Mm-hmm

... okay, you're gonna start here, end here, I, I would...

Ideally, I think the preferred method would be to start in pre-construction, follow through with project management, and end in the field, right? So you can get this, the full process, right? We have processes here.

So we can, you know, start in the beginning a- and see what the project's like at the very end, right, when we're closing out of stuff.

So it was an easier transition I'd like to imagine compared to my other colleagues who were in the, the training or the university program with me.

Thanks again.

Absolutely.

And I mean, uh, just before we end this, like, you know, my biggest thing would, like I've been saying the whole podcast, it, it goes back to that strive for constant growth.

So can you explain to us exactly how the deep tech thesis evolved for Julian and yourselves and how the deep checks kind of network rolls into that?

Personally, when I was building MedStack, which was the SaaS company that we worked on, we actually ran into the problem that led me to Deep Tech and led everyone at the fund to Deep Tech, which is that it's really, really easy to build software that competes with what you're doing right now.

So we have a defensibility crisis in software, and I experienced it firsthand building MedStack.

So we were the first to market.

We were building a privacy and security solution for digital health applications, and we were We built all the tooling for developers to build security into their products.

And then Amazon built it, Google built it, Azure built it.

And then a bunch of other companies were able to just raise tons of money and build a glorified checklist.

And what's the single biggest misconception you hear from software focused investors like ourselves?

Let's hear it from him real quick.

I'm Jacob.

Until recently I ran SuperMaven and now I am excited to run SuperMaven as part of Cursor.

I started Tab9 back in 2018 and launched in 2019 Tab9's deep learning version, which I believe was the first application of a large transformer language model to write code.

And since then, you know, the space has expanded by a huge amount.

You know, when I first started it, people were pretty skeptical that it would even be useful to use these models to write code.

And since then, it's taken off.

I'll let them explain real quick why they went in that direction.