I finally got around to reviewing this paper and it's as bad as I thought it would be. 1. Zero data or evidence. Just "we guessed right in the past, so trust me bro" even though they provide no evidence that they guessed right in the past. So, that's their grounding. 2. They
@DaveShapi Thanks for engaging. Some responses (mine only, not anyone else on AI 2027 team): >>> "1. They provide no evidence that they guessed right in the past". In the "Who Are We?" box after the second paragraph, where it says "Daniel Kokotajlo is a former OpenAI researcher whose
@abcampbell @DaveShapi Like a Manifold market?
@slatestarcodex @DaveShapi A very generous response given the tone of the initial post. Good job!
@slatestarcodex @DaveShapi imagine getting cooked this hard in public
@slatestarcodex @DaveShapi the research tab is legit pretty easy to miss fwiw. the way the page is designed with the link in the top right corner next to "about" implicitly communicates that it's not important and I don't need to pay attention to it.
@slatestarcodex @DaveShapi Well done
@slatestarcodex @DaveShapi Destroyed.
@slatestarcodex @DaveShapi lolllll👏🏻
@slatestarcodex @DaveShapi bodied
@slatestarcodex @DaveShapi I can’t believe 2027 has received as much hate as it has. Thought it was very well researched and written. Sad to hear so many lame haters.
@slatestarcodex @DaveShapi Maybe time for an apology from David, for going so hard so fast after them, probably because he didn't like their bottom line, from his current "fast acc will be fine" approach. If you're approaching something like a singularity, prediction often gets hard.
@slatestarcodex @DaveShapi @DaveShapi Please include this reply from @slatestarcodex in the thought-checking process you did with the AIs and see what they think. PS: Thanks for unblocking me.
@slatestarcodex @DaveShapi Pure class 🫡 thank you
@slatestarcodex @DaveShapi I think the big problem isn't the (cool, well researched) futurism, but the tl;dr being 'slow down or AI kills us all', which isn't an argument that anyone paying attention has missed. You & your co-authors hedge a lot, but it still felt like Persuasion.
@slatestarcodex @DaveShapi @DaveShapi take on their bet
@slatestarcodex @DaveShapi Tested latest mini OpenAI models for fast, token-efficient reasoning in an more efficient language, inspired by Neuralese. https://diego.horse/neuralese-...
@slatestarcodex @DaveShapi I just sent in several "bugs". Just wondering -- in cases where you judge it to not be a factual mistake, will the submitter be notified of the reasons for the disagreement?
@slatestarcodex @DaveShapi Dave is just a graphomaniac in denial of facts.
@slatestarcodex @DaveShapi Well, Pauli and vitamin C. Otherwise smart person taken over by complete snakeoil. I wonder if it is (in this case) caused by the nonsensical epistemology you subscribe to.