Hacker News Reader

Comment by sebzim4500 | Hacker News Reader

>and no obvious large-scale product-market fit

I'm afraid you are in as much an echo chamber as anyone. 200 million+ weekly active users is large scale pmf

Exactly. There's plenty of return on investment. Knowledge workers around the world are paying nice subscription fees for access to the best models just to do their work. There are hundreds of millions of daily active users already. And those are just the early adopters. Lots of people are still in denial about that they need this to do their work very soon. Countless software companies pay for API access to these models as well. As they add capabilities, the market only becomes larger.

OpenAI is one of a handful of companies that is raking in lots of cash here. And they've barely scratched the surface. It's only a few years ago that Chat GPT was first released. OpenAI is well funded, has lots of revenue, and lots of technology coming up that looks like it's going to increase demand for their services.

There's a very obvious product market fit.

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.

> It’s better at search than Google.

That’s hardly a high bar now.

> And programmers are dependent on it.

Entry level ones, perhaps.

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.

>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.

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.

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

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

> And programmers are dependent on it.

that is clearly not the case

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

but they're not making money and there are plenty substitutes. I bet they have practically zero paid customer retention rate. People say they love it, so what.

>I bet they have practically zero paid customer retention rate

Why do you think that?

I know a few people with a subscription but I don't think I know anyone who has cancelled. Even people who have mainly moved to Claude kept the plus subscription because it's so cheap and o1 is sometimes useful for stuff Claude can't do.

This is my first weekend with no subscription to a model since the day chatGPT4 came out.

I feel like I got to a point of increasingly diminishing returns from the models. I know what they are good at and what they are not good at so I started avoiding what they are not good at.

It reminds me of Jaron Lanier saying years ago that we have to be on guard to not start acting stupid to make AI seem smarter. I had surely started doing that.

This is quite a nice feeling. I can still learn things the model sucks at. I could feel a creep coming into my head that there was no point in learning something the model wasn't good at.

The models are amazing at reinventing the wheel in React. Mapping this as if it scales to all human activity though is completely delusional.

> [people moved kept subscription] because it's so cheap and

...

I'm not sure what your point is? Are you arguing that no company can sustain itself on $20 per month subscriptions? Or just that OpenAI can't?

Or that $20 a month isn't cheap?

Reply on news.ycombinator.com