Conversation
i don’t agree with everything here but it’s very strange to hear pmarca talk about how ai will change everything by doing complex cognitive tasks for us, potentially requiring significant autonomy and creativity, but then turn around and say it’s mindless and can never hurt us
It seems like Marc (without saying it) makes an argument for alignment theory over general AI fear. But it feels like that’s a pretty easy argument to make & not significant.
4
I sort of get it! It’s like, maybe intelligence is great but it’s not godly. The smartest person I know still can’t accomplish much, it’s the social organism of 8 billion humans that accomplishes things.
AI may be able to plug into the human social organism to accomplish things… Show more
2
1
11
I like Marc but he's slinging tech just like the other AI cheerleaders
3
I believe Marc's point is about the kinds of thinking machines will do and our access to their internals — not about their potential at the limit.
The argument is not that they're limited but that anthropomorphizing them leads to bad (and limited!) reasoning about them.
7
i think there’s an assumption we will build a tool that is incapable of being used as a weapon.
a sword that can only cut trees and clear a way for us, but suddenly disappears if you attempt to use it against another human.
fire that warms but doesn’t burn.
3
3
i suspect the word “mindless” is pulling a lot of weight here.
i think he means it won’t be agentic, but autonomy will have agency in some way, so maybe he just thinks it will always be utilised by good people and the room for error is small.
i think things like chaosgpt are… Show more
1
1
Show replies
Maybe he hasn't used it enough?
I spent ~1 month using ChatGTP every single day, and it's hard not to be a little afraid of AGI
Quote Tweet
Can ChatGPT build an entire web application? Is that possible?
I promise you haven’t seen ChatGPT do what I’m about to show you


Show this thread










