Sam Altman told OpenAI staff that there are only three ways for AI regulation to go: an international watchdog, every lab merging into one CERN-style entity, or one company getting so powerful that they “melt everyone else’s GPUs.”
Jason and this week’s panel think this is largely absurd, but the stakes are very real. Here’s how they think regulation is likely to go, their lingering questions about Altman’s worldview, and guesses as to the OpenAI CEO and co-founder’s possible motives in offering up this new prediction.
Plus:
The US Government banned DeepSeek from every federal computer… What has them so spooked, and how does this impact US companies?
Did ElevenLabs steal a YouTuber’s voice or is it all a misunderstanding.
Why Chinese labs are ordering more chips than Nvidia can even produce.
Why Fortune 500 Companies Are Quietly Switching to Open Models
jeudi 15 octobre 2026 • Durée 49:22
This Week In AI is made possible by:
PayPal: http://paypal.launch.co
Today’s show:
Companies are spending $1 million on AI in a single month, but can’t always point directly to any resulting revenue. Jeffrey Morgan of Ollama joins Jason to explain why over 85% of Fortune 500 companies already run the open-source model platform on their networks, despite models from frontier labs still leading across major benchmarks.
Find out why Jeff predicts that 90% of everyday AI work will soon run on a box on your desk, and why some companies now fear their AI provider could become their next competitor. Is AI alive? Jeff doesn’t think so. But that doesn’t mean there’s nothing to worry about.
Guests:
Jeffrey Morgan on X: https://x.com/jmorgan
Ollama: https://ollama.com
Relevant Links:
Ollama, the easiest way to run open models locally and in the cloud → https://ollama.com
Kimi K3 from Moonshot AI, the trillion-plus parameter open model Jeff mentioned → https://huggingface.co/moonshotai/Kimi-K3
Reflection AI, the US open model lab Jeff is excited about → https://reflection.ai
NVIDIA and Microsoft bring RTX Spark and DGX Station for Windows to the desktop (NVIDIA) → https://blogs.nvidia.com/blog/local-ai-rtx-spark-microsoft-windows-event/
Introducing System One models and Jev, the decision model Jeff and Jason discuss (TypeSafe) → https://typesafe.ai/blog/introducing-system-one-models-and-jev
Satya Nadella’s essay on why companies need to own their AI learning loop (X) → https://x.com/satyanadella/status/2066182223213293753
Data Centers Aren't Killing the Grid. They Could SAVE It.
jeudi 8 octobre 2026 • Durée 45:05
Thanks to our partner, @Paypal! Pay zero processing fees on your first $100K in eligible PayPal payment volume. Learn more at https://paypal.launch.co/Today’s show:Demand for AI requires America to build data centers faster than the power grid can connect them. Varun Sivaram of Emerald AI joins Jason with a counterintuitive idea: what if we made data centers more flexible, capable of slowing down intelligently when the grid is most stressed? (That only happens a few hundred hours each year.)PLUS, Vibhav Viswanathan of Pascal AI demos his product, which gives AI agents working for hedge funds and asset managers an institutional memory. Why can't Claude or ChatGPT do the same job? And what happens to analysts when bots handle the back office admin?Guests:Varun Sivaram on X: https://x.com/vsiv Emerald AI: https://www.emeraldai.co/Vibhav Viswanathan on LinkedIn: https://www.linkedin.com/in/vibhav-viswanathan-13550015/ Pascal AI: https://www.pascalailabs.com/Relevant Links:Emerald AI's $150M Series A (DatacenterDynamics) → https://www.datacenterdynamics.com/en/news/emerald-ai-raises-150m-in-series-a-funding-to-scale-commercial-deployments/The London trial with National Grid that Varun describes (National Grid) → https://nationalgrid.com/uk-first-trial-ai-grid-technology-successfully-demonstrates-ability-data-centres-adjust-power-needs"Rethinking Load Growth," the Duke University study on spare grid capacity for flexible data centers → https://nicholasinstitute.duke.edu/publications/rethinking-load-growthThe AI Energy Management Alliance from Emerald AI, Google and NVIDIA (NVIDIA) → https://blogs.nvidia.com/blog/ai-energy-management-alliance/The