Latest / The Smart Cookies podcast / Why does the INTERNET suck so bad now? With Cory Doctorow

Transcript

Why does the INTERNET suck so bad now? With Cory Doctorow · The Smart Cookies podcast

0:04
Welcome to the Smart Cookies Podcast. I'm here
0:06
with Cory Doctorow, journalist, science fiction
0:08
writer, and author of the amazing book, N -Shitification,
0:11
Why Everything Suddenly Got Worse and What to
0:14
Do About It. Corey, welcome to the program. It's
0:17
a pleasure to be on. Thank you for having me.
0:19
As a big believer in grand openings, I'm just
0:22
going to go for it. Why does the internet suck
0:25
so bad right now? Well, I think it's because
0:29
we used to have constraints that punished companies
0:33
that did bad things. And we got rid of those
0:36
constraints. So now whenever there's an individual
0:39
who's in a decision making position who has a
0:43
bad idea, there's nothing to stop them doing
0:45
it. And whenever there's a factional dispute
0:47
within a company where some people want to advocate
0:49
for users of the platform and other people want
0:52
to advocate for the shareholders at the expense
0:54
of those users, the people advocating for the
0:56
shareholders always win because the argument
0:58
is always like it would be icky to rip people
1:01
off in that way and not we would hurt. If we
1:04
rip people off that way, we lose workers or we'd
1:06
be fined or our competitors would take us over
1:09
or people would make new technology that would
1:11
make our technology obsolete. So in short, could
1:14
we just say this is capitalism then? Or is it
1:17
a tad more complicated than that? Well, I'm not
1:20
like the world's greatest true believer in capitalism,
1:23
but I think that it's unequivocal that we used
1:25
to have a different kind of capitalism. The platforms
1:28
used to be a lot better and now they're a lot
1:29
worse. I'm not saying that the... best way to
1:33
get platforms is to just use markets to get them
1:35
but if you are going to use markets to decide
1:37
how platforms behave then you need to make those
1:39
markets contestable you need to make the firm
1:42
small enough to regulate you need to invigorate
1:45
labor power so that workers can stick up for
1:47
their users and you need to make sure that the
1:52
regulatory environment doesn't stop people from
1:54
making technologies that make their lives better
1:56
because right now what we have is a situation
1:59
where platforms can exploit the flexibility of
2:02
digital tools to change the deal on you every
2:05
time you use them. So prices can change, costs
2:07
can change, search rankings and other aspects
2:12
can change in a way to shift value away from
2:15
you into the platform. But if you want to use
2:17
technology, right, if you want to reverse engineer
2:19
the app so that you and everyone who's gig working
2:22
near you can refuse all lowball offers and force
2:25
the wage up or if you want to reverse engineer
2:27
the app so that everyone you know who's on prime
2:30
can see whether prime users are being charged
2:32
extra because amazon knows that most people with
2:35
prime don't shop around or if you want to reverse
2:38
engineer the app so that you can take your ebooks
2:40
and audiobooks around with you or if you want
2:43
to reverse engineer your printer so that you
2:45
can use third -party ink in it and not use the
2:48
original hp ink which is ten thousand dollars
2:49
a gallon it's the most expensive fluid you can
2:52
buy as a civilian it would be cheaper to print
2:55
your grocery list with the semen of a kentucky
2:57
derby winning stallion right if you wanted to
2:59
if you wanted to change that technology that
3:01
would be within your remit that would be a matter
3:03
between you your devices and maybe a company
3:08
or an individual who is doing the technology
3:09
work that you hope to avail yourself of and the
3:12
manufacturer wouldn't get to step in and ask
3:14
the government to prosecute people who made your
3:17
technology better so that we would rather than
3:20
having a situation where we are defenseless and
3:22
they can do whatever they want with the technology
3:24
we'd have a situation where they were constrained
3:27
and we could use our technology in the ways that
3:29
benefited us ah all right i thought maybe it's
3:34
my Rose -tinted glasses of nostalgia. But I remember
3:37
the time of message boards, straightforward ad
3:40
-free websites, and customized MySpace profiles.
3:43
Extremely fondly. Fast forward then to a couple
3:46
of decades, and I have to battle through mountains
3:50
of AI slop and shitty ads, and then go through
3:53
three verification processes to figure out which
3:56
of the 12 streaming platforms I have to be subscribed
3:59
to if I want to watch a shitty Steven Seagal.
4:02
movie on a friday night so you're telling me
4:04
it's about the fact that everything's been so
4:07
monopolized um and and the fact that we're unable
4:11
to do to to tweak technology anymore that's causing
4:15
my experience on the internet let's concretize
4:18
that just a bit here right so um you have a company
4:21
google right obviously very important structurally
4:24
important internet company um They have had one
4:27
successful internal product to speak of in the,
4:30
and it was a product from the previous millennium.
