Sam Parr

You know what's in it, but you don't know what it looks like.

Okay.

You don't know the structure of how it's folded up into this like little tiny knot, a very unique structure.

When you say folding, figure out the shape of the knot.

Like imagine I said, hey, you're going to park this car at this address.

Cool.

But if you don't know what the garage looks like, you're just going to smash into the house, right?

Like you might know the location of it, but you don't know where to park the car.

So how do you park the drug that's going to attack this, that's going to either kill it or enhance it?

You need to know what the shape of it.

So the way they do it is one by one.

So they create this competition to be like, can anyone use computers to predict the protein folding?

Because doing this manually is untenable.

And for years, if you look at the graph, it was like, you know, like kind of this like 20%, 30% prediction accuracy for like a decade.

So Demis decides, he's like, this is what we're going to do.

We're going to throw our resources behind this.

And the first time they do it, they win the competition.

But they're like, great, we're trying to go to the moon and we just have like the tallest ladder.

The ladder doesn't get you to the moon.

And so they were actually incredibly disappointed.