Illia Polosukhin

And so being able to just like describe the front end and see changes you go has been really like kind of different experience.

So I leverage that.

And then I use every single model to test them out.

And then obviously near AI private for actual, you know, real use cases.

Yeah, no, that's awesome.

Well, I mean, we're trying to do that, you know.

I think the open source is effectively the principle.

I think, I mean, imagine if transformers were published and open sourced.

We actually wouldn't even maybe, like, because Google wasn't leveraging, like, Google was using it for translate, but it wasn't, like, leveraging it at, like, larger scale.

And so we wouldn't actually maybe have like this explosion either.

Yeah.

And then, I mean, the transformer itself is built on ton of research of other people, right?

Like the concept of attention came from, you know, researchers, I think from Meta, if I'm not mistaking, and from Montreal University as well.

The, you know, there's obviously all the work from before.

So like,

It's, you know, PyTorch like we're all using.

So like open source, I think it's the core of this and core of innovation.

And it's kind of sad that we are not there anymore.

Yeah.

I mean, I don't think it's a bad...