[foooorsyth]

It’s the fastest product to 100m users ever. Even if they never update their models from here on out, they have an insanely popular and useful product. It’s better at search than Google. Students use it universally. And programmers are dependent on it. Inference is cheap — only training is expensive.

To say they don’t have PMF is nuts.

[dom96]

> And programmers are dependent on it.

that is clearly not the case

[dingnuts]

>It’s better at search than Google

in what world? what it's good at is suggesting things to search, because half of what it outputs is incorrect, so you have to verify everything anyway

it does, slightly, improve search, but it's an addition, not a replacement.

[CamperBob2]

Two years ago half of what it output was incorrect. One year ago, maybe 30% of what it output was incorrect. Currently, maybe 20% of what it tells you is incorrect.

It's getting better. Google, on the other hand, is unequivocally getting worse.

[philipwhiuk]

Rubbish - there's no data that shows accuracy has improved by that much.

[CamperBob2]

I brought every bit as much data to the conversation as you did.

[ceejayoz]

> It’s better at search than Google.

That’s hardly a high bar now.

> And programmers are dependent on it.

Entry level ones, perhaps.

[disgruntledphd2]

I mean, these models are super useful for small defined tasks where you can check their output.

They're also useful for new libraries or things that you're not an expert in (which obviously varies by domain and person, but is generally a lot of stuff).

I'm a data person and have been using them to generate scraping code and web front-ends and have found them very useful.

But I wouldn't trust them to fit and interpret a statistical model (they make really stupid beginner mistakes all the time), but I don't need help with that.

Like, in a bunch of cases (particularly the scraping one) the code was bad, but it did what I needed it to do (after a bunch of tedious re-prompting). So it definitely impacts my productivity on side projects at least. If I was doing more consulting then it would be even more useful, but for longer term engagements it's basically just a better Google.

So yeah, definitely helpful but I wouldn't say I'm dependent on it (but I'd definitely have made less progress on a bunch of side projects without it).

Note: it's good for python but much, much less good at SQL or R, and hallucinates wildly for Emacs Lisp.

[numpad0]

s/programmers/front\ end\ html\ authors/