Comment by laylower
4 months ago
Reading this makes me even happier to pay for Anthropic.
Amodei and his sister saw through the behavior and called it out.
" “Eighty per cent of the charter was just betrayed,” Amodei recalled. He confronted Altman, who denied that the provision existed. Amodei read it aloud, pointing to the text, and ultimately forced another colleague to confirm its existence to Altman directly. (Altman doesn’t remember this.) Amodei’s notes describe escalating tense encounters, including one, months later, in which Altman summoned him and his sister, Daniela, who worked in safety and policy at the company, to tell them that he had it on “good authority” from a senior executive that they had been plotting a coup. Daniela, the notes continue, “lost it,” and brought in that executive, who denied having said anything. As one person briefed on the exchange recalled, Altman then denied having made the claim. “I didn’t even say that,” he said. “You just said that,” Daniela replied. (Altman said that this was not quite his recollection, and that he had accused the Amodeis only of “political behavior.”) In 2020, Amodei, Daniela, and other colleagues left to found Anthropic, which is now one of OpenAI’s chief rivals."
If you think Amodei is significantly different you’re going to be disappointed. There is nothing he has done that can’t be adequately explained as furthering his own interests. Remember how Musk doesn’t like Altman too? It’s because they’re all the same people, competing for the same thing.
I can go with the thesis that individuals need community control (boards, regulations, laws) in order to be accountable but is there some specific evidence that Amodei is the same? It seems like a "both sides" argument.
Community control typically doesn't work to constrain great men, a group has more weak points than an individual (look for example how easily Sam dismantled the OpenAI board decision by applying the right pressure).
The greatest and often only check on power has always been competition or opposition by other great men.
1 reply →
I don’t need to convince you, we’ve been through enough cults of personality, time will tell. But I’ve been right enough times to back myself. Maybe it’s because I grew up around a lot of people like them? They can’t hide that they would say whatever they think you want to hear.
Actually it’s funny: Their lack of empathy/emotional intelligence would also make them susceptible to thinking that talking to an LLM is like talking to a person, so maybe they really did think AGI was around the corner!
3 replies →
I mean he quit what he considered to be a problematic company, founded another one, that one’s models refused to do things that the previous company would do, then his new company refused to do the US government’s evil bidding while the other company happily went along with it.
Not small differences to me.
> I mean he quit what he considered to be a problematic company
Problematic why though? For the reasons publicly stated? Then why isn’t Anthropic just what OpenAI was “supposed” to be then? We know what that was from their charter, and Anthropic is not that.
> then his new company refused to do the US government’s evil bidding while the other company happily went along with it
You’re sure about that are you? I don’t see how you possibly could be, unless you’ve taken the PR at face value, before it was all quietly swept away under the next headline.
4 replies →
The only foundation model CEOs I trust are Demis and Yann
The problem with Altman isn't that he's "furthering his own interests". It's the deceitful behaviour he employs in the process.
There’s been enough divergence between words and actions from Amodei for me to also consider him deceitful, if that’s really the low bar you want to set. I’m not saying he’s worse than Altman, just to be clear.
1 reply →
He passed on a multi-billion dollar government contract on principle, so there's that.
Amodei is another Musk/Altman wannabe with their RSF. That randomly changes from 2.0 to 3.0. Where is the research on fighting and identify AI slop?