4:32
It was a search engine. It was really good when
4:35
they launched it 27 years ago, virtually everything
4:36
they've made in house since has been a flop and
4:39
almost all their successes, you know, mobile
4:42
video. Well, it's just, it's just empirically
4:45
true. Mobile video, ad tech, server management.
4:48
uh docs collaboration those are all other people's
4:51
companies they bought and operationalized they
4:53
are they are ops people not innovators and look
4:56
i'm an ex -ops guy that's fine but their story
4:58
is we are innovators and the thing is that under
5:01
the letter of antitrust law of competition law
5:03
that kind of acquisition is unlawful right when
5:06
you're acquiring a competitor to reduce competition
5:08
that is an unlawful acquisition and yet the doj
5:13
the ftc under multiple presidents democrat and
5:16
republican just waved through all those mergers
5:18
and so google was able to build up quite a giant
5:21
war chest and they started to buy all the shelf
5:23
space so they bribed apple more than 20 billion
5:27
dollars a year not to make a competing search
5:29
engine They bought every operating system, every
5:32
browser, every carrier, every platform, so that
5:38
if there was ever a search box in front of you,
5:40
it was wired into Google's server, which meant
5:41
that nobody really invested in a rival search
5:44
engine except for Microsoft, which just rolled
5:45
it out with its OS. And as a result, they became
5:49
not just too big to fail and too big to jail,
5:51
they became too big to care. So in 2019, when
5:54
the DOJ sued Google for antitrust violations,
5:57
They published a lot of memos from Google that
5:59
they'd received through discovery. And one set
6:01
of those memos that Ed Zitron from the Where's
6:05
Your Red At newsletter analyzed in depth talks
6:07
about how Google deliberately decided to enshitify
6:10
its service. So in 2019, Google had hit a wall
6:13
on search growth because they had 90 % market
6:16
share. So people with 90 % market share don't
6:18
grow. You could raise a billion humans to maturity
6:21
and make them Google customers, which is what
6:23
Google Classroom is, but it takes like 18 years.
6:28
They needed some other way to goose revenue.
6:30
And you have this epic battle between two Google
6:32
executives, and you can see it play out in the
6:34
emails and the memos. You have one guy called
6:36
Prabhagar Raghavan, who's the revenue guy for
6:39
Google search. He's an ex -McKinsey guy, came
6:41
from Yahoo. And you have another guy called Ben
6:43
Gomes, who's the tech guy for Google search.
6:45
And he has been with Google since the earliest
6:47
days. He was their data center guy. He started
6:50
with one server under a desk and built out the
6:52
global network of data centers and eventually
6:54
became their... search tech guy and so Prabhakar
6:59
Raghavan's got this great idea he says what if
7:01
we made search worse right if we turn off the
7:03
things that let you one shot a search right like
7:06
spell checking and this thing called query stemming
7:09
where like if you search for pants they do another
7:11
search for trousers in the background context
7:13
clues you know like if there's um You know, a
7:16
monkey that got loose on a highway and you search
7:18
for a monkey. It like elevates those queries
7:20
more highly because everyone's searching for
7:22
that. And that's the thing they're looking for.
7:23
All the stuff that kind of made Google magic
7:25
and let them one shot those queries. We just
7:27
turn all that off. Now you've got to search two,
7:29
three times before you find the thing you're
7:31
looking at. Well, each time you search, that's
7:33
another chance to show you an ad. And so here
7:36
you've got, you know, these two ideas and. if
7:40
google were in a competitive environment if google
7:42
was worried about regulators if google was worried
7:44
about their employees because remember you know
7:46
google hired these brilliant technologists you
7:50
know significant plurality of all the top phds
7:53
from every major university in the world uh they
7:57
would be in a position where this would be unthinkable
8:01
but by the time 2019 rolls around they're already
8:05
getting to the point where they're thinking they've
8:07
got enough workers uh they don't have to be quite
8:09
so deferential to them you know sure they'll
8:11
keep the free kombucha and the massages and the
8:13
ball pit and stuff but like yeah let them get
8:16
angry about this stuff um they're not worried
8:19
about competitors obviously uh they're not worried
8:22
about regulators for good reason i mean for one
8:24
thing google's been invading our privacy since
8:27
you know, the early 2000s with impunity because
8:29
the last time America got a new consumer privacy
8:32
law was in 1988 when Ronald Reagan signed a law
8:36
that bans video store clerks from revealing your
8:38
video rental history. That's the last technological
8:41
privacy threat that Congress has defended us
8:45
against as consumers. And so they're not worried
8:47
about regulators either. So they're sitting there
8:49
completely unrestrained. You have these two factions
8:52
and you can imagine those factions have been
8:54
at Google since the early years. I mean, I know
8:57
people who went to Google who were just scumbags
8:59
and who were just greedy assholes, but they lost
9:03
their arguments for the most part for a couple
9:07
of decades. And now they start winning the arguments,
9:09
right? The discipline, the fear of competitors,
9:13
of workers, of interoperable technology, of regulators,
9:16
that's gone away. And so when the clash occurs,
9:20
it cashes out in favor of a shittier service.
