Goalposts

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

2025 MarchIncipient

About whether today's reasoning models would have counted as AGI in 1990.

At least for me, my litmus test (today) would require an AGI model to respond with "huh I really don't know" instead of confidently making up garbage when it, in fact, has no clue.

I don't think I'd have expected differently even in 1990.

An AI model says "I really don't know" instead of confidently making things up when it has no clue.

Has this happened?

Yes 35 (27%)Not sure 33 (25%)No 62 (48%)

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