And now we're seeing perhaps a second era emerge bifurcated by the basis of Yeah.

Yeah, so in my paper, I talk about data-based algorithmic discouragement versus use-based algorithmic discouragement.

And so data-based algorithmic discouragement is exactly what you said.

The creation process of an AI model was somehow tainted, right? Cambridge Analytica is a great example of that.

where the collection of the data that was then used to train this model, so that creation pipeline was tainted by deceptive act, right? So anything that flows from that deceptive act, any profits or any benefits you get from that deceptive act now needs to be disgorged.

So it sounds in restitution, right? So it It sounds a lot like traditional monetary disgorgement, except we're not disgorging monetary profit, we're disgorging an algorithm, right? So that's how it started, how DFTC started using algorithmic disgorgement, how it's mostly used algorithmic disgorgement.

The South Korea case is also another instance of database algorithmic disgorgement.

So what was your arc to ending up at GW and how did you find your way to AI policy?

It also is deeply embedded in the culture of the city.

When I first moved to the city of Atlanta, Angela, this was in 2009, I scored big time.

Around the corner from the coffee shop where I used to do all my writing, there was a box of just stuff that people couldn't sell at a yard sale from over the weekend.

And I rummaged through that box and I managed to find two OutKast CDs.

And I put them right into the CD player of my 91 Nissan Maxima.

And one thing that really struck me as I was listening to Outkast while driving around was that Outkast was so good about putting a sense of place in the music.

They were calling out specific streets, specific landmarks, and they had an interlude where an airplane is descending onto Hartsfield Jackson Atlanta International Airport and calling out landmarks like the Georgia Dome.

Fame.

Does that naturally make them more risky?

I think there's an argument that yes, they can be riskier, right? When you think about mid-market versus large cap, um, mid-market tends to have, for example, smaller management teams.

They might not have as robust IT systems, and the-- sometimes they do have some credit risk, right? They can have some more concentration, whether it's by region or by customer.

However, I wouldn't say for just blanketing all middle market companies as more risky than large cap.

But that's why in the middle market you see structural differences when it comes to leverage levels versus large cap.

Typically, there's significantly lower leverage, and then two, the credit documentation is just tighter.

We're getting maintenance covenants in mid-market versus in large cap you don't.

Is there anything out there that worries you right now for the, let's say the second half of this year?

Because she was like, "Wow, I'm like your mirror image, because I'm a historian by day, but by night I'm a microbiology nerd."

I've been interested in, um, infectious disease for quite a long time, which, um, I don't, I don't find any kind of friends in my department.

Bald or whoever from like a thousand years ago? Uh, so we asked Christina, the Viking expert.

I don't know.

I've not-- You know, Stuf, I really don't know.

But, you know, technically, Mr. Bald is having this manuscript written for him.

It's in his possession, but that doesn't mean it's his work.

So it becomes a really interesting question, you know, Stuf, of who owns the IP on this.