Goalposts

Hacker News has set AI a lot of challenges over the years. Which ones has it met?

2024 Aprilanonylizard

About whether Gemini 1.5 reading Harry Potter proves anything, since the books are in training data.

This, Harry potter is a terrible example, even weak models know the rough story of Harry Potter unprompted.

If you want a real test, go test it on some Japanese light novel, or some harry potter fanfiction, and see if the model actually understands the plot details.

For reference, Opus/GPT-4 know the rough story of moderately popular light novels/mangas without any context given. They however do not precisely understand the fine-grained details of the story, like which character will win in a fight.

An AI given an obscure novel it hasn't seen in training correctly answers fine-grained plot questions, like which character wins a fight.

Has this happened?

Yes 53 (51%)Not sure 37 (36%)No 13 (13%)

Votes cast 1–2 October 2026: 100,590 votes from 9,694 people.