OpenAI O3 breakthrough high score on ARC-AGI-PUB

Re:OpenAI O3 breakthrough high score on ARC-AGI-PUB by fastball on Friday December 20, 2024 @08:58PM (#42476910)

What is your personal benchmark for AGI? Turing test was surpassed years ago, ARC-AGI was a next step some quite clever people in the space came up with as a successor, and has now surpassed as well.

So what is your benchmark?

[ Parent | on HN ]

Re:OpenAI O3 breakthrough high score on ARC-AGI-PUB by buzzy_hacker on Friday December 20, 2024 @09:33PM (#42477079)

> Turing test was surpassed years ago

I keep seeing this thrown around, but did anyone actually like go out and do this? I feel like I could distinguish between an AI (even the latest models) and a person after a text-only back and forth conversation.

[ Parent | on HN ]

Re:OpenAI O3 breakthrough high score on ARC-AGI-PUB by fastball on Friday December 20, 2024 @09:55PM (#42477186)

The models you've interacted with have guardrails. If you fine-tuned an LLM with the goal of "convince people you are human" (effectively the opposite of what the major players are doing with their fine-tunes), I am very confident even you would be fooled.

[ Parent | on HN ]

Re:OpenAI O3 breakthrough high score on ARC-AGI-PUB by adamtaylor_13 on Friday December 20, 2024 @09:12PM (#42476980)

Self-learning. Human level intelligence, which isn’t a set of facts but rather the ability to self-learn and correct which LLMs are completely unable to do right now.

[ Parent | on HN ]

Re:OpenAI O3 breakthrough high score on ARC-AGI-PUB by fastball on Friday December 20, 2024 @09:53PM (#42477173)

It is fairly clear to me that self-learning is already possible. Give an agentic LLM the ability to add things to its own context window, or fine tune itself, and it will.

[ Parent | on HN ]

Re:OpenAI O3 breakthrough high score on ARC-AGI-PUB by CamperBob2 on Friday December 20, 2024 @09:21PM (#42477028)

Self-learning

That's exactly what this particular benchmark requires.

[ Parent | on HN ]