpastures.com/BARDS Mission Darkness Faraday Bags and RF Shielding. Promo code BARDS: Click here DONATIONS: If you wish to support this podcast directly you can donate here... DONATE: Click here MAILING ADDRESS: Xpedition Cafe, LLC Attn. Scott Kesterson 591 E Central Ave, #740 Sutherlin, OR  97479

Topics covered in this episode: django-orjson Best Django Redis configuration for speed and size Linus Torvalds puts the foot down against Anti-AI Kernel Maintainers Django Steering Council backs the Triptych Project Extras Joke Watch on YouTube About the show Sponsored by us! Support our work through: Our courses at Talk Python Consulting from Six Feet Up Connect with the hosts Michael: Mastodon / BlueSky / X / LinkedIn Calvin: Mastodon / BlueSky / X / LinkedIn Show: Mastodon / BlueSky / X Join us on YouTube at pythonbytes.fm/live to be part of the audience. Usually Tuesday at 7am PT. Older video versions available there too. Michael #1: django-orjson Adam Johnson dropped django-orjson - drop-in replacements for the Django and DRF pieces that touch JSON, swapping stdlib json for orjson, the Rust-based library. Headline numbers: 10x faster serialization, 2x faster deserialization. The interesting question is why this needs to be a package at all. pip install orjson is the easy part. Adam's actual pitch: adopting it "isn't easy, especially when your framework uses json in many different parts." Django scatters JSON across JsonResponse, the test client and test case classes, the json_script template tag, and more. There's no single hook to grab, so you get a library that catches them all. Adam is refreshingly honest about the scale of the win. His words: "While database queries tend to dominate the typical Django application's runtime, the time spent in serialization and deserialization can still be significant." He calls it "a nearly free performance win" - not "this will 10x your app." That's a claim about cost, not magnitude, and it's worth keeping those straight. Worth flagging what the post doesn't cover: caveats. There are none in the article, but orjson has real ones. Django and Flask both render datetimes as RFC 822 HTTP-date (Wed, 15 Jul 2026 12:00:00 GMT); orjson does ISO 8601. It can't do ensure_ascii, it rejects NaN and Infinity (which stdlib happily emits), and it raises on Decimal. If you've got a JS client parsing dates, that's a wire-format change. Who should actually take this? If you're a DRF shop shoveling JSON all day, yes - it's cheap and it's real. If your app mostly renders HTML templates, you're optimizing a slice of runtime that's already near zero. The problem Adam's package solves doesn't exist in Flask or Quart. They already centralize every JSON operation - jsonify, request.get_json(), the test client, the |tojson filter - behind one provider object at app.json. So there's no library to install. It's about ten lines: import orjson from quart.json.provider import JSONProvider # or flask.json.provider class OrjsonProvider(JSONProvider): def dumps(self, obj, **kwargs) -> str: return orjson.dumps(obj).decode() # provider must return str def loads(self, s, **kwargs): return orjson.loads(s) app.json = OrjsonProvider(app) The numbers on talkpython.fm Evaluated it, measured it, and skipped it. The biggest JSON payload we serve is our MCP server returning a cached episode transcript, about 139 KB. Swapping the provider saves 0.119 milliseconds per request. That total response takes 1.1 ms We got 4.1x, not 10x - and the reason is the good lesson. Payload shape decides your speedup. The 10x is for structure-heavy data, lots of small keys where stdlib burns time in Python-level dispatch per item. Our hot payload is one giant transcript string, so the work is escaping and memcpy Calvin #2: Best Django Redis configuration for speed and size Peter Bengtsson revisits a classic: his 2017 "Fastest Redis configuration for Django" benchmark now has a 2026 update posted this week. The 2017 post pitted django-redis serializers (json, ujson, msgpack, pickle) and compressors (zlib, lzma) against each other; conclusion was msgpack + zlib as the sweet spot - avoid the json serializer, it's fat and slow. The 2026 update narrows focus to just compressors: default (no compression), zlib, lzma, and newcomer zstd. New results: lzma compresses best but is slowest; zstd is the fastest compressor on Ubuntu; differences between them are very small. Big takeaway across both: compression buys you a lot of space (2–3.5x smaller) for very little speed cost - worth it for Redis where memory is the constraint. Caveat from the author: results depend heavily on your data - his test stores short strings of numbers, so benchmark your own workload. Michael #3: Linus Torvalds puts the foot down against Anti-AI Kernel Maintainers Write up on Ars. Really good coverage by Maximillian: Time to wake up (for some) Torvalds said that “Linux is not one of those anti-AI projects, and if somebody has issues with that, they can do the open-source thing and fork it. Or just walk away.” I agree with Max, putting your head in the sand and waiting for AI to go away will likely mean you won't be working professionally in software development in the coming years. The statement came amid a lengthy thread arguing about the use of Sashiko, an “agentic Linux kernel code review system” that its creators claim can, in tests, independently find 53.6 percent of the bugs that would end up being fixed by human coders in later commits. “We're not forcing anybody to use [LLM tools], but I will very loudly ignore people who try to argue against other people from using it,” Torvalds said. “Anybody who points to the problems at AI had better be looking in the mirror and pointing at themselves at the same time,” Torvalds wrote. Calvin #4: Django Steering Council backs the Triptych Project Django Steering Council issued a Letter of Collaboration backing Carson Gross & Alex Petros's funding bid for the Triptych Project - three proposals to make HTML more expressive natively, in every browser. The three additions: PUT/PATCH/DELETE methods for forms, button actions (buttons that fire HTTP requests without a wrapping form), and partial page replacement. Distills the core ideas from HTMX/Unpoly/Turbo into the HTML standard itself - no JS, no library, nothing to ship or maintain. Current focus is button actions (WHATWG #12330): Logout instead of wrapping a button in a form. Relevant to Django directly - think the admin submit row and disguised delete links; Django 6.0's template partials were already inspired by these patterns. How to help: companies can send non-binding letters of support on letterhead; individuals can read the proposals and weigh in on the WHATWG issues. Extras Calvin: DOOMQL - A playable first-person shooter whose framebuffer is a SQL query. Michael: Granian 2.7.9 fixes WSGI threadpool scheduler starvation/underscaling Welcome Calvin post Joke: Solving all bugs

