If you run a startup, you need Brex.

Brex corporate cards give you up to 20 times higher credit limits with no personal guarantee.

And Brex's banking solution offers you the highest returning, lowest risk, treasury product available so you can grow your cash faster.

Brex also offers same hour liquidity so your funds are never tied up.

Forget same day liquidity. Go from invested to spending within tied up. Forget same-day liquidity.

Go from invested to spending within the hour.

Over 30,000 companies, including one in three U.S. startups, use Brex.

Join them at brex.com, backslash, TechCrunch's flagship podcast about the business of startups.

I'm Rebecca Boulon, and this is the episode where we bring on industry experts to help us explore a trend in the tech world and dive deep.

AI is entering a new phase where access to top talent is becoming as important or possibly more important than compute or data.

The market for AI researchers is so overheated. It's starting to resemble pro sports, complete with outsized contracts, talent agents, and unprecedented infrastructure needs. So today, we're talking to Didi Das' principal at Menlo Ventures. Didi's seen this shift from multiple angles. First as an engineer and product lead at Google, Facebook, and the AI startup glean, and now as an investor helping technical founders figure out how to build enduring companies in this new AI landscape. Didi, welcome to the show. Thank you for having me, Rebecca. It's a pleasure to be here. Yeah, really excited to hear your thoughts on this. I mean, you have posted on X about this a little bit. High level, what to make of these insane contracts that we're seeing mostly come out of meta. You know, it's actually pretty insane to me, too, just to see the kind of salaries that have been dished out.

You know, the most recent reports people are saying, I don't know if this is true or not, have been a billion dollars over four years, up from $100 million over four years, which is... I don't think that's true.

Which is insane as it is.

You know, I think what's really happening in the industry is

engineers were always valued as an asset that is somewhat dispensable. A lot of it was because

it's really hard to measure how good an engineer is. If I'm building a company, in the previous

era, I would be like, hey, I need some front-end engineers, maybe I need some back-end engineers. Maybe I need some machine learning people. I know exactly what skills I'm looking for. There's plenty of people which have those skills. And then once they join my company, it's actually quite hard for me to tell who is the outsized performer and who isn't. There's not many signals I have. With AI, everything's sort of flipped, where

it's all of a sudden, there's a very specific skill I'm hiring for, which is I need people who can

train, collect data, run infrastructure, run inference pipelines on these large AI models, and very, very

few people actually have that skill. And I know how big the reward at the end of the tunnel is.

So it kind of does, from a business point of view,

make sense for some people to dish out these crazy salaries,

...