AI safety's most concise summary ever, from Andrew Rousso. And they said it couldn't be explained...

@AndrewRousso It's bizarre to me that everyone in AI just axiomatically accepts that it's impossible to pause/slow down AI. OF COURSE IT'S POSSIBLE, we slow down dangerous technology all the time! Too often imo (eg nuclear power)

@AndrewRousso Absolutely ridiculously good! In 2 minutes he compressed most of the thoughts I had about AI in the past year or so. Genius! More of this please, sir!

@AndrewRousso Misses that China is the primary reason we can’t slow down

@AndrewRousso Right, sooo... Slow it down for safety, and then China is the one that develops AGI... because that's certainly a better prospect, right? Right?

@AISafetyMemes @AndrewRousso To slow down in the minds of OpenAI and the rest would risk losing the game entirely. So no one wants to do that even though it is probably the right move. Sam acknowledges the grave potential dangers, but is pressing ahead aggressively. He is Machiavellian enough to continue…

@AndrewRousso This will never get old. I mean, not in our remaining lifetimes.

@AndrewRousso Lmaooo the accuracy 😂

@AndrewRousso Somewhat subversive

@AndrewRousso

@AndrewRousso @tylertracy321