Competitive seething aside, OpenAI bros have good reasons to hate this. The real crime of Anthropic is, I think, that they've destroyed the commons. If a lab develops a much stronger model, what is the incentive to make it available to competition *that sabotages your work?*

@teortaxesTex It also nukes all safety collaboration or joint pause narratives.

"why tf would I willingly collab or accept a pause vote from a lab that wants to kidney shiv me?"

A coalition on "fuck CN/RU/EU", though unprincipled, was workable; now it's plainly "all labs for themselves gl".

@teortaxesTex I find it extremely ironic that Anthropic is doing this all in the name of “safety” while Dario simultaneously pushes out PR taking the position that Trump, the guy who bombs elementary schools, should decide how and where fully autonomous weapons are used.

@RoKahina @teortaxesTex Trump represents the will of the people , So when Dario says Trump should decide, he's actually saying the people should decide

@teortaxesTex OpenAI has also destroyed the commons. with the exception of edge cases like Ai2 all the labs have

@teortaxesTex But anthropic doing any less would be hypocritical of them

They’re walking the talk

@konstmish Yes indeed. Perhaps their strategy is making more money by selling their models to corporations rather than individuals. While the Fable 5 is very good, it runs into usage limit very quickly for it to be useful to researchers in a meaningful way.

@teortaxesTex

@AuroraSkye21259 @teortaxesTex lol

@FleischmanMena @teortaxesTex What has ru to do here, lol?