[nomilk]

Something surprising Sam Altman mentioned on the Lex Fridman podcast was that it only takes about 6 months to get a top physicist up to speed and productive researching AI.

So the team could largely be newbies to AI (with extremely good fundamental knowledge/skill), rather than folks we've all heard of from the AI scene.

[symfrog]

Does Sam Altman have any relevant technical experience to make that assessment? Sounds like something someone would say that just lost their key technical team members.

[klyrs]

For whatever it's worth, Scott Aaronson went from incisive skeptic to drooling fanboy in just about that long. Sam, likewise, seems prone to mistaking loyalty for expertise at this point in his career.

[perkolator]

There is nothing really surprising about this.

Jim Simmons that recently passed away did exactly this in finance decades ago. He didn't build a team of the brightest minds in finance.

He hired brilliant people that specifically did not work in finance. I think he even mentioned that astronomy was one of his favorite areas to hire from for finance.

I am sure we highly underestimate the indoctrination against new ideas that anyone at the top of a field has been subject to.

Humans love to turn everything into a high school prom king/queen popularity election though even when it is obvious the best people for the job didn't even go to the prom.

[rchaud]

Structured finance is a math gig where the only constraint is legality. And as history shows, math nerds have no clue how to constrain the runaway damage their structured instruments can cause.

I suppose I shouldn't be surprised that AI development appears to be heading the same way.