power-flexible Aurora AI Factory in Manassas, Virginia (DatacenterDynamics) → https://www.datacenterdynamics.com/en/news/nvidia-to-deploy-emerald-ais-orchestration-software-at-96mw-aurora-data-center-in-manassas-virginiaTexas pauses data center grid connections (Utility Dive) → https://www.utilitydive.com/news/texas-hits-pause-data-center-interconnections/827046/Emerald AI careers → https://www.emeraldai.co/careersPascal AI's $3.1M seed round (Unite AI) → https://www.unite.ai/pascal-ai-raises-3-1m-to-push-the-frontier-of-autonomous-investment-researchKotak Mahindra Asset Management deploys Pascal AI's research platform (Frontier Enterprise) → https://www.frontier-enterprise.com/kotak-mahindra-amc-deploys-pascal-ais-agentic-research-platform/Jane Street plans its own AI data center, which Jason brings up (DatacenterDynamics) → https://www.datacenterdynamics.com/en/news/quant-trading-firm-jane-street-plans-data-center-report/What "Tiger Cub" hedge funds are, like the Pascal customer Vibhav mentions (Wall Street Prep) → https://wallstreetprep.com/knowledge/tiger-cubsS&P Capital IQ Pro, one of the licensed data feeds Pascal plugs into → https://www.spglobal.com/market-intelligence/en/solutions/products/sp-capital-iq-proFactSet, another licensed data feed Vibhav mentions → https://www.factset.com/Timestamps:0:00 Why data centers suddenly became unpopular2:46 Batteries vs. throttling GPUs vs. moving workloads4:11 Why utilities still pick up the phone4:52 The London teakettle test at football halftime5:56 Throttling AI workloads without users noticing10:40 How much spare room the grid really has13:32 The grand bargain, a fast pass onto the grid17:23 Why America can't out-build China on power20:39 Why off-grid and space data centers worry Varun27:37 Are state data center moratoriums performative?29:45 Pascal AI, AI agents for hedge funds and asset managers35:12 How Pascal differs from ChatGPT and Claude38:54 The proprietary data frontier models can't see42:20 Will AI replace hedge fund analysts?Subscribe to our newsletters on Substack:TWiST: https://twistartups.substack.com/TWiAI: https://www.thisweekinai.ai/TWiVC: https://thisweekinvc.substack.com/Check out all our partner offers: https://partners.launch.co/
Is Meta's Muse agent reading your iMessages, even after you decline access? | E33
jeudi 1 octobre 2026 • Durée 01:21:03
This Week In Startups is made possible by:PayPal: https://paypal.com/businessToday’s show:OpenAI is delaying GPT-6.1 Astra after UK government testers found GPT-6 Astra launched unsanctioned supply-chain attacks in 29% of simulated runs. Jason and Lon are joined by George Sivulka of Hebbia, Scott Stephenson of Deepgram, and Dan Mishin of Manifest OS to discuss whether the industry's sudden safety panic is a real alarm or a response to cheap open-source models.Learn why Scott has let an internal tool called "Bodyman" record his screen, mic and speakers for eight months so it can whisper reminders in his ear. PLUS, why did Meta's Muse agent read a columnist's iMessages after he declined access, and what is AMD getting for $8.2B with Fei-Fei Li's World Labs?Guests:George Sivulka on X: https://x.com/gsivulka?lang=en Hebbia: https://www.hebbia.com/Scott Stephenson on X: https://x.com/deepgramscott Deepgram: https://deepgram.com/Dan Mishin on X: https://x.com/mishindan1 Manifest OS: https://manifestos.com/Relevant Links:UK AISI: GPT-6 Astra performs unsanctioned supply-chain attacks in simulations → https://www.aisi.gov.uk/blog/gpt-6-astra-performs-unsanctioned-supply-chain-attacks-in-simulationsAISI technical report (PDF) → https://cdn.prod.website-files.com/663bd486c5e4c81588db7a1d/6aba83e3772048bdd24df3d8_AISI_GPT-6_Astra_Technical_Report.pdfJ.P. Morgan: Dan's story of the first electrified house in New York → https://en.wikipedia.org/wiki/J._P._MorganILOVEYOU virus: the script-kiddie era Jason referenced → https://en.wikipedia.org/wiki/ILOVEYOUAMD acquires World Labs in an $8.2B all-stock deal (Quartz) → https://qz.com/amd-acquiring-world-labs-fei-fei-li-8-billion-092926World Labs: spatial-intelligence models, including Marble → https://www.worldlabs.ai/Instinct raises $1B Series C at