9:23
And that's in shitification. And it's not caused
9:25
by greed. Because we didn't invent greed in the
9:28
late 2010s. And it's not caused by VC because
9:33
VC companies back Google when it was good. And
9:36
it's not caused by laziness. It's not because
9:39
you weren't paying for the product. So you were
9:41
the product. I mean, Apple, you know, when they
9:43
turned off spying for Facebook, they turned on
9:45
spying for Apple to run their own ad network
9:48
with no opt out. You can pay for the product.
9:50
You cannot pay for the product. If the company
9:51
can make you the product, you're going to be
9:53
the product, right? So it's not about. A customer
9:55
reward perk. It's about discipline. It's about
9:58
fear. And so take away all the fear. Take away
10:01
all the discipline. And shitification follows
10:03
like night follows day. So it's like we have
10:05
a superior product. We've conquered the globe.
10:08
We have an idea what to do next. Let's make it
10:10
a bit shitty. That's almost like an ingenious
10:13
idea in a sort of psychopathic Lex Luthor sort
10:16
of way. Because, yeah, what else can you do?
10:19
But by the way. You write that n -certification
10:22
is a three -stage process. Could you break down
10:26
this process? Sure. I can use Facebook as an
10:29
example here. Yeah, yeah. An example would be
10:32
amazing. In stage one of n -certification, you
10:34
have platforms being good to their end users
10:36
and at the same time locking end users in. So
10:39
platform, it's a kind of company that mediates
10:42
between two or more groups of users, right? So
10:44
in the case of Facebook, the end users were us.
10:47
People wanted to find out what other people were
10:50
saying. And Facebook pitched people who were
10:53
already social media users in 2006. This is when
10:57
they opened up to the general public and dropped
10:59
the requirement that you have a .edu address
11:01
and said anyone can join Facebook. And they said,
11:05
hey, you know, sure, you've got an account on
11:08
MySpace, but obviously MySpace is owned by this
11:10
evil, crapulent, senescent Australian billionaire
11:12
named Rupert Murdoch, and he's spying on you
11:14
with every hour that God sends. Come to Facebook.
11:17
We will never spy on you at Facebook. And we
11:20
will only show you the things that you asked
11:22
to see. So pick your, pick your follows. We'll
11:24
show you what they post. So users pile in and
11:27
they lock themselves in. And there's a lot of
11:29
different ways that platforms can lock us in.
11:32
Amazon gets us to pay for a year shipping in
11:33
advance. So it'd be dumb to shop anywhere else
11:35
because we've already paid for the shipping.
11:38
Apple locks you in with proprietary file formats
11:40
and apps. Facebook. They play lock in on the
11:43
easiest setting because we do the lock in for
11:45
them through something economists call the collective
11:48
action problem, which is like how economists
11:50
describe the fact that you love your friends,
11:52
but they're a pain in the ass. So even though
11:54
you're all ready to leave Facebook, you can't
11:56
agree on when you should go or where you should
11:57
go. And so you all stay because you love each
12:00
other more than you hate Facebook. And so that
12:03
collective action problem and the high switching
12:05
costs of being denied the company, the people
12:07
who matter to you on Facebook mean that you're
12:09
stuck there. Once Facebook understands that,
12:12
they understand that they can make things worse
12:13
for you, not out of sadism, although some of
12:16
them are probably sadist, but in order to make
12:18
things better for business customers. And so
12:21
advertisers pile in because now Facebook is spying
12:25
on you and they're offering really low cost,
12:27
high fidelity ad targeting with an extremely
12:31
built out anti -ad fraud operation. So if you
12:34
give them a dollar to show an ad to a specific
12:36
user, that user is getting that ad crammed in
12:38
their face. And publishers crawl in because Facebook
12:42
tells them that if they post excerpts of their
12:44
content with links back to their website, they'll
12:46
stuff it into people's eyeballs, even though
12:48
they never asked to see it. Give them a free
12:50
traffic funnel, let them monetize it as they
12:52
see fit. And then those business customers, they
12:54
get locked in. And I think that if you're someone
12:57
who thinks of yourself as like an end user or
13:01
consumer, it's really easy to lose track of how
13:05
powerful a buyer can be instead of a seller.
13:08
So a monopoly is about a powerful seller, right?
13:12
We all know about that because we played the
13:13
shitty board game at the cottage in the summer,
13:15
right? But a monopsony is a powerful buyer. So
13:20
imagine you own a coffee shop and next door there's
13:23
an office building and 20 % of your sales come
13:26
from the people who work in that office building.
13:28
And the company goes out of business and the
13:29
next day you lose 20 % of your sales. That could
13:32
be like an extinction level event. Right. You
13:35
know, 20 percent drop off in overnight receipts.
13:38
That's persistent. You're going to have to lay
13:40
people off. You might have to get a bridge loan
13:42
to keep paying the loans on your equipment. You
13:45
might not make rent before you know you're out
13:47
of business. Meanwhile, say you're a coffee drinker.
