Andrej Karpathy on X: "100% Software 2.0 computer.Just a single neural net • Hacker News Board

Re: Andrej Karpathy on X: "100% Software 2.0 computer.Just a single neural net

#21

Post by poldk » Mon, Jul 01, 2024, 1:24 AM UTC

In implementation neural nets end up being a variety of different chunks of code like we already have. They’re never just one file with one magical pattern doing all the magic

Software industry is again re-inventing the marketing to justify sitting and doing largely the same job

Re: Andrej Karpathy on X: "100% Software 2.0 computer.Just a single neural net

#22

Post by DrMiaow » Mon, Jul 01, 2024, 1:40 AM UTC

That sounds a lot like what I have been building, but not one net. A system for evolving a system that can handle many nets. https://www.youtube.com/watch?v=sqvHjXfbI8o

Re: Andrej Karpathy on X: "100% Software 2.0 computer.Just a single neural net

#23

Post by dzhiurgis » Mon, Jul 01, 2024, 2:01 AM UTC

weare138 wrote: post #18

boznz wrote: post #17

Earlier quoted context omitted.

It is not magical if it is allowed by the laws of physics

That's why I'm introducing my new Quantum Heisenberg computer. It works until you observe it.

Sounds like most of my code :(

Re: Andrej Karpathy on X: "100% Software 2.0 computer.Just a single neural net

#24

Post by djaouen » Mon, Jul 01, 2024, 5:26 AM UTC

Well that doesn’t sound dangerous at all! /s

Re: Andrej Karpathy on X: "100% Software 2.0 computer.Just a single neural net

#25

Post by pointlessone » Mon, Jul 01, 2024, 6:20 AM UTC

We spent good 60 years to get OKish text rendering. I mean, some of us can tell the difference between Arial and Helvetica. Now imagine adding a new font to an OS like that: “Sure, boss! We only need to generate 5b examples for training and just 20 years of GPU time.”