Re:Large language models generate functional protein sequenc...
by teekert on Sunday May 14, 2023 @02:01AM
(#35935684)
It’s only a black box because our brains aren’t good enough.
Re:Large language models generate functional protein sequenc...
by detrites on Sunday May 14, 2023 @03:46AM
(#35936167)
But if our brains came up with the black box, can that really be true?
Re:Large language models generate functional protein sequenc...
by daveguy on Sunday May 14, 2023 @01:09PM
(#35939442)
Yes. All that is required to make that true is that our brains come up with a model that we cannot comprehend or that reflects a reality that we do not understand. That has been true since cycles and epicycles were used to describe the motion of celestial bodies and probably long before that.
Re:Large language models generate functional protein sequenc...
by throwawaymaths on Sunday May 14, 2023 @05:10AM
(#35936550)
> Who needs to simulate everything when you can just let a large enough black box approximate a solution
You are kind of describing science here. Take chemistry, it's a series of black boxes: chemical reaction patterns, bohr model, vsepr, orbital shell theory, perturbation theory
the new underlying model of the world. First it was: Who needs to understand an algorithm when you can just simulate everything? Now it's: Who needs to simulate everything when you can just let a large enough black box approximate a solution?