alt Hacker News...is Anthropic "meaningfully better" though? They're still fine being a defense contractor, and they lack the tools to enforce the ethics they want to uphold. They seemingly contribute even less to FOSS than OpenAI does (low bar) and split hairs over IP ownership when open models distill their results. Am I supposed to root for them because of their manufactured internet drama?
It's very reminiscent of the half-assed security theater that Google and Apple fought over. Neither one of them resisted government coercion in the end, they just took different routes to end up as federal asskissers.
I don't advocate for using either OpenAI or Anthropic. But you can say one is better while both are still unacceptable.
I don't see a viable case for arguing that they're equally bad, though you might quibble over the difference. (I couldn't care less how much these companies contribute to FOSS--that's small potatoes compared to everything else that's going on.)
My point is there are a lot of people invested in OpenAI's success who try to steer opinion around here, and they use arguments about ethics (or raise an eyebrow) for those purposes, not because they actually care about the ethics.
There was a rush of people flooding this thread to paint this post as new evidence against Amodei, when he's been upfront and consistent on these matters for a while. That doesn't mean anyone has to agree with him. But there's nothing new about his stance here. People suggesting that there is tells you what their motives are.