To assess whether Eliezer Yudkowsky is calibrated on AI doom, it seems relevant that in 2016 he said he'd be "pretty shocked" if an AI could pass an unrestricted one-hour Turing test before the end of the world.
12:35 AM · May 23, 2026 · 48.7K Views
Matthew Barnett posts analysis of Eliezer Yudkowsky's calibration on AI doom predictions centered on a 2016 statement about Turing test timelines before the end of the world
NathanpmYoung replies that the statement offers limited evidence on timelines.
12790922127103.6K
Sentiment
Positive users defend Eliezer Yudkowsky's forecasting record against isolated criticism, while negative users call out his inaccurate AI timeline predictions and claim his role damages the safety movement.
Pos
39.4%
Neg
60.6%
14 comments with sentiment.
Cluster Engagement
Digg Deeper
No Digg Deeper questions have been answered for this story yet.
Related links
Posts from X
Most Activity
Most Activity
VIEWS22.6KBOOKMARKS24LIKES87RETWEETS4REPLIES17
Alexandros Marinos 🏴☠️@alexandrosM
Unbelievable Yudkowsky quote. Not only he not believe neural networks had any real chance of a breakthrough, he didn't think any AI had a chance of appearing human like in conversation without the world ending soon thereafter.
Matthew Barnett@MatthewJBar
To assess whether Eliezer Yudkowsky is calibrated on AI doom, it seems relevant that in 2016 he said he'd be "pretty shocked" if an AI could pass an unrestricted one-hour Turing test before the end of the world.
37dViews 22.6KLikes 87Bookmarks 24
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)@teortaxesTex
I think this isn't such a bad take from Yud. I am pretty sure I could suss out any current gen LLM in an hour of an unconstrained Turing test. I know what to look for. They're not devious enough to scheme out *with human-like responses*. @allTheYud do you think you can?
Matthew Barnett@MatthewJBar
To assess whether Eliezer Yudkowsky is calibrated on AI doom, it seems relevant that in 2016 he said he'd be "pretty shocked" if an AI could pass an unrestricted one-hour Turing test before the end of the world.
37dViews 6.4KLikes 67Bookmarks 8
Matthew Barnett@MatthewJBar
@Jsevillamol This prediction was how he operationalized one of the three central premises of his argument with Bryan Caplan about AI doom. I think it's relevant to his track record, even if not conclusive.
Jaime Sevilla@Jsevillamol
@MatthewJBar He has previously owned this was a bad prediction, and has also has made some surprising claims that turned out to be correct eg >16% probability of IMO gold by 2025.
I love holding people accountable as much as anyone, but let's not bash people based in a single example.
37dViews 2.8KLikes 60Bookmarks 2
Minh Nhat Nguyen@menhguin
@MatthewJBar @inductionheads im iffy on eliezer bc it's clear many of his priors were formed well before the transformer paradigm. he argues against "alignment by default" but in ways often entirely unrelated to transformers as they are post-2022
Matthew Barnett@MatthewJBar
To assess whether Eliezer Yudkowsky is calibrated on AI doom, it seems relevant that in 2016 he said he'd be "pretty shocked" if an AI could pass an unrestricted one-hour Turing test before the end of the world.
37dViews 807Likes 26Bookmarks 0
Elizabeth Barnes@BethMayBarnes
Huh, seems pretty reasonable to me, maybe depends on exactly what you're imagining by 'no holds barred'? Seems plausible with spiky capability profile there'll still be things (that can be tested in conversation) where models are detectably worse than humans, or weird behavioral artifacts from training they can't suppress, even when they're pretty superhuman overall
Matthew Barnett@MatthewJBar
To assess whether Eliezer Yudkowsky is calibrated on AI doom, it seems relevant that in 2016 he said he'd be "pretty shocked" if an AI could pass an unrestricted one-hour Turing test before the end of the world.
37dViews 929Likes 16Bookmarks 1
Danielle Fong 🔆@DanielleFong
shocked
Matthew Barnett@MatthewJBar
To assess whether Eliezer Yudkowsky is calibrated on AI doom, it seems relevant that in 2016 he said he'd be "pretty shocked" if an AI could pass an unrestricted one-hour Turing test before the end of the world.
