zlacker

>>himara+(OP)
The main question at this point should be whether Sam's model or Ilya's model is more likely to succeed in the primary goal of the teams, which is (and should be): how do we stay on the leading edge long enough to figure out safety. All the adults in the room want safety [1], it's just how to really get it.

There are very wealthy competitors out there, any of which could end up beating OpenAI if they get half an edge. If you don't beat them, you don't get to figure out safety.

If Sam starts another company, you know deep in your soul he'll have all the backing he could ever dream of. Everyone who can sign a check is dying to get in on this. All the smart talent in the world would love to be employee number 1 through 1000. He's figured that you need the money if you want to stay in the game and he's world-class at making that happen. If OpenAI has all the purity of conviction and never gets another dollar because it all flows to SamCo...do they still win and figure out safety?

(Plus get some profits, attract staff who want to make bank, get full control of the board anyway, etc)

[1] We're nowhere near GPT controlling nukes, elections or the bond market, or desiring to. We need at least a couple massive algo changes before things take off. So some speed at this point isn't thaaat dangerous.

>>richar+Aw
>Everyone who can sign a check is dying to get in on this. All the smart talent in the world would love to be employee number 1 through 1000.

That's sound like the riad to disaster. People who can sign a check and are dying to get in on this, usually lack any moral or ethics.

Not a smart move to work for those people just because you fell for a clever salesman.

>>croes+XD
The alternative is to work for poor idealists, which isn’t generally very good for putting bread on the table.

Something tells me most people are going to go for fucktons of money and working on what they think is interesting. Even if it makes other people even more money.