News for hackers. Stuff that matters.
Re:GPT4 and the Multi-Modal, Multi-Model, Multi-Everything F... by hackerlight on Wednesday March 15, 2023 @05:07PM (#35174725)
> GPT 1/2/3/4 is not AGI.

While true, that's the wrong lesson. The right lesson is GPT 1/2/3/4 are progressively getting closer and closer to AGI. You can't blame us for extrapolating only a few years into the future and planning for it.

Re:GPT4 and the Multi-Modal, Multi-Model, Multi-Everything F... by catoc on Wednesday March 15, 2023 @05:16PM (#35174832)
That is quite a claim. Would be cool if true. Do you have any references to back this up? Why would scaling up turn a model that strings tokens together into something that can reason by itself, unprompted, with intent?
Re:GPT4 and the Multi-Modal, Multi-Model, Multi-Everything F... by pixl97 on Wednesday March 15, 2023 @05:32PM (#35175022)
Why is this whole 'unprompted' even important? A huge portion of human intelligence comes from external prompting (just research how very screwed up infants of exceptional neglect are). A model, or a human is going to need input/output from the external world. We're just arguing the semantics of the IO device at this point.
Re:GPT4 and the Multi-Modal, Multi-Model, Multi-Everything F... by hackerlight on Wednesday March 15, 2023 @05:37PM (#35175069)
> Do you have any references to back this up?

The reference is the GPT-4 release paper. It's getting progressively better at things that require intelligence, like writing code and doing unseen exams across various fields that humans find difficult. The generality and depth of intelligence is progressively improving.

> unprompted, with intent

Neither of these are necessary for AGI. Intelligence doesn't require intent. Intelligence doesn't require something to be "unprompted" (which human intelligence isn't, either, our brain is being constantly prompted by our external senses).