🕶️ Anthropic buys the dev shop behind OpenAI

Anthropic just bought the company that builds its API docs, Musk lost a $135B lawsuit on a calendar technicality, and somewhere a team is now 35% AI-powered. The gap between 'we shipped an agent' and 'the agent actually works in production' is where the real platform wars are being fought, and this week, boring plumbing beat bold narrative every time.

Email gets overwhelming fast: too much time digging through threads, chasing replies, and scheduling back-and-forth. Superhuman Mail, part of the Superhuman productivity suite, is built for long-term productivity, not quick fixes.

Its AI cuts through the repetitive stuff: surfacing key details from old threads, reminding you to follow up (and drafting it), and replacing scheduling chains with a simple booking link.

Net result: less inbox chaos, more actual work getting done.

In the agent era, the boring plumbing is turning into the real platform war. Anthropic just bought its plumber, acquiring Stainless, the devtools startup quietly powering official SDKs not only for Claude but for players like OpenAI and Google.

Folding that capability in-house tightens Anthropic’s grip on how agents reach real systems. Stainless already turns a single API spec into native-feeling SDKs, CLIs, and MCP servers across major languages, so this move makes Claude’s “connect to anything” story less slideware and more default behavior.

The deeper play is control of the connectivity layer: whoever owns the SDKs, CLIs, and servers owns the developer experience. As the team leans in on MCP and agent tooling from inside Anthropic, switching costs around Claude’s ecosystem just went up.

The Musk–Altman courtroom saga ended on a technicality: a California jury unanimously found his claims time-barred. In siding with OpenAI’s statute-of-limitations defense, the jury rejects late filing and turns allegations of “stealing a charity” into a procedural win.

The verdict means any court-ordered overhaul of OpenAI’s structure is off the table as a rumored IPO approaches, and the restructuring threat disappears for Microsoft’s biggest AI partner. Jurors focused narrowly on when promises were made and broken, not on the broader ethics of converting a nonprofit lab into a capped-profit vehicle.

For founder disputes, the case underlines how timing and paperwork can outweigh narrative: governance pivots, side letters, and charitable commitments all live or die on dated evidence. Musk has already said he will appeal—Musk signals quick appeal—but the first-round lesson is blunt: existential grievances filed too late are indistinguishable from none at all.