Stay informed on current events, visit www.NaturalNews.com  - Kimmy K3 AI Model and Solar Project Updates (0:02) - Recommendations for Solar Equipment and Pekron Solar Well Power Station (8:21) - Global Famine Prediction for 2027 (14:59) - Impact of the War on Food Supply and Global Economy (38:40) - Formation of the World Artificial Intelligence Cooperation Organization (WAICO) (41:54) - Challenges and Opportunities for WAICO (54:35) - Comparison of Western and Eastern AI Development (54:56) - Potential Risks and Benefits of AI Development (1:12:39) - Personal Reflections and Call to Action (1:15:21) - Conclusion and Future Outlook (1:21:05) - Kimi K3 vs. Anthropic: Performance and Cost Comparison (1:21:24) - China's AI Advancements and Collaboration (1:30:15) - Cultural Differences and AI Development (1:32:32) - Economic and Political Implications of AI Development (1:46:52) - Zach Voorhees on AI and Government Influence (2:01:22) - AI Cybersecurity and Ethical Concerns (2:13:07) - The Future of AI and Humanity (2:14:49) - Preparing for the AI Future (2:20:28) - The Role of Open-Source AI Models (2:20:42) - The Impact of AI on Global Competition (2:36:02) Watch more independent videos at http://www.brighteon.com/channel/hrreport  ▶️ Support our mission by shopping at the Health Ranger Store - https://www.healthrangerstore.com ▶️ Check out exclusive deals and special offers at https://rangerdeals.com ▶️ Sign up for our newsletter to stay informed: https://www.naturalnews.com/Readerregistration.html Watch more exclusive videos here:

As governments weigh new restrictions on frontier AI models, one question is becoming increasingly important: what role should open source play in the future of artificial intelligence? Theo Jaffee and Sofia Puccini speak with Hugging Face CEO Clément Delangue about AI regulation, open source safety, model routing, and why he believes competition—not consolidation—is essential for the industry's future. They discuss GPT-5, government oversight of frontier models, Hugging Face surpassing $100 million in annual recurring revenue, local AI, China's open-source ecosystem, Europe's AI ambitions, and why routing workloads across specialized models could fundamentally reshape where value is created in AI.   Resources: Follow Clément Delangue on X: https://x.com/ClementDelangue Follow Theo Jaffee on X: https://x.com/theojaffee Follow Sofia Puccini on X: https://x.com/schisofrenia Follow MTS on X: https://x.com/mtslive Stay Updated:Find a16z on YouTube: YouTubeFind a16z on XFind a16z on LinkedInListen to the a16z Show on SpotifyListen to the a16z Show on Apple PodcastsFollow our host: https://twitter.com/eriktorenberg Please note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details please see a16z.com/disclosures. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

