So we need to adapt.

It's really interesting because if you look at, I mean, we're very much focused there on what's happening in the UK and absolutely right.

But this is the same problem we have all over the world.

And it's something I'm very worried about is the fact that, you know, wherever people are taking base models or taking anthropics models or whatever, they're losing that ability to adapt them for the local languages.

I think in India, for instance, I think there's 140 different dialects.

at least that's probably conservative.

And many of those won't be represented in any form of televised programs, documents or whatever that could easily be used for training.

RAG, for instance, in terms of the data you use.