[thefz]

> it doesn’t really get its own code

It doesn’t really get its own anything, as it is unable to "get". It's just a probabilistic machine spitting out the next token

[threatofrain]

"Getting things" is a matter of performance, not about the underlying hardware. If I'm an idiot who knows nothing about programming, but every time I slam the keyboard we get good programs, then how useful is it to discuss whether I am in fact empty-headed?

[thefz]

But the discussion here is that it does not output good programs at all.

[threatofrain]

So we might discuss their performance along a gradient and think on their year over year improvement. Current industry performance is of such magnitude that it has persuaded the world to adopt ChatGPT workflows as much as they have. Adjacent to code, one might look to Terry Tao and how he relates to ML workflows in math.

[hitarpetar]

I guess in your arbitrary hypothetical it wouldn't be useful

[CamperBob2]

It's just a probabilistic machine spitting out the next token

As are you and I. Did you have a deeper point to make?

[Kudos]

Hey, I think everyone understands how they work by now and the pedantry isn't helpful.

[BoredPositron]

Its a tale worth repeating because a minuscule percentage of people know or pretend to know how it works. Our view might be a bit skewed here on hackernews but normal people believe llms are thinking machines.

[thefz]

Then if it can't really reason on its own creation how do you expect it to be correct in what it does if it's simply regurgitating code parsed online?

[hitarpetar]

actually I'm not sure everyone does