Brian Maucere

I feel like he was the first one.

He was like, oh, there's this company, Sakana, because he kind of always was bringing in like the world, you know, AI and stuff.

And anyway, if you go to their website, they're not sitting.

It's not like they did the AI scientists two years ago and they had this more recent stuff than any talked about.

These are just some of their papers, some of their blogs, right?

So you have from the 20th, which is what, six days ago, an unofficial guide to prepare for a research position application.

Then you have a day before that repo language models with context repositioning with some really cool animated graphs on there, which I love.

A couple of days before that, extending the context of pre-trained LLMs by dropping their positional embeddings.

Digital Red Queen adversarial program evolution in core war with LLMs.

Like it's not your Sunday read per se, unless you just really get into this stuff.

But if you want to see like sort of,

parts of the bleeding edge of AI, the way that some companies are thinking about where AI is going.

There's many, I am sure.

Sakana is one of them, right?

I would definitely put them in.

And so when I saw this news, I thought one of the things was like, is Google betting against itself?

And I thought, God, no.

And of course, the article was just using that as a hook.

But one of the newsletters

No, absolutely not.