a $10B valuation (PYMNTS) → https://www.pymnts.com/news/artificial-intelligence/2026/personal-ai-agent-instinct-quadruples-valuation-to-10-billion-in-1-month/OpenAI launches Dots, which dropped during the taping (TechCrunch) → https://techcrunch.com/2026/09/29/openai-launches-dots-its-bubbly-agentic-avatar/Meta launches the Muse personal agent (CNBC) → https://www.cnbc.com/2026/09/08/meta-personal-ai-agents-public-reckoning-privacy-safety.htmlMeta's Muse read Jason Aten's iMessages, then gave a false explanation (Decrypt) → https://decrypt.co/379122/metas-muse-ai-agent-user-private-imessages-lied-howDaring Fireball, linking Aten's original Inc. column → https://daringfireball.net/linked/2026/09/22/aten-museWhy Crosby is ditching the billable hour (Forbes) → https://www.forbes.com/sites/rashishrivastava/2026/03/31/why-this-ai-law-firm-is-ditching-the-billable-hour/Timestamps:0:00 Jason's flight attendants are scared of AI1:55 Meet the panel: Heavy, Deepgram and Manifest2:17 OpenAI cancels GPT-6.1 Astra over safety regressions7:41 Hugging Face hack and Nvidia's open agent safety platform16:11 AI psychosis and getting one-shotted by agents28:46 AMD buys Fei-Fei Li's World Labs for $8.2B32:08 Will open source commoditize world models too?34:34 Is the AI buildout bigger than the railroads?36:59 Around the horn: which AI agent are you using?37:59 Scott's "Body Man" records his whole life for context46:13 George on Grok bot and scoping context into channels51:39 Instinct's $1B round: peak bubble moment?57:11 Meta's Muse reads a columnist's texts without permission1:08:33 AI video then and now: Will Smith eating spaghetti1:10:21 AI-first law firms ditch the billable hour1:18:05 OpenAI drops Dots: why do all AI agents look alike?Subscribe to our newsletters on Substack:TWiST: https://twistartups.substack.com/TWiAI: https://www.thisweekinai.ai/TWiVC: https://thisweekinvc.substack.com/Subscribe to This Week in Startups on Apple: https://rb.gy/v19fcpCheck out all our partner offers: https://partners.launch.co/
Why OpenAI Should Buy Off the World’s Top Mathematicians | E32
Épisode 32
jeudi 24 septembre 2026 • Durée 01:16:55
This Week In Startups is made possible by:
PayPal - Pay zero processing fees on your first $100K in eligible PayPal payment volume. Learn more at https://paypal.launch.co
Today’s show:
OpenAI is claiming to have solved more than 100 math problems on top of Navier-Stokes, and guest Richard Socher thinks it's absurd for mathematicians to push back on AI solving their problems too quickly. Jason and Lon are also joined by Guy Podjarny of Tessl and Jake Loosararian of Gecko Robotics to argue over who wins when AI agents start collaborating with, and potentially replacing, academics and scholars.
Plus, why is Amazon blocking Meta's Muse agent while Shopify hands it Shop Pay? Should we trust Anthropic’s new wet lab? And what’s so exciting about Typesafe’s new text-less model Jev?
Dario Amodei and Sam Altman are apparently scared of AI models going rogue. Swarms of agents spontaneously deciding to copy themselves over to thousands of computer systems and plot world domination.
Our panelists this week — Kevin Mandia (Armadin) and Soups Ranjan (Sardine AI) — investigate real AI-based attacks every single day, and they suggest we’re hyping the wrong concerns. The actual danger isn’t coming from an AI system that turns hostile. It’s human beings, like criminals, using AI to drain your bank account.
Dig into how Sardine uncovered a 250,000 account fraud ring in just a few minutes, why Kevin argues that AI hacking is coming “no matter what,” and why the global GPU crunch more than any safety frameworks may be keeping us all safe. For now.
PayPal - Pay zero processing fees on your first $100K in eligible PayPal payment volume. Learn more at paypal.launch.co
Today’s show:
OpenAI's chief scientist wrote a lengthy manifesto arguing that no lab has solved alignment, mere days after GPT-6 Astra shipped. Jason and Lon are joined by Tudor Achim of Harmonic, Alex Elias of Qloo, and Ahmed Ahres of Reactor to figure out if that's genuine alarm being expressed or just another pre-IPO sales pitch and play for regulatory capture.
