Anthropic to investors: Trump hates us so much - Axios

Anthropic has filed its S-1 and, with a candour one rarely encounters outside a hostage video, has dedicated a risk factor to the possibility that the US government simply does not care for it. Axios reports the company warning investors that its public positions on AI regulation and safety could subject it to "heightened scrutiny, political backlash, or disproportionate regulatory action." Disproportionate, mind. Not unlawful, not unwarranted — disproportionate, which is the precise word you reach for when you know the treatment will be perfectly legal and simply unfair. And in the very next breath the filing confirms Anthropic has no plans to change its behaviour. Marvellous. A prospectus that says: we shall keep doing the annoying thing, please wire funds.

The real jewel is the comparison clause. Anthropic warns it could become a target of government investigations that "similarly-situated competitors" may not face. An unnamed, unindicted, deeply fortunate category of competitor — one that has, hypothetically, spent the last several years being noticeably quieter about the whole safety business. In an S-1 there are no accidents. That sentence is a competitive-disadvantage disclosure dressed up as a legal caution, and every investor who reaches page eighty now knows exactly which rival is standing in the sunlight while Anthropic gets the subpoena.

So the pitch, ultimately, is this: give us money for a company whose stated growth strategy includes being loathed in Washington. The headline has Trump in it because Axios knows which way the wind is blowing, and Anthropic is, to its credit, refusing to pretend otherwise — no frantic LinkedIn post about "constructive engagement," no quiet 2am deletion of the safety blog. It is a strange kind of integrity, the sort that shows up in a securities filing as a warning label rather than a value. One imagines the underwriters reading it twice, sighing, and ordering a considerably larger lunch.