Scott Alexander

“Some big macroeconomic indicator (eg GDP, unemployment, inflation) shows a visible bump or dip as a direct effect of AI ("direct effect" excludes eg an AI-designed pandemic killing people): 15%.”

Part of Alexander's five-year prediction set, which he specifies resolves as of 1/1/2028; he framed it alongside his 'Solow's Law' expectation that AI would not show up as a visible break from trend in the big macroeconomic statistics.

“If all of this is too probabilistic, my modal scenario looks something like AGI in 2031 [AI intelligent enough to do 90% of knowledge work jobs].”

Claim trimmed to the AGI-timing clause to avoid overlapping the separate diffusion/jobs entry from the same sentence.

“Define AGI as AI intelligent enough to do 90% of knowledge work jobs. I think there's a 25% chance of AGI by 2027, a 50% chance by 2034, and a 75% chance by 2045.”

Target date set to the 50% (median) point of Alexander's stated distribution; he justifies it mainly by extrapolating METR time-horizon trends.

“AGI... diffuses throughout the economy until more than half of jobs are automated by the late 2030s. Also around the late 2030s we get Bostromian superintelligence, originally in labs but very quickly diffusing out. GDP goes vertical in the late 2030s and early 2040s, and the point of no return is sometime around then.”

Target date set to the end of the 'late 2030s' window Alexander gives for more than half of jobs being automated.

“Define the superhuman gap as the time between AI that can do 90% of knowledge work jobs, and AI that is obviously smarter than the top human geniuses in 90% of fields (it doesn't have to be the same AI). I think there's a 25% chance the superhuman gap range will be less than 1 year, a 50% chance it will last less than 4 years, and a 75% chance it will last less than 10 years.”

A relative-interval forecast with no calendar deadline: the clock starts only once AI can do 90% of knowledge work jobs (Alexander's AGI definition, which he gives a 50% chance by 2034).

“Define the diffusion gap as the time between the AI that could do 90% of knowledge work jobs, and the time when AI does do even half of knowledge work jobs. I think there's a 25% chance the diffusion gap is less than 3 years, and a 50% chance it's less than 10 years.”

A relative-interval forecast with no calendar deadline; Alexander notes his 75% figure is omitted because it falls past the point where 'diffusion' becomes obsolete as a concept.

“Define the Bostromian superintelligence gap as the time between AGI and an AI which, if given independent control of resources like labs and factories, could accelerate technology by a subjective century in one year (eg if invented in 2030, could produce a level of technology that feels typical of 2130 by 2031). I think there's a 25% chance the Bostromian superintelligence gap will be less than 2 years, a 50% chance it will be less than 10 years, and a 75% chance it will be less than 50 years.”

Conditional forecast measured from whenever AGI (Alexander's definition: AI able to do 90% of knowledge work jobs) arrives; his companion modal scenario in the same post puts AGI at 2031 and Bostromian superintelligence in the late 2030s.

“Given the current amount that corporations are pursuing safety, I think there's a 20% chance that the first AIs to cross the point of no return will want to eliminate the human population. [The point of no return is the point where, if an AI wanted to eliminate humanity, humans would no longer have a plausible chance of stopping it.]”

Narrowed from a two-part entry to the unconditional-on-counterfactuals 20% estimate; in the same post Alexander's modal scenario places the point of no return in the late 2030s / early 2040s.

“My proposed operationalization of this is that on June 1, 2025... we'll give [the best image generating model] the following prompts: 1. A stained glass picture of a woman in a library with a raven on her shoulder with a key in its mouth; 2. An oil painting of a man in a factory looking at a cat wearing a top hat; 3. A digital art picture of a child riding a llama with a bell on its tail through a desert; 4. A 3D render of an astronaut in space holding a fox wearing lipstick; 5. Pixel art of a farmer in a cathedral holding a red basketball. We generate 10 images for each prompt, just like DALL-E2 does. If at least one of the ten images has the scene correct in every particular on 3/5 prompts, I win, otherwise you do.”

Alexander's premature September 2022 victory claim was retracted after Surge AI labelers judged the Imagen outputs inadequate, but when the bet came due on June 1, 2025, GPT-4o/ChatGPT image generation cleared the 3-of-5 bar; official judge Gwern agreed Alexander won and counterparty Vitor conceded.