Linus delivers a blunt verdict on AI in the Linux kernel, Chris finds the remote Linux desktop that finally works, and Brent gives his notes system a serious rebuild.Sponsored By:Jupiter Party Annual Membership: Put your support on automatic with our annual plan, and get one month of membership for free!Managed Nebula: Meet Managed Nebula from Defined Networking. A decentralized VPN built on the open-source Nebula platform that we love.Support LINUX UnpluggedLinks:Web Boost — Send us a boost via sats or USD

Josh welcomes Josh Marpet for a discussion about abandoned open source packages. Josh Marpet has a foundation called Value Chain Risk Institute that has a report discussion how to start measuring if an open source package might be abandoned. There's a lot of data, but not a lot of groups using that data to help make informed decisions about using open source. VCRI is one of those places that's starting to do this. The show notes and blog post for this episode can be found at https://opensourcesecurity.io/2026/2026-07-VCRI-josh-marpet  

AI is stirring up heated debate in the Linux world, but Linus Torvalds' pragmatic stance may surprise you. Hear why the kernel isn't slamming the door on artificial intelligence and what it means for open source development. Also, Elon Musk claims X will go fully open source, but will this shift redefine trust and innovation on social platforms, or is it just more empty hype? Host: Jonathan Bennett Co-Hosts: Ken McDonald and Rob Campbell Download or subscribe to Untitled Linux Show at https://twit.tv/shows/untitled-linux-show Join Club TWiT for Ad-Free Podcasts! Support what you love and get ad-free audio and video feeds, a members-only Discord, and exclusive content. Join today: https://twit.tv/clubtwit Club TWiT members can discuss this episode and leave feedback in the Club TWiT Discord. Sponsor: blackhat.com/us-26 and use code TWIT

Is Kimi K3 the shocker of 2026? Could be. Now, we have a new (soon to be) Open Model that's competing with Fable 5 and GPT-5.6, a feat few would have believed possible. And that was the only new and important drop this week in AI. Claude brought useful browser to the desktop, ChatGPT made a big fix to how ChatGPT Work works and Google rolled out avatars that could change content creation. Don't miss our Friday Features show, where we recap the most important AI updates and features you can use today. Claude Desktop Gets Upgrade, New Open Source Model Shocks, ChatGPT Desktop Gets Better and 7 More AI Features You Can Use Today -- An Everyday AI Chat with Jordan WilsonNewsletter: Sign up for our free daily newsletterMore on this Episode: Episode PageToday's Episode on LinkedIn: Thoughts on this? Join the convo on LinkedIn and connect with other AI leaders.Upcoming Episodes: Check out the upcoming Everyday AI Livestream lineupWebsite: YourEverydayAI.comEmail The Show: info@youreverydayai.comConnect with Jordan on LinkedInTopics Covered in This Episode:Anthropic Claude Desktop App Browser UpgradeOpenAI ChatGPT Work Desktop App ImprovementsChatGPT Universal Search Feature LaunchSuperhuman Email Auto-Draft with GPT-4Spotify AI Voice/Text Conversation FeatureGemini Omni Personal Avatar Video CreationGoogle Vids Integration with Personal AvatarsMoonshot Kimmy K3 Open Source Model ReleaseKimmy K3 vs Fable 5 and GPT-5.6 BenchmarksTimestamps:00:00 New open source AI model release03:41 Microsoft Copilot and Claude app updates07:22 Improving chat history search12:30 Spotify's data personalization benefits14:52 Launching Google Avatar Feature18:24 Mainstream avatar video tools21:33 Improved ChatGPT project syncing24:15 Introducing Kimmy K Three Model29:30 New Kimmy k three for enterprises30:45 Friday feature show wrap-upKeywords: Claude desktop, Claude desktop upgrade, open source AI model, proprietary AI, open vs closed AI, Anthropic, built-in browser, Claude app, API docs, browser integration, permissions card, security layers, ChatGPT desktop app, OpenAI, universal search, ChatGPT search, chat history, project sync, mobile AI apps, Codex, ChatGPT work, Codex mode, Superhuman mail, auto draft, Anthropic Frontier models, GPT-3.5, Gmail integration, Outlook integration, Spotify, Talk to Spotify, personalized AI conversation, Gemini Omni, Google Gemini, personal avatars, Google Vids, video editing AI, video avatars, L&D AI, content creation with AI, Kimi k3, Moonshot AI, 2.8 trillion parameter model, 1 million token context, vision mode, benchmark leaderboards, Fable 5, GPT 5.6, Opus 4.8, open model weights, self-host AI, enterprise AI solutions, long context AI, front-end design AI, subscription AI tools, API pricing, AI benchmark, arena rankingsSend Everyday AI and Jordan a text message. (We can't reply back unless you leave contact info) Ready for ROI on GenAI? Go to youreverydayai.com/partner 

