David Solomon

You asked me a question and I'm like, okay, 2022.

Yes, but that's the wrong starting point.

Because if you look at 2022, a bunch of what we've done from 2022 to 2025, and now in 2026, was correcting the stupid things that we did in 2020 and 2021.

And so I think you've got to look at it through a 10-year lens.

And what I would say is at this point,

If you look at a 10-year lens, it shifted to more heavily weighted engineering.

My guess is that's going to shift differently going forward, given the power of these tools and our ability to code.

And you're going to see nuanced changes that probably to some degree reduce the number of people that we start with.

over the next few years, but probably not what you and I would call dramatically.

So I don't want to be predictive because if I throw out a number, I'll hear everybody's going to start holding me to the number.

But I just told you we were running it at 2,500, which by the way, looks similar to what we were running at pre-COVID, but lower than the 3,000 plus we were running in 2021.

My guess is it's going to be in the next three years.

That will contract a little, but not in a way when you and I would say, wow, this is a huge, huge change.

We're still going to hire a lot of people out of school.

That is a great question, Tracy.

I can tell you some examples of things we're doing where I know we'll be able to quantify some significant productivity gains.

but they're more in the category of remaking operating processes.

So what we're doing under what I know you guys have read about, 1GS 3.0.

So if we redo our client onboarding, anti-money laundering, KYC processes, and we had a process that had 3,800 people touching it, not 100% of their job touching it, and now we'll have a process with a few hundred people touching it, that's a productivity, when it's done, I can measure the productivity gain of that process, okay?

That's different than saying to me, OK, you've got a bunch of people who work in investment banking.