[gauravphoenix]

Why doesn't FB create an API around their model and launch OpenAPI competitor? It is not like they don't have resources, and the learnings (I am referring to actual learning from users' prompts) will improve their models over time.

[whimsicalism]

Because they would prefer this to be commoditized rather than just to be another entrant into this space.

[dbish]

They don’t run a cloud services company and get a ton of data elsewhere already. Not worth the effort (yet) imho. I could see them getting into it if the TAM truly proves out but so far it’s speculation that this would be huge for someone outside of selling compute (ex aws/azure)

[ilaksh]

For starters, the performance isn't competitive with gpt-3.5 and would lose money if they made it price-competitive.

OpenAI has a huge headstart on productizing these models by fine-tuning and optimizing their use quite a bit. And they have probably not actually made money when you consider the massive investment.

But this is still very big for people that need or want to run their own fine-tuned models privately.

[obblekk]

GPU constraints. Time spent serving an API is time not spent training new models (for publishing or R&D).

[minhazm]

They're going to license it to cloud providers and make money from that. This lets them avoid all the other headaches associated with becoming a cloud services provider while still getting some income to pay for this work.

https://blogs.microsoft.com/blog/2023/07/18/microsoft-and-me...

[ipsum2]

There's a million different language model (not wrapper) companies offering APIs already. OpenAI, Anthropic, Cohere, Google, etc. It wouldn't be profitable.

[whimsicalism]

There are really only three companies offering good language model APIs: OpenAI, Anthropic, and Microsoft Azure by serving up OpenAI's models. That is it.

[anonylizard]

That's like saying there's 3 competing search engines (Google, Bing, brave?). Or three competing video hosts (Youtube, tiktok, instagram). Or 3 competing cloud providers.

LLMs are infrastructure level services, 3 is a lot of competition already.

[dontupvoteme]

Because their strategy is to give open source the capability to fight their competitors, which is great for us.

[avisser]

hypothesis: FB's legal team successfully lobbied the execs not to with the position that Facebook does not need to give any governmental body an excuse to investigate them. And the legality of LLMs (re: copyrights) is a fine excuse to drag Zuckerberg in front of Congress.

[berkle4455]

To reduce the valuation of OpenAI.

[esafak]

Why? So researchers don't go there?

[berkle4455]

It's not just researchers, this is free for 99% of commercial use.

[dooraven]

because Facebook is a consumer company and this is an enterprise play.

They enterprisesh plays they've tried Workplace / Parse / Neighborhoods (Nextdoor clone) haven't been super successful compared to their social / consumer plays.

[typon]

Because Ads are more profitable than APIs