WEBVTT


00:00:00.140 --> 00:00:03.980
One of the things that, that you can do with AI is surveillance pricing

00:00:03.981 --> 00:00:06.121
and algorithmic wage discrimination.

00:00:06.300 --> 00:00:07.820
So surveillance pricing.

00:00:07.821 --> 00:00:10.261
Your rabbi on this is a woman called Lindsay Owens.

00:00:10.262 --> 00:00:12.581
Has got a book coming out called, uh, gouged.

00:00:12.582 --> 00:00:14.441
She's the best. She's amazing.

00:00:14.540 --> 00:00:16.660
And then algorithm wage. Wage discrimination is.

00:00:16.661 --> 00:00:18.341
Vena Dubois, a lawyer and a scholar,

00:00:18.342 --> 00:00:20.381
has written about this, and they're both the same thing.

00:00:20.382 --> 00:00:21.021
It's, it's a,

00:00:21.022 --> 00:00:25.661
it's when you use an algorithm to segment an audience

00:00:25.662 --> 00:00:27.341
based on behavioral surveillance data.

00:00:27.342 --> 00:00:28.061
Right, I,

00:00:28.062 --> 00:00:28.861
I've seen what you do.

00:00:28.862 --> 00:00:31.321
I've followed you around and your use of a service.

00:00:31.740 --> 00:00:34.860
And then you start running experiments on the segments to see what

00:00:34.861 --> 00:00:37.601
the highest price they'll accept or the lowest wage they'll accept.

00:00:37.700 --> 00:00:39.780
So on the wage side, this started with Uber,

00:00:39.781 --> 00:00:41.981
where drivers who accepted low ball offers.

00:00:41.982 --> 00:00:43.101
And remember, when you're an Uber driver,

00:00:43.102 --> 00:00:46.541
you have 20 seconds to decide whether the offer is good or bad, right?

00:00:46.542 --> 00:00:47.021
So you're, like,

00:00:47.022 --> 00:00:48.621
looking at the mileage, you're looking at the time,

00:00:48.622 --> 00:00:49.381
you're driving in traffic.

00:00:49.382 --> 00:00:51.061
You have 20 seconds to go yes or no, right?

00:00:51.062 --> 00:00:51.861
Oh, wow.

00:00:51.862 --> 00:00:53.741
Okay. So if you accept a low ball offer,

00:00:53.742 --> 00:00:58.241
your next offer is lower, and it keeps pushing you down, right?

00:00:58.420 --> 00:00:59.220
Uh, now,

00:00:59.221 --> 00:01:01.081
if you're picky, it's higher.

00:01:01.340 --> 00:01:03.300
And then if you start taking the low ball offers,

00:01:03.301 --> 00:01:05.061
they sneak into there, they push it down again.

00:01:05.062 --> 00:01:07.701
So whatever it is, that's your side hustle that lets you Be picky,

00:01:07.702 --> 00:01:09.781
eventually you lose it. Wait, that's.

00:01:09.782 --> 00:01:14.601
That's the Uber algorithm doing that to its own drivers.

00:01:14.740 --> 00:01:15.860
To drivers, yeah.

00:01:15.861 --> 00:01:18.781
So this is algorithm to push their wages down.

00:01:18.782 --> 00:01:20.741
That's how. That's how Uber got profitable,

00:01:20.742 --> 00:01:22.221
right? For 13 years,

00:01:22.222 --> 00:01:24.981
they lost $31 billion of the Saudi royal family's money.

00:01:24.982 --> 00:01:26.781
41 cents on every dollar they brought in.

00:01:26.782 --> 00:01:29.021
They pissed away. They eventually, like,

00:01:29.022 --> 00:01:30.901
put all the other cab companies out of business.

00:01:30.902 --> 00:01:32.901
Induced a lost decade in transit investment.

00:01:32.902 --> 00:01:34.381
Became the only game in town.

00:01:34.382 --> 00:01:37.141
Started screwing the drivers and started raising prices on riders.

00:01:37.142 --> 00:01:39.141
And they decoupled riders from drivers.

00:01:39.142 --> 00:01:42.021
You remember, it used to be if you got surge priced for the ride,

00:01:42.022 --> 00:01:43.701
you were paying for. They got surge,

00:01:43.702 --> 00:01:45.321
paid for the ride, they delivered.

