Great Agenda from @TiffanyLi158545 this morning on the cyber startups trying to get agents under control. In other news, a scooplet: new Sol, Astra and cyber models from OpenAI could be coming as soon as Friday...

New: OpenAI is all-hands-on-deck to get its next flagship model in a place where leaders feel like they can safely release it. At the same time, they're dealing with the fallout from 3 safety researchers' firings. w/ @aaronpholmes @amir:

OpenAI’s Sora could be coming back…ish. A stealthy startup cofounded by Hollywood mogul Jeffrey Katzenberg and ex Sora lead Bill Peebles is in talks with OpenAI to tweak and sell the model to film studios. w/@coryweinberg

Exclusive: Jensen Huang's dealmaking machine ran non-stop the past summer–buying Hugging Face, backstopping a giant OpenAI data center, investing in a range of AI startups and infrastructure firms. Huang sure kept his team (and reporters😉) very busy. But what's driving this historic spree is a leader who lives in constant anxiety about Nvidia's future. More unreported details about OpenRouter, Perplexity, Figure and what's on Nvidia's radar next w/ @_pheebini in @theinformation 🎯

A NY City Council hearing yesterday was more theater than anything else. But, amidst all the hubbub, whistleblower researchers had interesting things to say about how quickly AI research is being automated.

Proximal, a year-old startup that provides coding data to the labs, has raised at a $300m valuation from General Catalyst. In the last 10 months, it's surpassed $200m in annualized revenue, highlighting the labs' neverending appetite for data:

Inference providers are hitting the funding markets. Fal has spoken to investors about raising at $15B, while Fireworks has considered raising at $30b. Both would nearly double their valuations from rounds earlier this year. w/ @validapau @aaronpholmes

The scoops don't stop! TypeSafe, which is developing a cheaper and faster alternative to frontier AI called Jev, is raising $1B+ at a $10B+ valuation. This is just a week after it raised its last round at $200m. w/ @coryweinberg @validapau

We have a new acronym for you: SAFA. The Google, OpenAI and Anthropic AI safety body is starting to take shape, with a tentative name of the Standards Authority for Frontier AI and hopes to launch by the end of the year or early next. W/ @steph_palazzolo Lots of details here including: - Former White House AI adviser Sriram Krishnan was approached for CEO, and David Friedberg and Condoleezza Rice considered for chair - There's an active debate over whether the body should conduct testing itself, amid concerns about lack of resources for CAISI - After the effort for a FINRA-like body lost steam in the White House, SAFA would proceed independently at first, with optimism for a public-private partnership in the future - SAFA could work together with a previous industry consortium, the Frontier Model Forum - There's widespread industry skepticism about the project from outside the three AI labs You can read all about it here:

Had an amazing time at our AI Agenda Live conference in SF yesterday! Got to interview folks like @embirico, @amasad & @TYehoshua, watch @Jessicalessin in conversation with @koraykv, & listen in on so many more discussions between my colleagues and AI leaders. Until next year!

In this morning's AI Agenda, @amir reports on OpenAI's work to develop its answer to Grok Bot, as well as what it's doing to respond to Instinct. Plus, I share more on what OpenAI researchers have seen to make them so nervous about the progress of AI.

As OpenAI looks to respond to safety fears, one solution could lie in the recent past: earlier this year, OpenAI and Anthropic were negotiating a legally-binding deal to stress-test each other's models. w/ @amir:

Jensen, Musk, Zuck: They're on our new list of Tech's Most Influential People, which looks at who will sway the industry over the next 10 yrs. But we went deeper than the obv. choices: @dwarkesh_sp, @JTLonsdale, @YasminRazavi, @ScottWu46, @miramurati—you're in great company!

New in this morn's AI Agenda: OpenAI is close to solving another Millennium Prize problem, the Hodge Conjecture. It could take extra time to announce, though, as the company figures out how to navigate its rocky relationship with the math community.