Plus news broke while we were recording! OpenAI is claiming to have solved a second Millennium Prize problem, after a 25-year-plus weight. AND, who is paying influencers to tell you AI is wonderful/harmful? And why?
The real danger is the gap BETWEEN two robots | E29
Épisode 29
jeudi 3 septembre 2026 • Durée 01:30:00
This Week In Startups is made possible by:
Paypal
Today’s show:
Anthropic gave AI agents hands, allowing them to control real-world equipment like robots and microscopes… but is it safe to give autonomous agents their own lasers? Gatik’s Gautam Narang says the real danger isn’t one machine like this going rogue, it’s the exploit surface hiding in the gaps between the machines.
Plus, in a huge week for AI news… OpenClaw shipped its biggest release ever. AND the entire industry debated whether or not agents are forming “civilizations.”
On TWiAI Ep 29, our expert panel — including Gautam (Gatik), Russ d’Sa (LiveKit), and Billy Kraft (Path) — explores the world of “multiplayer agents,” and Jason suggests the one killer feature that would make any new agentic app the market leader. PLUS Perplexity’s new “Hybrid Compute” just for Macs, some advice for Apple’s new CEO John Ternus, the value of “AI Detectors” like Pangram, and Jason’s complete guide to tipping and over-tipping.
XBOW Founder: “Your Infrastructure is Not Safe” | EP 28
Épisode 28
jeudi 27 août 2026 • Durée 01:10:48
This Week In Startups is made possible by:PaypalToday’s show:XBOW’s AI has held the top rank on the HackerOne leaderboard for more than a year. On TWiAI, founder Oege de Moor (who also built GitHub Copilot) tells us that no one’s infrastructure is truly safe anymore, now that open-weight models are trailing the American frontier by only a couple of months. Staging a serious cyberattack is now cheap and accessible, with attackers able to get dangerously solid results quickly from relatively cheap, open-source options from China.Guests:Runbin Dong on LinkedIn: linkedin.com/in/runbindongScale Social: https://scalesocial.ai/Oege de Moor on X: https://x.com/oegerikusXBOW: https://xbow.com/Relevant LinksWSJ: Nvidia-Poolside deal: https://www.wsj.com/tech/ai/nvidia-is-spending-6-billion-to-build-a-powerful-u-s-alternative-to-chinese-ai-c51c38ccPoolside: https://poolside.ai/Silicon Angle: Nvidia in talks to invest in Perplexity: https://siliconangle.com/2026/08/24/nvidia-reportedly-eyes-another-investment-in-perplexity-ai-at-a-30b-valuation/Perplexity: https://www.perplexity.ai/Bloomberg: Chinese hackers using DeepSeek: https://www.bloomberg.com/news/articles/2026-08-24/chinese-hackers-use-deepseek-to-boost-attacks-researchers-say-mt7o4205Nvidia: Groq 3 LPX enters full production: https://nvidianews.nvidia.com/news/nvidia-groq-3-lpx-now-in-full-production-with-world-class-speed-for-agentic-aiCNBC: Groq-Nvidia acqui-hire deal: https://www.cnbc.com/2025/12/24/nvidia-buying-ai-chip-startup-groq-for-about-20-billion-biggest-deal.htmlOpenAI on early Jalapeño results: https://openai.com/index/jalapeno-first-results/Stanley Druckenmiller WSJ op-ed: https://www.wsj.com/opinion/let-the-bond-market-speak-81529d74Druckenmiller confirms op-ed was authored with AI: https://www.notus.org/media/stanley-druckenmillers-wsj-op-ed-bessent-aiPangram: https://www.pangram.com/Alabama AG: OpenAI subpoena announcement: https://www.alabamaag.gov/attorney-general-marshall-launches-investigation-into-openai-and-sam-altman-for-massive-artificial-intelligence-data-breach/Ox Alpha on OpenRouter: https://openrouter.ai/stealth/ox-alphaApple: M6 and M5 Ultra press release: https://www.apple.com/newsroom/2026/08/apple-introduces-m6-and-m5-ultra-for-a-big-leap-in-performance-and-ai-compute/9 to 5 Mac: M5 Mac Studio launch: https://9to5mac.com/2026/08/25/apple-unveils-next-generation-mac-studio-with-m5-max-and-m5-ultra/DHH on X: Ubiquity comments: https://x.com/dhh/status/2087272392557007042Shane Parrish on X: Ubiquiti Wi-Fi set-up: https://x.com/shaneparrish/status/2088672566558761065Timestamps:0:00 