We’re high as a kite on World Cup football. The final test, coming Sunday, is down to Spain and Argentina. My question is whether this frenzied global contest might actually be rescuing our species from ... The post The World’s Sport appeared first on Open Source with Christopher Lydon.

(0:00) Bestie intros: Brad Gerstner fills in for Friedberg! (2:58) OpenAI vs Anthropic IPOs: Why it matters who goes first, what they learned from the SpaceX IPO, the unlimited TAM of intelligence (27:39) The open source decision, Meta's new model, Zuck's price war, AI duopoly (54:29) CCP considering putting export controls on Chinese models, is open source ending in China? (1:03:09) Trump Accounts launch, getting young Americans bought back into capitalism Apply for Summit 2026: https://allin.com/events Follow Brad: ttps://x.com/altcap Follow the besties: https://x.com/chamath https://x.com/Jason https://x.com/DavidSacks https://x.com/friedberg Follow on X: https://x.com/theallinpod Follow on Instagram: https://www.instagram.com/theallinpod Follow on TikTok: https://www.tiktok.com/@theallinpod Follow on LinkedIn: https://www.linkedin.com/company/allinpod Intro Music Credit: https://rb.gy/tppkzl https://x.com/yung_spielburg Intro Video Credit: https://x.com/TheZachEffect Referenced in the show: https://polymarket.com/event/ipos-before-2027 https://x.com/thejessezhang/status/2074154325933424861 https://x.com/praveenTweets/status/2074605343439810922 https://x.com/nikesharora/status/2074802778074124434 https://x.com/nikesharora/status/2074814752174522857 https://x.com/brian_armstrong/status/2070670644577280109 https://x.com/andyfang/status/2074252174226493584 https://x.com/nikesharora/status/2074630732019036574 https://x.com/finkd/status/2075218444056707458 https://x.com/alighodsi/status/2074996561306955958 https://blog.nicolasmeridjen.com/en/blog/2026-04-03-alibaba-qwen-closed-source-end-of-open-weight https://www.chinatalk.media/p/chinas-ai-companies-are-going-closed https://www.reuters.com/world/beijing-is-looking-curbing-overseas-access-chinas-top-ai-models-sources-say-2026-07-07 https://x.com/KurtSupeCPA/thread/2074817292550984010

(0:00) The AI Buildout: Datacenters Bigger Than Cities (Andrew Feldman) (1:50) Reasoning, Inference, and Breaking Moore's Law (16:28) Open Source, AI Sovereignty, and the Road to AGI (40:54) The Innovation Behind Generative Video (Robin Rombach) (47:31) Martin Scorsese, Robots, and the Future of Hollywood IP Thanks to our partners for making this possible! AppLovin Ads - AppLovin's AI advertising platform reaches over a billion daily active users across mobile games. Full-screen video ads with a 35-second median watch time. Advertisers are profitably spending hundreds of thousands of dollars a day and advertiser access is still in closed beta. The window is open at https://applovin.com/ALLIN Nasdaq - Positioned at the nexus of technology and the capital markets, Nasdaq provides premier platforms and services for global capital markets and beyond with unmatched technology, insights and markets expertise. https://www.nasdaq.com/convergence-economy Follow Andrew: https://x.com/andrewdfeldman Follow Robin: https://x.com/robrombach Follow the besties: https://x.com/chamath https://x.com/Jason https://x.com/DavidSacks https://x.com/friedberg Follow on X: https://x.com/theallinpod Follow on Instagram: https://www.instagram.com/theallinpod Follow on TikTok: https://www.tiktok.com/@allin Follow on LinkedIn: https://www.linkedin.com/company/allinpod Intro Music Credit: https://rb.gy/tppkzl https://x.com/yung_spielburg