[phreeza]

And then the gnashing of teeth 2 years later when the terms or the availability of the API change, making half of the startups nonviable.

[rvz]

Exactly. Building an entire startup on top of someone else's API (Especially when the big players can also access it) doesn't seem to be a great idea.

The only winner here is always OpenAI.

[msoad]

with enough resource you can recreate what the paper describes. I'm still not sure if this stuff useful to sell as a service though

[rosstex]

You could also just steal OpenAI's cluster in a midnight heist, that might be easier.

[kyrieeschaton]

There are multiple groups working on reimplementations of GPT-3. It's not that difficult, just requires a certain level of compute resources which a lot of BigCos and large organizations have on tap already.