I think it's a bit silly to list founder/CEO-types and take their opinions as if they know the topic well, unless they have the appropriate AI-related background.
The technical nature of the problem is a tiny piece of the claimed impact, and an even smaller piece of the question of whether the proposed policy would, in fact, resolve any real problem and do so without causing greater harms than it prevents.
From what I've read, Yann LeCun himself allegedly "signed" the letter. His name was removed from the list after he issued a public statement denying he'd signed, and in fact actually disagreed with it.
I guess crazy things like that happen when you consider the content of what a person says, rather than their cultural status or designated authority.
Musk obvious knows a fair amount and is reported to be behind this petition.
He also happens to have claimed he can build AGI faster within Tesla than OpenAI can. He had publicly stated that he resigned from OpenAI’s board because he has a conflict of interest.
He’s obviously thrown money around near this subject, but I don't see any reason to believe he understands much about it.
> He also happens to have claimed he can build AGI faster within Tesla than OpenAI can
Musk has basically never been anywhere close to accurate on claims of how fast he can deliver anything,
>- Elon Musk
the quintessential twitter-bro
>- Steve Wozniak
currently fronts 'eco-friendly' crypto scam
>Jaan Tallinn, Skype co-founder
"Tallinn participates in the effective altruism". Somehow crypto bros love that movement. Tallinn even invested together with SBF in various "ventures", some directly competing with OpenAI (Anthropic).
Which one is there the strongest case for trusting their judgement on this issue and what is that case?
(And are these actual verified signatories, because the letter had a problem with false signatures?)
Now: Bengio talks about the reasons he signed it and the reason is more in line of "machines pretending to be humans" i.e. everything from disinformation, security issues to companies interjecting LLMs where a human should be in the loop, causing all kinds of problems.
Elon Musk is literally a twitter-bro that has shown his lack of technological understanding more times than I can count. Also his "visions" are nothing but badly regurgitated sci-fi tropes, distilled and unrealistic, sprinkled with the power of money on top targeted to people who don't understand technology.
Marcus is constantly saying that the technology is hallucinating and too-early to be deployed.
Hopfield I have no qualms with.
But what do Woz, Jaan, Evan and Chris have to do with AI development?
- Yoshua Bengio
- Elon Musk
- Steve Wozniak
- John Hopfield
- Gary Marcus
- Jaan Tallinn, Skype co-founder
- Evan Sharp, Pinterest co-founder
- Chris Larsen, Ripple co-founder