Porby has clearly defined AGI in their own way, and whatever ambiguous model existing today that you think of as AGI is not a match with their definition of strong AGI.
Yup.
For what its worth, I haven’t noticeably updated my timelines since the post; GPT-4 and whatnot are pretty much what I expected, and I’d be pretty surprised if GPT-5 eats us.
(Edit: my P(doom by date) has actually gone down a bit for technical reasons as I’ve continued research! I’ll probably add a more complete update to this post as a comment when I get around to editing it for the openphil version of the contest.)
Yup.
For what its worth, I haven’t noticeably updated my timelines since the post; GPT-4 and whatnot are pretty much what I expected, and I’d be pretty surprised if GPT-5 eats us.
(Edit: my P(doom by date) has actually gone down a bit for technical reasons as I’ve continued research! I’ll probably add a more complete update to this post as a comment when I get around to editing it for the openphil version of the contest.)