Steve Brown

So I could say that it's deciding what to put into the context for the medical question is a form of RAG.

I mean, I'm not using the RAG tools for that.

I'm actually writing algorithms to help you kind of curate it.

But actually what I'm doing is I'm letting you do it.

I'm saying like, hey, create a scenario.

Let's help you populate that scenario with the stuff that's relevant to that scenario.

Well, right now, I mean, this is in private...

data rapidly evolving with more and more of the automation being added every day.

But here's where it's going.

It's like I've got a long medical record right now.

If my question is diagnostic and I'm trying to figure out, hey, what's going on?

What's really going on?

there's a subset of that information that's super important and some other stuff that maybe is less important.

So I don't want the important stuff to get lost in the context with the unimportant stuff.

So I do want to curate that.

And I'll use AI to curate, like right now you can manually curate it, but I'll use AI to curate that.

And generally the more recent stuff is probably more important than the old stuff.

There's like some basic things.

But if my scenario is,

treatment, you know, am I on the right treatment?