13:50
Right. There's five coffee shops on your block
13:52
and one of them goes under. Okay, well, you shop
13:55
somewhere else, right? Monopsonies are so much
13:57
easier to get and maintain than monopolies. And
14:00
once the buyer is structurally important to a
14:02
firm, they can really squeeze the firm. So you
14:04
have these business customers who are now locked
14:06
into Facebook, advertisers, publishers. They're
14:08
there because we're there. We're there because
14:10
we're holding each other hostage. The advertisers
14:12
and the publishers can't afford to leave. So
14:13
Facebook phase three starts to squeeze them.
14:16
For advertisers, they see ad fidelity dropping,
14:19
ad prices going up, ad fraud explodes. So in
14:22
2017, Procter & Gamble dropped its $200 million
14:26
per year surveillance advertising. That's the
14:28
pharmaceutical giant, right? No, that's a different
14:30
company. Procter & Gamble is a consumer packaged
14:32
goods company. If you go to the grocery store...
14:34
Shampoos and shit. Yeah. Well, and cakes and
14:38
breads and canned goods. I see. Everything in
14:40
your grocery store comes from, to a first approximation,
14:42
either Unilever or Procter & Gamble. Like all
14:44
the packaged goods come from one of two companies.
14:47
Procter & Gamble is one of them. So Procter &
14:49
Gamble had been spending $200 million a year
14:51
on surveillance ads, on what they call behavioral
14:53
ads. And they dropped that to zero and they saw
14:55
a 0 % drop in sales. Because to a first approximation,
14:59
$200 million worth of ads was just going down
15:01
the fraud hole. Publishers, meanwhile, are discovering
15:04
that they have to post longer and longer excerpts
15:07
of their content to Facebook. just to be seen
15:09
by the people who've subscribed, let alone to
15:11
be recommended, they're becoming more and more
15:13
substitutive for their own website. And eventually
15:16
it has to be the whole article. And what's worse,
15:17
if you put a link back to your own website, well,
15:21
Facebook downranks you because that link might
15:22
be malicious. And so now you're just a commodity
15:25
backend supplier to Facebook. And your only path
15:27
to monetization is that corrupt advertising system.
15:32
So this is the end of stage three of inshittification.
15:34
Facebook has harvested all of the value. They
15:37
have left a kind of homeopathic residue of things
15:40
you've asked to see in your own feed. They've
15:43
left a void that they filled with stuff that
15:45
advertisers and publishers are paying to place
15:48
in your feed. Those people are getting ripped
15:50
off to the tunes of billions. They're not having
15:52
a good time either. And then, you know, final
15:55
stage of inshittification, the platform is a
15:56
giant pile of shit, but that's a very brittle
15:58
equilibrium. The difference between like, I hate
16:02
this place, but I can't stop coming back. And
16:03
I hate this place. I'm never coming back. It's
16:05
very thin. And you get one like scandal, one
16:09
Cambridge Analytica, one privacy breach, one
16:11
live stream mass shooting people, both for the
16:14
exits. And then the platform owners, they panic
16:16
and being tech bros, they call it pivoting. So
16:19
Mark Zuckerberg gets out of his sarcophagus one
16:22
morning and says, you know, hearken to me, brothers
16:23
and sisters, for I've had a vision in the night.
16:26
It turns out the future is not as I told you.
16:29
arguing with your racist uncle using the primitive
16:31
text interface that I invented to non -consensually
16:33
rate the fuckability of Harvard undergraduates.
16:36
It is actually for me to convert you and everyone
16:39
you love into a legless, sexless, low polygon,
16:42
heavily surveilled cartoon character so I can
16:44
imprison you in a virtual world I stole from
16:46
a 25 -year -old satirical cyberpunk novel that
16:49
I call the metaverse, right? And that is the
16:51
final stage of enshitification. And how is that
16:55
going for him? Well, it lasted long enough to
16:58
keep the stock market believing that maybe he
16:59
could keep growing even after he dominated the
17:02
sector. And then when they lost confidence that
17:04
that would be the way that he could keep growing,
17:05
he said it was AI. And then when they lost confidence
17:08
in AI, he said it was super intelligence. And,
17:10
you know, who knows what it's going to be next,
17:12
right? Given the way Trump is going, it's probably
17:14
going to be med beds, this imaginary technology
17:16
that you lie in and fixes all of your illnesses.
17:19
And Facebook's going to pivot to med bed. By
17:22
the way, speaking of AI, does it? emergence at
17:25
least the way it's designed and utilized now
17:27
only accelerates this process of n -sheetification
17:30
because it certainly seems like that to me so
17:33
one of the things that ai is pretty good at is
17:35
um uh like doing minor variations to kind of
17:40
edge people towards some state or another. So
17:43
there's this thing called algorithmic wage discrimination.
17:45
That's a term coined by the scholar Veena Dubal.
