Conversation

Also props to the reporter (Brook Silva-Braga) and the whole CBS team for asking great questions and for being open to the fact that the situation we're in is weird and likely bad. (And respect to Geoff Hinton for being really honest.) I agree with Brook! It seems concerning!
6
229
Here's the rest of the CBS team just reacting like actual sensible humans, it's incredible to see. And, I have to say, it gives me hope? Mainstream people often react better than we maybe thought they would. (They almost sound like rationalists/EAs in 2015 here. Quite funny!)
9
224
If the timeline is 20 years, or even 5, what does a 6 month moratorium on GPT-5 buy us? Does it delay the problem by 6 months, or not at all? Does it get us 6 months closer to a solution, or 6 months further?
10
26
I guess the actual real honest answer is: “we don't know!“ (My guess would be that it gives us slightly more time, but not enough. There have been good criticisms of the moratorium, and good criticisms of those criticisms, which I'll link to.)
6
41
Show replies
Did you use chatGPT to generate this out-of-context edit or something because it's really "Tucker Carlsonny"
1
4
Huh? Nope? I watched the interview, came across that section and thought: “my gosh, Hinton seems to have updated, why haven't I come across this on Twitter yet?“ I took the most memorable bit from the interview, but I think that's fair, I linked to the rest. Can't upload 40 mins?
1
2
Exactly! (One of my favourite things is trying to convince AI ethics and AI alignment people that they're actually basically on the same side. I think there is so much overlap here. “How can we make this safe for humans“ is something we all agree on, and it's already a problem.)
2
25
Show replies
He is answering like a scientist, not PR. So yes, it’s not inconceivable. He can’t rule anything out conclusively. What’s the agenda you’re trying to breathe fire under?
4
42
Show replies