3 ms·
> It is the kind of innovation where anybody will catch up. Technically? Probably yes. But where is your startup going to find millions of users for extra tra
by huijzer 4y ago
> It is the kind of innovation where anybody will catch up.
Technically? Probably yes.
But where is your startup going to find millions of users for extra training data?
- wslh 4y ago> But where is your startup going to find millions of users for extra training data? The (big) community. There are existing decentralized approaches to accomplish this. And I am not mentioning the data that could be consumed beyond DMCAs.
- robotresearcher 4y agoThe startups will use the existing LLMs to bootstrap theirs, getting access to the data by proxy. This will transfer cash to the early LLMs but give a means for startups to get going.
- robotresearcher 4y agoSome startups will use the existing LLMs to bootstrap theirs, getting access to the data by proxy. This will transfer cash to the early LLMs but give a means for startups to get going.
- seydor 4y agoeverywhere? what moat does openAI have?
- ipatec 4y agoMicrosoft Office suite