17:48
She originally used it to describe how Uber pays
17:50
its drivers. So Uber will periodically lowball
17:53
a driver by just a little bit, offering them
17:56
a little bit less per mile or per minute than
17:58
they would normally get. And if the driver accepts
18:01
the ride, then that's the new normal. And then
18:03
they start to lowball them again. So they're
18:05
just tight, tightening the dose down a little
18:08
bit in these small random increments. And the
18:11
idea is like, if you're picky as a driver, it's
18:13
because you've got a side hustle, right? You've
18:15
got something else that lets you be selective.
18:17
But if Uber can like trap you in this lowering
18:21
wage, then maybe whatever it is you used to do,
18:23
you can't do anymore. And so now you're dependent
18:25
on, on Uber. So that's the thing that these things
18:27
are really good at. um this is even worse when
18:30
it's applied to other labor markets so preferentially
18:33
in the united states hospitals source their nurses
18:35
as contract workers uh they want to do union
18:38
avoidance so they don't want to hire staff nurses
18:40
so they understaff the wards and then they hire
18:43
day nurses and these day nurses uh they used
18:46
to get their work through staffing agencies and
18:48
there'd be a few in every town and these days
18:50
uh those have all been replaced by four giant
18:53
apps each of which calls itself uber for nursing
18:55
and just like uber They try to do algorithmic
18:59
wage discrimination. So like I said, we haven't
19:01
had a new privacy law in this country since 1988
19:03
with the Video Privacy Protection Act. And so
19:06
we have this totally unregulated data sector,
19:08
data broker sector, and they will sell anyone
19:11
any information about anyone. So the way that
19:14
these nursing apps work is before you're offered
19:16
a shift, the app will go and connect to a data
19:19
broker and look up your credit. And if you're
19:21
carrying a lot of credit card debt, and especially
19:23
if it's delinquent, you get offered a lower wage,
19:25
right? Because you're probably more economically
19:26
desperate. What that means is that the nurse
19:29
who's inserting your catheter probably drove
19:31
over until midnight the night before and then
19:33
skip breakfast to make rent because their wages
19:35
are being forced down. Right. And so keeping
19:37
people in that desperation equilibrium, it's
19:40
the kind of thing that you can automate well
19:42
with technologies that these days we call AI.
19:45
You know, a few years ago we called machine learning.
19:48
It's just, you know, it's a pretty straightforward
19:50
algorithm, but it's like the kind of thing that
19:52
that AI is pretty good at. Mostly, I'm not as
19:55
worried about AI. As a technology, because it's
19:59
such a disgusting investment bubble that is going
20:02
to burst and take the economy with it. 31 % of
20:05
the S &P 500 is seven AI companies, none of which
20:09
are profitable, all of which are hundreds of
20:12
billions of dollars into CapEx they will never
20:14
recoup on, all of which are losing billions of
20:17
dollars a year. But aren't they saying that Uber
20:21
is still not profitable? all these years depends
20:25
on who you ask i mean uber has found some profit
20:28
so what uber did for a lot of years so uber the
20:30
way that uber got lock -in was the saudi royal
20:33
family gave him 31 billion dollars which they
20:35
funneled through masayoshi uh we work or not
20:39
we work um softbank which funded we work and
20:41
uber uh is funding open ai uh was funding yahoo
20:46
in the years when it bought every promising web
20:47
2 .0 startup and then killed it this guy is like
20:50
a menace and so masayoshi sun gives 31 billion
20:54
dollars in saudi royal money to uber uber loses
20:56
41 cents on the dollar for 13 years because they
21:00
are charging so much less for rides than it costs
21:03
to provide them cities start to under invest
21:05
in transit because why would you ride the bus
21:07
when you get an uber so cheaply and all the other
21:10
cab companies go out of business and so now in
21:12
most places it's just uber or uber and lyft and
21:15
so they have this gigantic lock -in and so the
21:18
question was always well can uber eventually
21:20
make money so for a long time they relied on
21:22
these weird accounting tricks like they would
21:24
say OK, well, we entered all these other markets
21:28
like we went into China and Indonesia and India
21:30
and Japan and whatever. And we started these
21:33
companies and they were all unsuccessful. So
21:36
we sold them to other unsuccessful competitors.
21:40
So like Didi bought Uber China. Right now, those
21:44
companies, they're also not public, but we are
21:48
now holding a bunch of their non -public, non
21:50
-liquid stock. And this year, we think that stock
21:54
has doubled in price. Not because anyone's buying
21:56
it or selling it. It is illiquid. It's a private
21:58
company. But we are going to book that as $11
22:01
billion in revenue for this year, right? So they're
22:04
just like making up numbers. This is how WeWork
22:07
operated too. They just made up numbers and added
22:09
them to their balance sheet. More recently, because
22:12
Uber has doubled the price of rides and halved
22:14
the price of driver compensation on average by
22:18
playing these games with pricing. where, you
22:21
know, they'll do to you what they do to drivers,
22:24
where periodically they'll charge you more. And
22:26
if you accept the ride, they'll charge you more
22:28
again next time. And next time they're trying
22:30
to find the equilibrium of your willingness to
22:31
pay, you know, and they're using behavioral cues.
