Jeff Klune

And so that's what we've done here.

We've created AI that literally does the entire process with no humans touching the system to go from the original idea and coming up with a creative idea all the way through to a paper that successfully navigated peer review and actually outperformed 55% of the human-authored papers that were submitted to the same venue and got accepted and published through peer review.

So it's not as good as the best human scientists or even the kind of the top tier of human scientists yet, but it is remarkable that it's even possible.

Actually, I think this exact same playbook can disrupt and will disrupt almost every scientific discipline.

Really?

It's going to start with the places where getting experimental data automatically is really easy.

So the easiest place is in a computer or like in AI and computer science.

We do our experiments within software.

And so that was where we started.

And that's the easiest.

There are increasingly people, you know, every day I get an email from people who are hooking these ideas up to automated laboratories.

where the systems that can automatically do pipettes and test tube to automatically do experiments in synthetic biology or material science, chemistry.

And soon the system will be able to do experiments that are suggested to humans who go off and do a little bit more work and a little bit more of a complicated domain.

And eventually they could be suggesting experiments for the most expensive types of experiments we have, like experiments on the Hubble telescope,

So I think we're gonna see things across the spectrum with the lowest hanging fruit first and increasingly into more and more difficult forms of science.

Wow, that is very interesting.

Yeah, so actually the process is very similar.

So one thing that is known, and Newton captured really well, is that scientists see further by standing on the shoulders of giants, those that came before.

So scientists, human scientists, basically read what is currently known, the experiments that have already been conducted, the ideas that have been floated, and often they can then say, huh, oh, if that works, or if the world works this way, then I can try this idea, or what if I do that?

And so it's often the case that humans will look at the same basic set of facts and ideas and independently come up with the same next idea.