Re:OpenAI O3 breakthrough high score on ARC-AGI-PUB
by braden-lk on Friday December 20, 2024 @01:22PM
(#42473411)
If people constantly have to ask if your test is a measure of AGI, maybe it should be renamed to something else.
[ Parent
| on HN ]
Re:OpenAI O3 breakthrough high score on ARC-AGI-PUB
by OfficialTurkey on Friday December 20, 2024 @01:31PM
(#42473490)
From the post
> Passing ARC-AGI does not equate achieving AGI, and, as a matter of fact, I don't think o3 is AGI yet. o3 still fails on some very easy tasks, indicating fundamental differences with human intelligence.
Re:OpenAI O3 breakthrough high score on ARC-AGI-PUB
by cchance on Friday December 20, 2024 @01:34PM
(#42473532)
Its funny when they say this, as if all humans can solve basic ass question/answer combos, people seem to forget theirs a percentage of the population that honestly believe the world is flat along with other hallucinations at the human level
Re:OpenAI O3 breakthrough high score on ARC-AGI-PUB
by jppittma on Friday December 20, 2024 @01:45PM
(#42473657)
I don't believe AGI at that level has any commercial value.
Re:OpenAI O3 breakthrough high score on ARC-AGI-PUB
by Jensson on Friday December 20, 2024 @06:28PM
(#42476182)
Humans works in groups, so you are wrong a group of human is extremely reliable on tons of tasks. These AI models also work in groups, or they don't improve from working in a group since the company uses whatever does the best on the benchmark, so it is only fair to compare AI vs group of people, AI compared to an individual will always be an unfair comparison since an AI is never alone.