es: is AI progress actually going to continue and even accelerate as fast as Kokotajlo et al say it will?
Ironically we don’t think it’ll go quite that fast either, as you can see from Footnote 1 on the first page of AI 2027. I am feeling bad for not proclaiming that more often to ward off misconceptions. We have a lot of uncertainty about timelines!
:( Jonas was telling me to name it AI 2028… I should have listened to him… Eli was telling me to name it “AI Endgame...” I didn’t like the sound of that as much but maybe it would have been better...
It’s safer to underestimate AI takeover timelines rather than overestimate it as it could make humans more aware and act faster to prevent them.
It seems that people have misunderstood what I wanted to say which was partly my own mistake. should have used the word timeline above instead of scenario.
Ironically we don’t think it’ll go quite that fast either, as you can see from Footnote 1 on the first page of AI 2027. I am feeling bad for not proclaiming that more often to ward off misconceptions. We have a lot of uncertainty about timelines!
You did name it “AI 2027” ;)
:( Jonas was telling me to name it AI 2028… I should have listened to him… Eli was telling me to name it “AI Endgame...” I didn’t like the sound of that as much but maybe it would have been better...
It’s safer to underestimate AI takeover timelines rather than overestimate it as it could make humans more aware and act faster to prevent them.
It seems that people have misunderstood what I wanted to say which was partly my own mistake. should have used the word timeline above instead of scenario.