22:36
At one point, there was an Uber executive who
22:38
said that they were using whether your battery
22:41
was about to die to raise prices. They later
22:43
walked that claim back. I don't know if it's
22:45
true or not. One of the things about Uber's app.
22:48
like all apps is that because it is locked up
22:51
because it has a digital locker in it what if
22:53
you're a gamer you might call digital rights
22:55
management it's illegal to reverse engineer it
22:58
and try and look inside and see how it works
22:59
because in 1998 bill clinton signed a law that
23:02
bans reverse engineering where there's drm involved
23:05
and so we can't really study how uber works we
23:08
can only make indirect observations so there's
23:11
people like the rideshare guy who's got a popular
23:13
youtube channel who will get a bunch of uber
23:15
drivers in a room who all have different degrees
23:18
of how often they drive and what kind of rides
23:21
they accept and all of them will get a page for
23:24
the same ride and they'll show their phones and
23:26
you can see the range of compensation and offer
23:28
most of us do don't do that as riders but i bet
23:31
if you did it as a rider you'd find the same
23:33
thing so there's a possibility that uber is now
23:35
profitable it's inched into profitability by
23:38
squeezing riders and drivers so so very hard
23:44
Before I interrupted you, you said that the AI
23:46
sector is a massive bubble right now. Yeah. So
23:51
I think those foundation models are not long
23:54
for this world. I think the only reason they
23:57
exist is because other people are giving AI grifters
24:00
billions of dollars to keep the lights on at
24:02
the data centers. And those people are going
24:04
to expect a return on their investment. And when
24:06
the market craters, they're going to pull out
24:08
if they don't lose it all. And I think most of
24:11
those servers will just be shut off. I think
24:12
the foundation models are doomed. If you want
24:15
to plan for the future of AI, I think you should
24:17
plan for what you would do if you could buy GPUs
24:20
at 10 cents on the dollar, hire a lot of unemployed
24:23
applied statisticians, and optimize open source
24:26
models that have barely been touched yet in terms
24:29
of optimization and have lots of room for improvement.
24:31
That's the future of AI. And it'll be a bunch
24:34
of like marginally useful utilities that are
24:36
kind of cool. I have a local model on my computer.
24:39
I have this spreadsheet I have to parse once
24:41
every three months of sales that I have to break
24:44
out in this very specific way that's really hard.
24:47
And it takes me four hours. And I broke it down.
24:50
I did it yesterday because it was the first day
24:52
of the quarter. We're recording this on October
24:54
2nd. So it was October 1st yesterday. And I used
24:57
a model running locally on my computer, Lama.
24:59
And I just broke it down with Lama. And I broke
25:03
it down one step at a time so I could check the
25:04
numbers at each step. It took me half an hour.
25:07
Right. That's going to be like a really useful
25:08
thing. You don't have to pay for it. Lama's free.
25:11
Right. And it'll get faster and better with more
25:14
open source hacking. So it's not a business.
25:17
It's a consumer surplus. Back to my experience
25:22
on the shitty Internet these days. Why do all
25:26
these companies want us to download an app instead
25:29
of just using a website? For example, if I want
25:31
to have a look at the time. They are literally
25:34
making me have to download some clunky fucking
25:37
app where I have to sign up with my complete
25:39
biography as well as surrender my mother's maiden
25:42
name or something. And click through a license
25:45
agreement. Yeah, exactly. Of course. Before they
25:47
let me have a look at the dial, Jesus Christ.
25:50
Why is that happening? So I mentioned earlier
25:53
this law, the Digital Millennium Copyright Act
25:55
that Bill Clinton signed in 1998, which makes
25:58
it a crime to break DRM. punishable by a five
26:01
-year prison sentence and a $500 ,000 fine. And,
26:04
um, sorry, break the RM for, for us. So if there's
26:10
any kind of digital lock on a service, so if
26:13
you, if you have a program that's open, that
26:16
doesn't have any access controls that tries to
26:18
stop people from changing how it works, like,
26:20
you know, a browser like Firefox, right? Then
26:23
anyone can modify how it works. So like, There's
26:26
ad blockers for browsers. 51 % of all web users
26:30
have installed an ad blocker. It's the largest
26:32
consumer boycott in human history. And that is
26:37
possible because the browser is open. If the
26:39
browser were closed, if they had done something
26:41
to try and stop you from modifying it, and you
26:43
went in and you broke that, you opened the lock,
26:46
that is a crime, even if the thing that you're
26:49
doing is legal. So even if you just want to block
26:52
ads or protect your privacy. So an app is just
26:55
a website. skinned in the right kind of ip to
26:57
make it a felony to protect your privacy while
26:59
you use it and uh any kind of modification to
27:03
an app is is illegal because you have to break
27:06
the drm the digital rights management the digital
27:08
lock in order to get at it uh and and this is
27:11
why firms really like apps because if you're
27:14
using an app you can't like install another app
27:17
a counter app that checks its math or refuses
27:20
lowball offers or uploads the stuff that you've
27:23
seen to a central repository so other people
27:25
can compare the offers they're getting to the
27:27
offers you're getting or any other things you
27:29
might do to defend yourself you know as a worker
27:31
you might say well i'm going to install an app
27:34
along with all the other workers that i know
27:37
that are going to refuse any kind of lowball
27:39
offers from the gig work platform, that is going
27:43
to monitor, say, seven gig work platforms and
27:45
see which one is paying the most at any given
27:47
moment, breaking down by mile, by estimated time,
27:51
and so on, that lets me get the most out of the
27:53
platform, that lets me defend my interests at
27:56
the expense of the shareholders of the company
27:58
that made the tool. By the way, in a recent talk
28:02
you gave for, I believe, CloudFest, you said,
28:05
Entitification is a result of specific policy
28:08
choices, which means we can reverse those choices.
