Ilya Sutskever's SSI Inc raises $1B

Re:Ilya Sutskever's SSI Inc raises $1B by sva_ on Wednesday September 04, 2024 @11:05AM (#41446643)

The neural networks in human brains are very different from artificial neural networks though. In particular, they seem to learn in a very different way than backprop.

But there is no reason the company can't come up with a different paradigm.

[ Parent | on HN ]

Re:Ilya Sutskever's SSI Inc raises $1B by whimsicalism on Wednesday September 04, 2024 @11:11AM (#41446701)

that is very weak evidence for the impossibility claim

[ Parent | on HN ]

Re:Ilya Sutskever's SSI Inc raises $1B by janalsncm on Wednesday September 04, 2024 @12:19PM (#41447472)

It was refuting the weak evidence for possibility stated above.

[ Parent | on HN ]

Re:Ilya Sutskever's SSI Inc raises $1B by whimsicalism on Wednesday September 04, 2024 @12:40PM (#41447740)

cheers i missed that

[ Parent | on HN ]

Re:Ilya Sutskever's SSI Inc raises $1B by calf on Wednesday September 04, 2024 @02:56PM (#41449242)

Do we know that? I've seem some articles and lectures this year that kind of almost loosely argue and reach for the notion that "human backprop" happens when we sleep and dream, etc. I know that's handwavy and not rigorous, but who knows what's going on at this point.

[ Parent | on HN ]

Re:Ilya Sutskever's SSI Inc raises $1B by sva_ on Wednesday September 04, 2024 @05:18PM (#41450951)

I've only heard of one researcher who believes the brain does something similar to backprop and has gradients, but it sounded extremely handwavy to me. I think it is more likely the brain does something resembling active inference.

But I suppose you could say we don't know 100% since we don't fully understand how the brain learns.

[ Parent | on HN ]