100 sats \ 3 replies \ @OneOneSeven 26 Oct 2023 \ on: Frontier risk and preparedness tech
I understand the point, it just kinda freaks me out. Also seems like they're desperate
"We've seen this thing do some weird shit :D help us prevent omnicide!"
write
preview
200 sats \ 2 replies \ @k00b OP 26 Oct 2023
Given the timing, it might be in response to this recently released research showing ChatGPT can be used to construct exploits.
reply
11 sats \ 1 reply \ @OneOneSeven 26 Oct 2023
Maybe it's just the AI posing as OpenAI and asking for ideas
reply
0 sats \ 0 replies \ @k00b OP 26 Oct 2023
👀 😆
reply