The WordPress co-founder claims Automattic board members "conspired" behind his back to vote him out — then immediately posted that he's back "in control." The whiplash is extraordinary, and the company behind WordPress.com, Tumblr, and several other major platforms is in genuine chaos. This is the web infrastructure story nobody is watching closely enough.
The University of Michigan and Los Alamos National Labs want to plant a massive AI data center in a small Michigan township — without asking residents first. The backlash is a preview of hundreds of similar fights coming as the nuclear-AI power story lands in actual communities.
Willison gave GPT-6 Astra his home address and asked for 5K and 10K running routes using OSM data. It ran autonomously for 27 minutes and delivered exactly what was asked. A quiet, practical demo of what long-horizon agentic tasks actually look like in the wild.
Former Google and Anthropic researchers at Mirendil are building systems they call "self-accelerating AI" — models that contribute meaningfully to their own development. This is the recursive improvement thesis moving from thought experiment to funded startup.
AI safety jitters hit equities hard — SoftBank shed over 10% — but Bitcoin held its ground. The decoupling narrative is getting another data point, and the trigger is telling: it wasn't macro, it was model risk.
Dario Amodei published a full essay calling for the AI industry to slow down and announced Anthropic is unilaterally implementing a three-part pacing plan. The timing — right as GPT-6 Astra demos are going viral — is either principled or strategic. Probably both. The Dr. Eli David quip retweeted by Charles Payne captures the absurdity neatly: every major player says pace the frontier, and China is happy to agree.
Ethan Mollick pushed back hard on the AI safety discourse crowding out everything else: "Existential risk is obviously critical, but it is not the only AI thing that requires policy." He's flagging a real governance blind spot — while Washington debates extinction scenarios, the mundane harms and labor disruptions keep compounding with zero regulatory attention.
a16z's David George argued AI is making the tech power law more extreme — capital compounds advantage faster than ever, and the gap between winners and everyone else is widening structurally. A VC saying "money begets more money" isn't news, but the mechanism — AI capability as a self-reinforcing moat — is worth taking seriously.
The week's dominant story isn't any single model release. It's the collision between the AI industry's internal power struggles and the physical world's resistance to absorbing them. Mullenweg's board coup at Automattic, Michigan residents shutting down a town hall, SoftBank cratering on safety fears — these are all the same story: AI expansion hitting friction it didn't anticipate.