Automattic’s 33-Hour Coup, and can AI labs police themselves?

A week after one of Anthropic’s own researchers set off alarms with a bleak warning, CEO Dario Amodei is trying to turn panic into policy. His plan is to “pace the frontier” of AI development, with independent safety evaluators and some kind of coordination among labs in democratic countries. The pitch has found fans. It has also found a very familiar skeptic in Nvidia’s Jensen Huang, who isn’t buying the idea that the industry can sort this out cleanly on its own.

That tension is the real story here. Amodei is asking AI companies to agree not just on technical standards, but on what “slowing down” even means and who gets to enforce it. That’s a lot to ask from firms that compete on speed, model quality, and bragging rights. The moment you put safety in the hands of the same people shipping the products, the whole thing starts to look less like governance and more like a group project with a prize attached.

The broader conversation on TechCrunch’s Equity podcast moved quickly from AI policy to old-fashioned boardroom drama. WordPress parent Automattic was back in the spotlight after Matt Mullenweg’s 33-hour ouster, a sequence the hosts say could end up giving him even more control over the technology that powers much of the web. That’s a neat little reminder that power in tech doesn’t only show up in model weights and safety memos. Sometimes it shows up in who survives a board fight.

And the deals keep coming. The episode also touches May Mobility’s SPAC and its more than $300 million in potential funding, plus DoorDash’s $425 million investment in Wonder. The through line is pretty clear: whether the subject is AI, web software, or food delivery, the money is still chasing the same promise — automation, faster and cheaper, with a lot of bold language wrapped around the edges.