4 ms·

I still don't understand why people are building things on top of OpenAI's APIs. It's obvious that in the early days, they'll operate at a loss, with many free

by rpigab 3y ago

I still don't understand why people are building things on top of OpenAI's APIs. It's obvious that in the early days, they'll operate at a loss, with many free features and cheap premium features. It's a good time to try it out and see what it can do, but you can't build anything that relies on it, it's a rugpull waiting to happen. They'll add moderation, raise prices, and you'll have to do the same, your users will leave and the ones who stay will hate you. This is only the beginning. Use self-hostable open source alternatives even if they're less easy to use or require infrastructure, and plan for high electricity and storage bills. Whatever you build will last, as long as you respect your users and don't peek at the cleartext logs of users who didn't opt-in to that. It won't be easy. If OpenAI keeps a monopoly on LLMs, they're gonna become like Microsoft did with Windows, but the enshittification process will be a thousand times faster.

gwern 3y ago
You build it on top of the API to bootstrap and prototype with a very high quality model that takes minutes to get started with. Then once you have operated for a while, you have a ton of logged data to train/retrieve/analyze and can swap in a competitor, downgrade the OA model, finetune instead, or something. In OP case's, this worked perfectly: he built a working product quickly and easily, and... discovered that it was a product he didn't want to work on regardless of token cost, cut his losses, did the post-mortem, and moved on with his life. Imagine if he had wasted a year & a ton of money building his own LLM "because I'll need to economize on tokens and have to build a custom LLM" before he could even launch!
urbandw311er 3y ago
I’m torn on this. Yes, it’s never clever building a business on top of a single platform dependency (eg Twitter, Facebook, Play Store, etc) Equally, if you take the example of the App Store, market pressures have actually forced Apple to reduce their cut. Time will tell I guess. And also whether OpenAI end up with a monopoly, part of a duopoly etc.
pixl97 3y ago
People still build on Microsoft platforms. It turns out actually having a usable platform is more important to some businesses than having freedom and no usable platform.
lmm 3y ago
> It's obvious that in the early days, they'll operate at a loss, with many free features and cheap premium features. It's a good time to try it out and see what it can do, but you can't build anything that relies on it, it's a rugpull waiting to happen. So you make hay while the sun shines. And sure, you probably want a plan for after, but that shouldn't stop you from doing what works now.
- rpigab 3y ago
  It's very rare to make hay in the early days of your product, as is the case here. You need growth and long-term vision, but you don't get those if you're building on top of an unstable giant who doesn't like people creating sexbots with their product. But yeah, sometimes it works, high risk, high reward maybe.

by rpigab 3y ago

gwern 3y ago

You build it on top of the API to bootstrap and prototype with a very high quality model that takes minutes to get started with. Then once you have operated for a while, you have a ton of logged data to train/retrieve/analyze and can swap in a competitor, downgrade the OA model, finetune instead, or something. In OP case's, this worked perfectly: he built a working product quickly and easily, and... discovered that it was a product he didn't want to work on regardless of token cost, cut his losses, did the post-mortem, and moved on with his life. Imagine if he had wasted a year & a ton of money building his own LLM "because I'll need to economize on tokens and have to build a custom LLM" before he could even launch!

urbandw311er 3y ago

I’m torn on this. Yes, it’s never clever building a business on top of a single platform dependency (eg Twitter, Facebook, Play Store, etc) Equally, if you take the example of the App Store, market pressures have actually forced Apple to reduce their cut. Time will tell I guess. And also whether OpenAI end up with a monopoly, part of a duopoly etc.

pixl97 3y ago

People still build on Microsoft platforms. It turns out actually having a usable platform is more important to some businesses than having freedom and no usable platform.

lmm 3y ago

> It's obvious that in the early days, they'll operate at a loss, with many free features and cheap premium features. It's a good time to try it out and see what it can do, but you can't build anything that relies on it, it's a rugpull waiting to happen. So you make hay while the sun shines. And sure, you probably want a plan for after, but that shouldn't stop you from doing what works now.

rpigab 3y ago
It's very rare to make hay in the early days of your product, as is the case here. You need growth and long-term vision, but you don't get those if you're building on top of an unstable giant who doesn't like people creating sexbots with their product. But yeah, sometimes it works, high risk, high reward maybe.