37dViews 2.7KLikes 19Bookmarks 1
Jaime Sevilla@Jsevillamol
@MatthewJBar He has previously owned this was a bad prediction, and has also has made some surprising claims that turned out to be correct eg >16% probability of IMO gold by 2025.
I love holding people accountable as much as anyone, but let's not bash people based in a single example.
38dViews 399Likes 19
Eliezer Yudkowsky@allTheYud
@xriskology @MatthewJBar I've previously abjured the things that I've said before the age of 23. Also nobody should listen to you because you were once wrong when you were five years old.
37dViews 126Likes 6Bookmarks 1
Matthew Barnett@MatthewJBar
To be clear, I'm not sure if current AI could actually pass this test. A lot hinges on how such a test is conducted. I do think the prediction will ultimately end up being wrong though.
Matthew Barnett@MatthewJBar
He said this in the comment section of this post: https://www.econlib.org/archives/2016/03/so_far_my_respo.html
37dViews 446Likes 4Bookmarks 1
Trinley Goldenberg@mattgoldenberg
@prerat the strong version of the turing test is that there's no input that could let you distinguish, the weakest version is that the AI can use tricks like pretending to be a human that don't speak very good english
pick your fighter
37dViews 126Likes 4Bookmarks 1
Matthew Barnett@MatthewJBar
@davidmanheim @Jsevillamol I'd welcome a more comprehensive evaluation of his predictions. What other falsifiable predictions has he made that directly pertain to AI doom (as opposed to unrelated predictions about other topics)?
David Manheim@davidmanheim
@MatthewJBar @Jsevillamol Of course it's relevant, and filtered to support your motivated position about his accuracy. If you wanted to do any kind of evaluation of his track record properly, you'd want to collect a large set and evaluate them, instead of picking out an example where he performed poorly.
37dViews 289Likes 5Bookmarks 0
tautologer@tautologer
@MatthewJBar current AI definitely does not pass this test
37dViews 81Likes 6
Dr. Émile P. Torres (they/them)@xriskology
@MatthewJBar He said the Singularity would happen in 2021, later updating this to 2025. And he predicted that nanotech would suddenly emerge and kill everyone by 2010. No one should take him seriously.
37dViews 231Likes 6
David Sartor@DavidSartor0
@ApriiSR @MatthewJBar I expect him to be proven wrong about this one in the future but it definitely hasn't happened yet, they're really not that close to winning here.
38dViews 93Likes 5
entirelyuseless@entirelyuseles
@MatthewJBar I agree he is not calibrated, but no AI has passed an unrestricted one hour Turing test.
37dViews 108Likes 3
Aprii 🩷💎🔎💜@ApriiSR
@DavidSartor0 @MatthewJBar i have some amount an of intuition that like, they've knocked out of the park the stuff that was supposed to be the hard part
this is not a perspective i'm totally sure of, though
38dViews 90Likes 2
entirelyuseless@entirelyuseles
@thePartyPartyUS @MatthewJBar Do you understand the phrase "Turing Test"?
37dViews 8
Low fat sweaty cat 🐬/acc@Silicon_alien
@teortaxesTex @allTheYud is the models coming across as human a goal any of the labs have? i feel like they care more about avoiding a 4o situation. if they really wanted human like models they could do reverse pangram during post training
37dViews 24Likes 3
David Manheim@davidmanheim
@EmileAndHisBots @MatthewJBar His forecasting track record is actually excellent: https://www.metaculus.com/accounts/profile/108770/
37dViews 23Likes 3
Bronson Schoen@BronsonSchoen
You’re trying to imply much broader claims about AI risk are false because Eliezer believed in faster takeoff than seems likely. Given that I think you’re well aware this isn’t a crux of his reviews on AI risk in the long term, this seems like another “gotcha”. I expect that you’ll claim this is about epistemics or prediction records or something, but this seems like punditry.
37dViews 163Likes 1
Load more posts