🚀 AI Deep Dive Episode 6: Are We Building World Models The Wrong Way? @rocketalignment and World Mechanics co-founder @soniajoseph_ explore how models represent physics, whether interpretability could improve training and why language alone may not be enough for physical AI. 00:00 Intro 01:39 What counts as a world model? 07:12 Looking inside models for physical understanding 11:28 Interpretable training and the bitter lesson 19:24 Physics, language and understandable explanations 28:07 Surgical robots and manufacturing 31:10 Language models and recursive self-improvement 38:13 JEPA and specialized world models 45:25 What video models actually learn about physics

Sam Altman says the company needs to figure out when AI becomes too powerful to safely develop or release. Internally dubbed “pacing 2.0,” the effort would establish limits on AI systems that can continuously improve themselves. Read more: thein.fo/3WCUYPR

Nvidia has committed more than $170 billion in investments, guarantees and other financial support to help AI cloud providers and data center developers expand. As investors grow wary of these companies’ debt and growth plans, Nvidia could face pressure to provide even more support. Read more: thein.fo/4hWbkLU

OpenAI and Anthropic face growing calls to slow AI development by restricting the computing power available to researchers. The proposal could buy time for better safeguards, but it would challenge a key competitive advantage in the race to build more powerful models. Full story: thein.fo/4jEmS7T

Exclusive: Private equity firms weighed a $100 billion-plus takeover of Microsoft's Xbox division, which has struggled with falling revenue and thin profit margins as Microsoft pours capital into AI. Read more: thein.fo/4i70cKR

What worries you most about Anthropic’s prospectus? 🗳️ An $80.7 billion net loss 🗳️ Two customers at about 20% of revenue 🗳️ Compute spending trajectory 🗳️ Whether growth can keep compounding Cast your vote in The Information Polls app, and see what the rest of the tech industry thinks. Download now to vote: thein.fo/4eKOhkZ

AI agents aren't just changing how companies work. They're changing how companies are built. Leaders from @blocks, @Cloudflare and @Klaviyo joined The Information's Co-Executive Editor @amir Efrati at AI Agenda Live 2026 in San Francisco to discuss how agentic AI is reshaping company structures, accelerating development and redefining the role of employees. Read more: thein.fo/4BBFmfR

Sam Altman told OpenAI staff that one approach to AI safety could involve merging the biggest AI labs into a single organization. But with OpenAI and Anthropic racing toward IPOs, coordinating with rivals could prove difficult. Full story: thein.fo/47r0RlB

Today on TITV: Elon Musk says Tesla and SpaceX have formed committees to explore a potential merger. Would combining the companies make financial sense? Plus, Netflix's slowest revenue growth in three years, a $48 million bet on AI marketing startup Different and prediction markets ahead of the midterms.

The Information | TITV | October 21, 2026 nitter.siddz.com/i/broadcasts/1qxvveNWv…

Today on TITV: Elon Musk says Tesla and SpaceX have formed committees to explore a potential merger. Would combining the companies make financial sense? Plus, Netflix's slowest revenue growth in three years, a $48 million bet on AI marketing startup Different and prediction markets ahead of the midterms. 📺 Tune in at 10 am PT / 1 pm ET on: thein.fo/3VRWSfn

SCOOP w/ @aaronpholmes: A group of PE firms in recent months explored a deal for Xbox that could have valued it up to $150 billion.

Nvidia has already put more than $170 billion on the line to support AI cloud providers and other major customers. Now it's considering another major investment in Firmus, which is seeking at least $2 billion in private funding after its IPO fell apart. Full story: thein.fo/3UzBXxf

Hackers used AI models from Anthropic, SpaceX and Moonshot AI to automate cyberattacks on more than 400 financial targets in Latin America. Exclusive research shows how the attackers bypassed AI safeguards to steal sensitive data with little human involvement. Read more: thein.fo/4z5PZ8S

CoreWeave shares have fallen roughly 20% in a week as investors question its debt-heavy approach to financing AI data centers. With borrowing costs climbing, the company’s troubles are raising concerns about how the broader AI boom will be financed. Read more: thein.fo/4zGZ6wu

Nvidia is in talks to invest many $100s of millions in Australian neocloud Firmus, which abruptly pulled its IPO early this month. More details in @theinformation: theinformation.com/articles/…

Exclusive: DeepSeek is helping Huawei design its next generation of AI chips, with Huawei engineers working inside DeepSeek’s offices. The collaboration could help China train advanced AI models without relying on Nvidia. Read more: thein.fo/4jqD8cr

OpenAI is preparing a cheaper model in its mid-tier Terra family as it looks to undercut Anthropic on price while offering developers more choices. Full story: thein.fo/4irv4XQ

Hundreds of AI cloud providers are racing to build data centers, but executives expect only seven to 10 to survive. Nvidia is helping keep the market afloat through investments, credit support and customer referrals. Read more: thein.fo/4dPplsi

Exclusive: Emil Michael said he would not allow Anthropic into the Defense Department as long as its software gives AI models control over whether to end conversations with users. Full story: thein.fo/4hVruFo Watch the interview: thein.fo/4joPTnQ

Could superintelligence become the ultimate military deterrent? @DeptofWar CTO @USWREMichael argues that AI could give the U.S. an overwhelming intelligence and technological advantage, potentially preventing conflicts before they begin: “Imagine if you added, super intelligence to that intelligence agency, the things they can understand about what the adversary is doing to protect you from that adversary making a move.” Read more: thein.fo/4i8FdHA