News for hackers. Stuff that matters.

infgeoax

Sharing some random thoughts on AI.
Re:Getting 50% (SoTA) on Arc-AGI with GPT-4o on Tuesday June 18, 2024 @08:03PM (#40723361)
Hmm, but is it really "generalizing" or just pulling information from the training data? I think that's what this benchmark is really about: to adapt to something it has never seen before quickly.
Re:Getting 50% (SoTA) on Arc-AGI with GPT-4o on Tuesday June 18, 2024 @01:34AM (#40714356)
Yeah humans can generalize much faster than LLM with far fewer "examples" running on sandwiches and coffee.
Re:Getting 50% (SoTA) on Arc-AGI with GPT-4o on Monday June 17, 2024 @11:05PM (#40713702)
Plus it's unclear if the amount of data used to "train" a human brain is really less than what GPT4 used. Imagine all the inputs from all the senses of a human over a lifetime: the sound, light, touches, interactions with peers, etc.
Re:Getting 50% (SoTA) on Arc-AGI with GPT-4o on Monday June 17, 2024 @10:58PM (#40713654)
Isn't chatGPT already proven to be smarter than many of us in many ways?
Re:Getting 50% (SoTA) on Arc-AGI with GPT-4o on Monday June 17, 2024 @10:55PM (#40713640)
Agree. When I first saw ARC, my reaction was this could possibly be the kind of problem that gives us evolutionary pressure.
Re:Getting 50% (SoTA) on Arc-AGI with GPT-4o on Monday June 17, 2024 @10:52PM (#40713624)
But it's his EPR paper inspired the Bell's inequality and pushed the field further. Yes he was wrong about how reality works, but still he asked the right question.