28:11
At the moment, though, it doesn't really seem
28:13
like we're going that way. The infamous photo
28:16
of these extremely charismatic big tech gentlemen
28:20
from Zuck to Bezos swarming Trump at his inauguration
28:24
seems to prove that rather than the state taming
28:27
the tech giants, the tech giants have pretty
28:30
much captured the state. I think that's true.
28:33
I mean, you know, we've had this antitrust resurgence
28:37
that started under Trump one. We kind of forget
28:40
that, you know, the case against Google and Facebook
28:42
started under Trump one. And it was not clear
28:47
exactly why Trump did that. Some of it was just
28:49
capriciousness. Some was petty grievance. But
28:53
the thing is that around the time Trump was bringing
28:56
those antitrust cases, we were seeing an antitrust
28:59
resurgence in Australia. in Japan and South Korea,
29:03
in the EU, in Canada, where I'm from, in the
29:07
United Kingdom, and also in China, where there's
29:11
been pretty vigorous antitrust against tech companies.
29:14
No one really has a theory about why that was
29:16
all happening and why it's continued to happen,
29:18
because while it's slowed down in the US and
29:20
Trump, to the extent he's doing antitrust right
29:22
now, is attacking companies that displease him.
29:26
which are companies that are guilty right they've
29:28
done bad things every major company is an antitrust
29:31
violator but he's not going after them because
29:34
they violated the law he hasn't sorted his list
29:36
by the companies that hurt americans the most
29:38
he's sorted his list by the companies that make
29:40
him the most angry personally but all around
29:45
the world we've seen this surge in antitrust
29:46
and if you're a political scientist this is a
29:49
mystery because it is a bedrock of political
29:51
science theory that the preferences of billionaires
29:54
are the ultimate determinant of our policy. Things
29:57
that billionaires want happen. Things that billionaires
30:00
reject do not. And it doesn't matter how popular
30:02
or unpopular they are with the rest of us. There's
30:04
a big empirical study, Princeton study, that
30:07
analyzed thousands of policy outcomes and came
30:09
to this conclusion that public policy preferences
30:11
have no bearing on policy outcomes when they
30:15
conflict with billionaires' policy preferences.
30:18
And yet here we are. Right. The single most important
30:22
pro billionaire policy we have is monopoly. I
30:25
mean, monopoly is the concentration of power
30:28
and wealth. So, of course, it shits out billionaires
30:31
at the rate of knots. Right. So we we now have
30:35
this anti monopoly surge all around the world.
30:39
It's like gravity has reversed itself, like pigs
30:41
have started flying, like water has started flowing
30:44
uphill. And it's still going on without Trump.
30:48
And in some ways. The fact that Trump is such
30:51
an isolationist is to the benefit of other efforts
30:54
to resist American tech giants. You know, in
30:58
the European Union, they're pursuing this thing
31:00
called Eurostack, where they just want to get
31:03
the hell off of American platforms. And it's
31:05
not because of patriotism. It's because Trump
31:09
has weaponized US platforms to attack European.
31:12
institutions so for example when the chief prosecutor
31:15
at the international criminal court brought charges
31:18
against benjamin netanyahu for genocide that
31:22
chief prosecutor's outlook account was killed
31:26
and lost all their email all their documents
31:28
all their files that they used for these legal
31:30
cases and when microsoft was asked by european
31:35
governments to make privacy guarantees they said
31:38
we cannot promise you that we will not share
31:41
government and corporate data on microsoft servers
31:45
belonging to europeans with the us government
31:47
even if it's on european servers and we cannot
31:50
promise you we'll tell you if we're doing it
31:51
and so you have europe being really interested
31:55
in getting the hell off of the american tech
31:58
stacks and the biggest impediment to that is
32:02
that it's illegal in europe to circumvent to
32:04
jailbreak to break drm And it's because in 2001,
32:08
the US Trade Representative convinced the European
32:10
Union to pass a law just like the Digital Millennium
32:13
Copyright Act. In fact, that is something the
32:16
US Trade Representative has done to all of America's
32:18
trading partners. So, you know, if you're an
32:21
American, the fact that these companies rip off
32:24
all their users, including their American users,
32:27
may be beneficial to you because at least it's
32:29
propping up the stock market. Maybe you've got
32:30
a 401k. But if you're a foreigner. you don't
32:35
even get that benefit. You just get ripped off.
