4 ms·
John Carmack thinks there's a 60% chance AGI will happen in 2030 or before. There was an article showing that most predictions in AI has been too pessimistic an
by senttoschool 4y ago
John Carmack thinks there's a 60% chance AGI will happen in 2030 or before. There was an article showing that most predictions in AI has been too pessimistic and that breakthroughs happened sooner than predicted.
- anigbrowl 4y agoI agree with him, but this article is still trashy clickbait.
- ggm 4y agoThis is specious accuracy in odds. Carmack has abosolutely no basis to assign a probability to this breakthrough, and it's as easy to argue its 1000:1 against on the same evidence. If he really said this, he's even more self-promoting than I thought. "There is a 50% chance of fusion power in my lifetime" -based on what evidence?
- senttoschool 4y agoBased on reputation. Ever heard of it? The world runs on reputation as much as actual competency.
- ggm 4y agoI don't mind reputation existing but his basis of reasoning is utterly specious at this point, because nothing. No single thing, whatsoever, from the AI sector would suggest anyone else thinks this is plausible, Aside from other self promoters like Kurzeweil. This is literally proposing the million monkeys are on the brink of Shakspear. Pick up the phone, find Scott Aarenson, and ask him to assign a probability that AGI is coming. I hope you listen to his reasoning why this is both an insanely silly question, and unanswerable at any confidence. Scott also has "reputation" -and peer review. Right now Carmack has "I can code inner loops fast" I am coming back late to edit this: you have literally invoked "an appeal to authority" here. Your basis of reasoning is "Because it's Carmack it can't be wrong"
- senttoschool 4y agoAn appeal to authority is a core part of human behavior. Also, I never said this: "Because it's Carmack it can't be wrong". Of course he can be wrong. Anyone can be.