Is this real? Come on guys the
drama unfolded “a few days ago”
where Sam future was turning
invisible. Now everything is out of
A sudden set in stone? Wake up.
It’s a drama show … part 2 to be
continued
probably because people who clicked on this particular link were the people who were interested in the drama and your comment added nothing of value to the discussion they wanted to have
At one company we worked we had a courtyard and for some reason there was a separate office outside the main complex all on it's own - we suspect it was a repurposed visitors or security center.
We then noticed that people moving into it never came back to the main building but left - we started calling it the "departure" lounge :).
He is in the "departure" lounge - he just needs to figure it out that his time at OpenAI is over.
He will be welcomed in Google, Meta,.. almost anywhere, even MS. It's not the end of the carrier, rather new beginning, or turning point. Staying at OAI may require sort of apology, which may be impossible depending of his ego size.
Exactly, he could go and train grog. But seriously, some AI companies raise with ease 100MM or even more. Guess Ilya would have no trouble to raise the same amount or even more. He knows for sure what is needed to get to GPT4 level.
> I deeply regret my participation in the board's actions. I never intended to harm OpenAI. I love everything we've built together and I will do everything I can to reunite the company.
Ilya did soemething that is highly unforgivable. He betrayed his friends for an ideal. Greg and Sam trusted him. That is the only reason he is being pushed out and I believe it is a good thing for both. Ego is in the way for everyone right now but in the end, Ilya will be better off. Hopefully he can walk away with some money and take some time to think where he wants to go next. Maybe teach the next group of kids? We need great minds in the Universities. Even if he has crazy woke ideas.
Yudkowsky-derived "Safety" is probably now dead, especially as Meta completely dismantled their safety group while this shitstorm was playing out.
"Safety" now means "least risk for shareholders"
...and if you really push the question you'll get LeCun's response of "Don't worry, our AI will just kill theirs" which is hilariously dystopian but also pragmatic.
In my opinion "safety" has always been about that and it's also a projection of power. If you say "we must ensure safety from this machine" you are asserting the machine has a high level of power. People are attracted to power. Investors are attracted to power. Speaking of "safety" is implicitly speaking of power.
I'm specifically stating that to speak of safety is to speak of power. And to speak of power is to attract investment and those who seek power. Speaking of "Safety" is perhaps not to actually seek to 'protect humanity', but rather to advertise dominance.
I remember going to a talk in Berkeley by Nick Bostrom about AI Control theory after reading his book Superintelligence (in 2014?). What struck me at the time was AI control theory was mostly pointless.
If you put all these guard rails around your AI to keep it from turning everyone into paperclips (etc) then it's just not YOUR AI that turns everyone into paperclips/computronium, it's the person that doesn't put those guard rails in.
Worse by putting those guard rails in you're making your AI less effective that the other guys so it can't even outcompete.
27 comments
[ 2.8 ms ] story [ 34.9 ms ] threadWe then noticed that people moving into it never came back to the main building but left - we started calling it the "departure" lounge :).
He is in the "departure" lounge - he just needs to figure it out that his time at OpenAI is over.
> I deeply regret my participation in the board's actions. I never intended to harm OpenAI. I love everything we've built together and I will do everything I can to reunite the company.
https://twitter.com/ilyasut/status/1726590052392956028
He might get a "sideways" promotion to do something that will frustrate him and make him leave on his own volition.
Reminds of another quote: "If you aim at the king, better not miss"
"Safety" now means "least risk for shareholders"
...and if you really push the question you'll get LeCun's response of "Don't worry, our AI will just kill theirs" which is hilariously dystopian but also pragmatic.
If you put all these guard rails around your AI to keep it from turning everyone into paperclips (etc) then it's just not YOUR AI that turns everyone into paperclips/computronium, it's the person that doesn't put those guard rails in.
Worse by putting those guard rails in you're making your AI less effective that the other guys so it can't even outcompete.