you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
this post was submitted on 24 May 2026
17 points (94.7% liked)
TechTakes
2586 readers
42 users here now
Big brain tech dude got yet another clueless take over at HackerNews etc? Here's the place to vent. Orange site, VC foolishness, all welcome.
This is not debate club. Unless it’s amusing debate.
For actually-good tech, you want our NotAwfulTech community
founded 2 years ago
MODERATORS
So I was poking around the AI 2027 blog and discovered that they seem to be working on making another scenario, this time titled “AI 2030” they haven’t made any posts about it (that I can find anyways) but if its just an AI 2027 rewrite but with moved back timelines I’d imagine people will be less charitable with them
They've been using the excuse that not everyone who participated in the "AI 2027" project agreed on 2027 as the year it all happens. But if that's the case, why the hell did they call it "AI 2027"?
Gotta love the ex post facto of it all.
Within a month after it was out, they were already building up excuses (calling 2027 their modal number, and admitting their timelines had already slipped back a few months). Also, if you read between the lines of various statements they made, they all but admit they picked 2027 for maximum clout/influence. (Lying is okay if its to stop the AI apocalypse! Or maybe they were all more short-term sort of grifters). Even Eliezer recognized setting a hard and early date would damage the grift for everyone!
Maybe because a lot of the press around it was about Daniel Kokotajlo as a forecaster, and at the time he had a 40% chance of AGI by the end of 2027 (according to them)? idk, still does feel a bit disingenuous
If they advertise themselves as a team of forecasters, but then pick a number that doesn't line up with their forecasts because one team member has a gut feeling or vibes it should be sooner, then that is just another reason not to trust them and to treat them like the clowns they are. Of course, even that reading is pretty charitable, the real reason they picked 2027 is to balance urgency and hype generation with a bit of cushion for when the prediction doesn't pan out.
Bingo. They probably hyped up Kokotajlo as a forecaster BECAUSE he had it earlier than the rest of them, so their prediction could have credence to it
So in AI'27 they predicted 2030 will have "1T Wildly Superintelligent copies thinking at 10000x human speed", "wildly superhuman" coding ability, and "brain uploading", with "biosphere destroying mirror life" on the horizon.
Now they are predicting "maybe it will be able to write C++ in 2030 without constantly falling over (50% probability)".
Seems like a bit of a step back, but I guess we'll see what they put in their fun interactive website once it's ready.