This uncanny moment: AI has us all lurching between hype and hysteria, but what are the real risks and opportunities?

How do you make the value case? How do you convince sort of

It's a good question.

And it's fundamentally, it has to come down to believing that advancement of research and then getting that research into the hands of people is valuable.

So I would narrow it down to say that there's three main outcomes.

And one is companies formed and things that are going out into the world that look like revenue.

It's great.

Two is open source releases and things that are more in the academic space of your typical collaborations, talks, and papers.

And three is new institutions formed, which wouldn't have existed before.

So it's a really catalyzed journey for a lot of these folks who, prior to this experience, wouldn't have formed a company or wouldn't have spoken to somebody in a climate lab across the world.

So a collaboration that would have never occurred before.

If they had just stayed in the frontier lab that they were in.

It's so

funny because I feel like you speak in the language of business, but I feel like I can hear the way this like background in philosophy and

framing.

Different

models factors in.

And I feel like it's such an interesting way to both like approach business and venture capital and investment, but also to be able to tell the stories of these kind of like inventions and innovations we're trying to catalyze.

Yeah.

And something that's hard to measure that we don't measure enough, and I don't think it's really a metric that's out there, is how the infrastructure of these science miracles or these breakthroughs are valuable in themselves and they compound over time.

So things like data and open source models and benchmarks and things like this are not necessarily measured or valued in and of their own right because they don't fall into papers and they don't fall into revenue.

But fundamentally, they're super necessary for us as a society to advance scientific progress, especially as we get more digital tools.

And it falls in between both categories of funders as to who's actually going to fund these kind of data sets that need to get created for chemistry modeling and groups that need to be able to do materials discovery.

And then also the data that needs to get labeled for climate interventions and understanding all of the satellite data that we have.

So both on the spectrum of too much data and too little data, there is a ton of infrastructure that must get built that isn't getting built.

And it doesn't have this clear ROI question other than it will compound and it eventually will come back to meaningful progress.