Industry contemporary
Anthropic
Founded by people who left OpenAI over where it was going, and built around the bet that safety research and frontier capability have to happen in the same building.
Anthropic was incorporated in January 2021 by senior people who had just left OpenAI, principally the siblings Dario Amodei - OpenAI's VP of Research - and Daniela Amodei, its VP of Safety and Policy. The stated reason for leaving was directional disagreement, specifically over OpenAI's 2019 arrangements with Microsoft.
The other founders named in the record are Jared Kaplan, Jack Clark, Chris Olah, Ben Mann, Sam McCandlish and Tom Brown. Sources describe the group as seven former OpenAI employees while listing eight names including the Amodeis; the discrepancy is in the sources and is left visible rather than resolved by picking one.
The premise is a specific technical claim rather than a slogan: that you cannot do useful safety research on systems you are not building, because the problems only appear at the frontier. Dario Amodei had co-authored 'Concrete Problems in AI Safety' in 2016, which set out failure modes - side effects, unsafe exploration - that were theoretical then and are engineering concerns now.
The funding history includes a fact the company would presumably rather not carry: of the $580M Series B in April 2022, roughly $500M came from Alameda Research, the trading firm affiliated with Sam Bankman-Fried, months before its collapse. That is in the public record and belongs in any honest account of how the company was capitalised.
It is structured as a public benefit corporation with a governance trust, and its products are the Claude models. It is included on this timeline because the generative-AI era is not a single company's story, and because a site that teaches people to check their sources should name the one it was built with.