> I wonder why LLMs like GPT-4, Bard and Claude are so geo restricted at first?

Managing scale while maximizing profit potential? Also, US/UK probably lets them put their strongest linguistic foot forward initially, and there may be additional training done before rolling out to regions with other dominant languages.

> I understand some places have regulatory challenges

That’s probably not the main issue.

They want places with tech startups who will pay for their API.

Thats where there is lots of money to be made. And if they are GPU constrained, then launching in the countries with the highest proportion of future paying customers makes sense.

I'd guess Anthropic considers these 2nd tier markets, so it's not a question whether it's too difficult but whether it's a priority at the moment.

I would say that they want English language only, and not EU.

The hilarious part of that is that the UK has basically all the regulations that they are probably worried about.

Even more hilarious is that everyone in their target audience speaks English.

I don't think GPT4 was geo restricted?

> I wonder why LLMs like GPT-4, Bard and Claude are so geo restricted at first? I understand some places have regulatory challenges but can’t see SG, UAE, or Chile being too difficult.

I'm amused by the inclusion of Chile in this list. I'm a Chilean and I do have access, but through the Anthropic Console, as I already had API Access.

It made me laugh when Google announced how strong Bard 2 is at over a hundred human languages and then restricted the deployed chat app to like… three.

That’s not even region locking, it’s capability locking while simultaneously advertising that very thing!