The interpretation of the law is classified? That’s stupid and everyone who protected that classification, regardless of whatever the interpretation is, is a traitor!
Probably the actual classified artifact is an NSA policy document that details the NSA's own interpretation of the law and thus forms part of its governance.
No, it’s a secret FISA court decision that the public can’t see but a member of the Senate Intelligence Committee can, but he can’t tell us what he’s seen. But he can ask questions to get the surveillance state to pull another Clapper-esque whopper and get away with it.
This is why I'm never giving a penny to OpenAI again, now matter how much damage control Altman tries to do with "look, we reworded the contract to have redlines too!". Yeah, legal redlines that the administration can bypass with their secret memos and secret rubberstamp courts. This isn't even a Trump thing: the Bush DOJ wrote secret memos making torture legal, the Obama DOJ wrote secret memos making it legal to assassinate American citizens. Non-technical redlines which aren't under the vendor's control aren't worth a piss squirt.
> This is why I'm never giving a penny to OpenAI again, now matter how much damage control Altman tries...
Altman is like Musk: he showed his true colors long before the current politically-inflected drama.
Musk was over-promising about self-driving, so much and for so long it became pretty clear he was a shameless liar. There are also so many reports of Altman lying (e.g. that's apparently why he got fired) and engaging in Machiavellian manipulations that you can be pretty sure he's a shameless liar too.
Contra the popular memes, I don’t think they’re losing money with every query sent (the money pit is capex on new models and hardware, but I don’t think inference itself is unprofitable), so this wouldn’t actually work.
I was already paying for Claude Max before the War Department fiasco, so there’s not much more I can do to hurt OAI apart from complain about it online, although I did persuade several people on various group chats I’m on to switch.
I think it's a lost cause. Anthropic is still getting used at Palantir[0], their software is used in strike planning whether they consent or not. We can support them all day and fight OpenAI to the last breath, but ensuring AI is used responsibly is not up to any of us. It's the government's job to hold itself accountable, and they can't do it. By digging in their heels, Anthropic is preparing for an unwinnable fight against an enemy that doesn't play fair.
Considering how many lines Anthropic has crossed, it all feels like forced outrage to me. I feel ethically justified supporting none of these companies, it's reminiscent of the forced duopoly between iOS and Android.