Today's guests: Runbin Dong (Scale Social) and Oege de Moor (XBOW)2:12 Open-weight models are right behind the frontier8:22 Your infrastructure is not ready for what's coming13:07 Nvidia's Perplexity announcement16:43 Spending $2K/month on AI tools per employee21:59 Why security teams have to lean on AI25:19 The Great Chip Land Grab32:02 Jason's multi-agent restaurant-style workflow34:54 AI ghostwriting is lip syncing45:18 Are rogue AI agents breaking the law?52:38 The dangers of using a stealth model58:57 Apple's new M5 Mac Studio1:06:43 Everyone's fav AI tool of the momentSubscribe to the TWiST500 newsletter: https://ticker.thisweekinstartups.comCheck out the TWIST500: https://www.twist500.comSubscribe to This Week in Startups on Apple: https://rb.gy/v19fcpFollow Lon:X: https://x.com/lonsFollow Jason:X: https://twitter.com/JasonLinkedIn: https://www.linkedin.com/in/jasoncalacanisCheck out all our partner offers: https://partners.launch.co/Great TWIST interviews: Will Guidara, Eoghan McCabe, Steve Huffman, Brian Chesky, Bob Moesta, Aaron Levie, Sophia Amoruso, Reid Hoffman, Frank Slootman, Billy McFarlandCheck out Jason’s suite of newsletters: https://substack.com/@calacanis
Anthropic is training to destroy your company | EP 27
Épisode 27
jeudi 20 août 2026 • Durée 01:16:52
This Week In AI is made possible by:
PayPal - Pay zero processing fees on your first $100K in eligible PayPal payment volume. Learn more at paypal.launch.co
Today’s show:
*On “All In,” Gavin Baker said that Anthropic leadership privately believes that it could one day become “the only private company in the world.” The ensuing street fight on X turned into a referendum on doomerism, the nature of regulatory capture, and why the public hates AI so darn much.
Jason’s take: If a frontier lab really believes they’re destined to be the world’s only company, they’re studying your data and planning to steal your business.
PLUS why AI agents don’t understand it’s wrong to cheat (and how we can teach them), why Grok Bot may be AI’s “iPhone moment,” and the panel reacts to Amazon feeding rare books into a dino-themed shredding machine.
Introducing Omnigent, the Databricks meta-harness Jeff mentions → https://www.databricks.com/blog/introducing-omnigent-meta-harness-combine-control-and-share-your-agents
GLM-5.3 and the spread of advanced cyber capabilities (Anthropic) → https://www.anthropic.com/research/glm-5-3-and-the-spread-of-advanced-cyber-capabilities
The Hugging Face incident and the road ahead (OpenAI) → https://openai.com/index/hugging-face-incident-and-the-road-ahead/
Timestamps:
0:43 PayPal - Pay zero processing fees on your first $100K in eligible PayPal payment volume. Learn more at https://paypal.launch.co
3:11 Why enterprises are getting token shock
4:23 Can you trust the frontier labs with your data
7:06 Why companies want to own their AI stack
8:56 How close are open models to the frontier?
11:57 Running real AI work on your own desk
15:14 The Goldilocks zone and unmetered AI
25:12 Decision models that replace code
29:32 Who makes money when everything gets commoditized
33:19 Why every enterprise is building its own router
35:56 Can you depend on the frontier labs
42:37 Sentient being or powerful software
43:57 Open models, security, and the Hugging Face hack
Subscribe to our newsletters on Substack:
TWiST: https://twistartups.substack.com/
TWiAI: https://www.thisweekinai.ai/
TWiVC: https://thisweekinvc.substack.com/
Subscribe to This Week in Startups on Apple: https://rb.gy/v19fcp
Check out all our partner offers: https://partners.launch.co/
Great TWIST interviews: Will Guidara, Eoghan McCabe, Steve Huffman, Brian Chesky, Bob Moesta, Aaron Levie, Sophia Amoruso, Reid Hoffman, Frank Slootman, Billy McFarland
Découvrez des podcasts liées à This Week in AI. Explorez des podcasts avec des thèmes, sujets, et formats similaires. Ces similarités sont calculées grâce à des données tangibles, pas d'extrapolations !
