playnext.fm

The AI Daily Brief: Artificial Intelligence News and Analysis · Nathaniel Whittemore

Work AGI is the Only AGI that Matters

March 25, 2026·26 min·2 clips
Sam Altman told staff OpenAI has finished training Spud, a model that might accelerate the economy in weeks.
The AI Daily Brief covers a dense news cycle across OpenAI, SpaceX, and the ongoing AI policy battles. CNBC obtained documents resembling an OpenAI IPO prospectus, including standard legal risk disclosures acknowledging Microsoft as responsible for a substantial portion of financing and compute, as well as risks from Elon Musk litigation, capital expenditure, and geopolitical exposure to Taiwan. OpenAI clarified this was not an actual IPO prospectus, and Microsoft called the language standard, though it represents a preview of what future public disclosures will likely look like. Separately, SpaceX is finalizing paperwork for a $75 billion IPO, which would be the largest in US history, with retail investors potentially receiving double the usual 20% allocation and no standard six-month lockup for insiders. The Fundrise Innovation Fund, which holds pre-IPO shares in Anthropic, OpenAI, and SpaceX, briefly traded at 16 times its net asset value, implying an Anthropic valuation of $5 trillion against its February $380 billion raise, exposing how meme-stock dynamics can decouple ETF prices from underlying value. SoftBank is pushing past its self-imposed 25% loan-to-value ratio to fund another $30 billion OpenAI commitment, with CEO Yoshimitsu Goto acknowledging the potential for a temporary policy breach. On the OpenAI organizational front, Sam Altman steps down as Helion Energy chairman and narrows his scope to fundraising, compute supply chains, and data center buildout. The safety team moves under Chief Research Officer Mark Chen and the security team under President Greg Brockman. CEO of Applications Fiji Simo takes the commercial lead. Altman sent staff a memo noting that pre-training on the Spud model is complete and suggesting a very strong model is weeks away that could accelerate the economy. OpenAI announced Sora's shutdown, with the official Sora account thanking its community while noting compute constraints made the product unsustainable relative to demand. In the Pentagon versus Anthropic case, a federal judge in Northern California called the Pentagon's conduct troubling, noting that designating Anthropic as a supply chain risk appeared to punish the company for seeking public scrutiny, potentially violating the First Amendment. The Pentagon's lawyer argued in court that Secretary of War Pete Hegseth's tweet banning all military contractors from working with Anthropic was beyond the scope of his authority, a reversal that the judge found unconvincing.

As heard by us

A brisk, grounded case for judging AGI by the work it can change.

Work AGI sits at the center of this episode: the idea that the major labs are now trying to make advanced models useful inside companies, not just impressive in isolation.

Read the full review in PlayNext →

Why you'd press play

Press play when AGI arguments need a dose of workplace reality.

Read the full recommendation in PlayNext →
Listen to the show on
Apple PodcastsSpotify
Play clips from this episodeMore from The AI Daily Brief: Artificial Intelligence News and Analysis