Starting to suspect that Anthropic's putative security and safety considerations are largely posturing and performative.

I think self-consistency is such a rare virtue nowadays that people can't conceive of an org that sticks to its mission

agree with them or not, all steps from Ant's leadership are very predictable after conditioning on their core values

the anthropic silent limits thing is not "dishonest", not "a terrible idea", not "evil", not "dangerous" etc etc. it is just very disingenuous (but also very on brand) for them to claim they do it for "safety concerns". that is all.

@soldni Yeah, I mean I think they believe they are somehow morally superior and virtuous and anything they do must be correct because they are the “good guys” and in that way they are consistent but sort of vacuously so

@soldni Respectfully man, what the fuck are you talking about? In what universe is it “harmless” to help to us military bomb children? How is it a positive outcome for humanity?

They are consistent with their actual goal: monopolize access and control over AI development

fully owning this is confusingly written post, I should have just posted "everything Ant has ever done leads to this"

@natolambert But the public is not directly going to be using any of these models (outside of individual developers). Eventually things are going to be abstracted out as products and that's all the public is going to see. My two cents.

the silently failing strategy is a bit weird. I can speculate it is to prevent jailbreak, but not doing same for other safety risks is one rare case of non-consistent strategy.

@soldni What are their core values? Is harmless not one of them??

@pjreddie that the definition of harmlessness, while based on public input and 3rd parties consultation, is ultimately up to them unless compelled by law

@natolambert They are already working on their next super model. It's called Fairy Tale

@tunguz

@soldni Also sorry I think I came in too hot with my initial reply I think we basically agree! This is not (or shouldn’t be) surprising behavior from anthropic

@soldni Yeah I agree with this, I guess I think they have very consistent positions and values they are just completely divorced from a lot of their stated values/goals

@tunguz Starting? It’s run by the guys who said ChatGPT 2 was dangerous for the world.

@natolambert tech bros on X are much dumber than genpop

@pjreddie ugh, I shouldn't have used "virtue" since it has positive connotations; I just wanted to point out they are consistent in their positions

@ScienceOrMyth @tunguz All the frontier labs are somewhat aware of what the others are doing. If you don't, you are just not part of it.

@natolambert I understand the cynicism by default, but I don't understand what they could do differently, that would distinguish between legitimate safety restrictions and so-called safety washing. Their actions are compatible with both

@soldni *the US military

Some day I’ll proof read these before smashing the tweet button