00:01:45.540 --> 00:01:46.620
Now it's totally decoupled.

00:01:46.621 --> 00:01:47.981
That's decoupled. It's. It's.

00:01:47.982 --> 00:01:48.421
It's like,

00:01:48.422 --> 00:01:52.041
it's bizarre world Marxism where you pay according to your ability,

00:01:52.140 --> 00:01:53.820
and they work according to their need.

00:01:53.821 --> 00:01:57.661
It's enshittification. It goes back to the other book. It's.

00:01:57.662 --> 00:01:59.301
It's that enshittification.

00:01:59.302 --> 00:02:02.061
That is enshittification. So I.

00:02:02.062 --> 00:02:03.661
I'm gonna make you really angry here.

00:02:03.662 --> 00:02:04.381
Oh, don't,

00:02:04.382 --> 00:02:05.541
Corey. What?

00:02:05.542 --> 00:02:08.581
No. Groundwork Collaborative in the Roosevelt Institute.

00:02:08.582 --> 00:02:11.101
So Groundwork Collaborative is Lindsay Owens outfit.

00:02:11.102 --> 00:02:14.261
And the. And the Roosevelt Institute published report on nurses. Right.

00:02:14.262 --> 00:02:16.381
So used to be contract nurses were hired through,

00:02:16.382 --> 00:02:17.301
like, a local body shop,

00:02:17.302 --> 00:02:18.541
a staffing agency. Right.

00:02:18.542 --> 00:02:20.481
The travel nurses and people like that.

00:02:20.482 --> 00:02:21.021
Yeah. Yeah.

00:02:21.022 --> 00:02:24.301
Now it's four. Four national apps that compete with each other.

00:02:24.302 --> 00:02:26.061
They All call themselves Uber for nursing.

00:02:26.062 --> 00:02:27.541
So maybe you can see where this is going.

00:02:27.542 --> 00:02:28.361
Oh, boy.

00:02:28.380 --> 00:02:32.000
They buy credit card data from data brokers.

00:02:32.260 --> 00:02:34.700
And the more outstanding credit card debt you have,

00:02:34.701 --> 00:02:36.181
the lower the wage you're offered.

00:02:36.182 --> 00:02:38.821
Oh, my fucking god.

00:02:38.822 --> 00:02:41.761
Because it's a desperation premium,

00:02:42.340 --> 00:02:45.580
right? The company that you're working for

00:02:45.581 --> 00:02:49.161
looks to see your level of desperation

00:02:49.260 --> 00:02:52.720
so that they can lowball you. And. And so.

00:02:54.140 --> 00:02:54.860
So, John,

00:02:54.861 --> 00:02:57.021
like me, you have heard of Tennessee,

00:02:57.022 --> 00:02:59.701
Ernie Ford, you know that there are coal bosses

00:02:59.702 --> 00:03:01.901
who would have done this shit if they could have, right?

00:03:01.902 --> 00:03:04.701
But they weren't going to be able to hire the army of Pinkertons

00:03:04.702 --> 00:03:06.141
to follow every coal miner around

00:03:06.142 --> 00:03:07.541
and figure out who's in the most debt.

00:03:07.542 --> 00:03:10.621
And then get the boiler room full of guys in green eye shades

00:03:10.622 --> 00:03:14.181
to Mark up the ledgers and change their wages 10 times a minute, right?

00:03:14.182 --> 00:03:15.941
This is the thing you need AI for.

00:03:15.942 --> 00:03:20.221
This is what automation gets you on the wage side,

00:03:20.222 --> 00:03:23.081
and the same thing is happening on the price side.

00:03:23.100 --> 00:03:23.820
Where, again,

00:03:23.821 --> 00:03:26.181
like it used to be, demographers would do this kind of cute

00:03:26.182 --> 00:03:27.621
madman style thing where they'd be like,

00:03:27.900 --> 00:03:31.020
we got the little old ladies who own seven cats who live in Sheboygan,

00:03:31.021 --> 00:03:31.941
and we've got this and that. We.

00:03:31.942 --> 00:03:33.401
They would go do focus groups.

00:03:33.402 --> 00:03:37.421
AI just naively buckets people into groups based on behavioral data

00:03:37.422 --> 00:03:39.341
and then automatically runs experiments to see

00:03:39.342 --> 00:03:41.286
The highest price they will accept