32:37
And so to keep that rip off going, to get the
32:39
most profitable lines of business of the most
32:41
profitable companies in the S &P 500 alive, the
32:44
US trade representative has forced every American
32:47
trading partner to agree not to develop technology
32:50
that would be wildly profitable and save money
32:54
for the people who live in those countries, like
32:55
an alternative app store, an alternative firmware
32:58
for printers or for cars or for tractors. They'll
33:01
let you use generics or get generic service.
33:04
And that was all under the threat of tariffs.
33:07
But a deterrent only works to the extent that
33:09
you don't use it, right? If you say, you know,
33:12
I'm going to burn your house down unless you
33:14
do what I tell you. Well, once you've burned
33:16
my house down, why should I do what you tell
33:18
me? And so now that we are living with tariffs,
33:21
why would we continue everywhere in the world
33:25
to toe the line on... the American priorities
33:31
for tech policy. And so we've seen some very
33:34
muscular antitrust coming out of the EU. Really?
33:37
Could you give us a couple of examples? Because
33:40
from where I'm sitting, I just see these guys
33:42
getting more powerful by the day. So some good
33:44
news would be nice. something called the Digital
33:47
Markets Act. It's a big, gnarly hairball of a
33:49
law. It's got some pretty cool features. So you
33:52
may know about the GDPR, the General Data Protection
33:55
Regulation, the privacy law in Europe. It has
33:58
been pretty much a catastrophe. It was enacted
33:59
nine years ago. And it says that you just can't
34:03
spy on people unless you have their permission.
34:05
And if you ask their permission and they ignore
34:07
you, you just have to let them keep going. You
34:09
can't just like throw up a cookie consent dialogue
34:11
that says, you know, do you want to use cookies?
34:13
Yes, no. Show me choices, blah, blah, blah. And
34:16
then if they and then just make them go through
34:18
like 11 clicks to avoid the cookies or just have
34:22
them say, yes, I accept all the cookies. That's
34:24
illegal. The problem is that the GDPR was structured
34:27
so that the cases under GDPR would be heard in
34:31
the courts of the member state. where the company
34:35
was headquartered. So if it was an Irish company,
34:38
it would be heard by the Irish data commissioner.
34:40
If it was a German company, it'd be heard by
34:42
the German data commissioner and so on and so
34:44
on. So all the tech companies in Europe pretend
34:47
that they're Irish because Ireland is a tax haven.
34:50
And so if they pretend they're Irish, then their
34:53
money floats in a state of untaxable grace somewhere
34:55
in the Irish sea. And what that means is that
34:59
Ireland has to keep these companies happy. Because
35:02
if you're Apple, And you can pay enough lawyers
35:04
to pretend that you're Irish this week. Next
35:06
week, you could pretend you're Cypriot or Maltese
35:09
or Dutch or Luxembourgian. And so they don't
35:14
enforce any laws against companies in Ireland,
35:17
not just tax laws, because they don't want these
35:20
companies to go. So when GDPR cases went to Ireland,
35:23
as these tech companies violated the hell out
35:25
of the GDPR, they just died there. The Irish
35:28
data regulator, like... barely gets out of bed.
35:32
And when he does, he sits around in his underwear
35:34
all day eating breakfast cereal and watching
35:35
cartoons. And so it took like nine years to move
35:39
some of these cases to the federal courts where
35:41
they're finally being heard. But under the Digital
35:43
Markets Act, all these cases are going to start
35:45
in the federal courts. So they're going to bypass
35:47
the Irish court. So it's a very procedural thing,
35:52
but it basically means that within a year, we
35:54
should probably just have the end of spying.
35:57
on users in Europe and the end of these bullshit
36:00
cookie consent dialogues. That's huge, right?
36:03
Because a lot of this stuff is just downstream
36:04
of spying on people. There's another piece of
36:07
this that I'm really excited about. It's the
36:08
interoperability mandates. And it says that companies
36:11
cannot block third parties that want to plug
36:14
into their services. So if you have an... I feel
36:16
like tweak it again. If you don't want to tweak
36:18
it again. That interoperability means if you
36:21
want to tweak something again, right? Yeah, if
36:24
you want to plug one thing into something else,
36:26
if you want to use like someone else's shoelaces
36:28
and your shoes, someone's gas in your gas tank,
36:30
someone's light bulb in your light socket. So
36:33
what they're saying is like you can't block people
36:35
from plugging their own instant messaging into,
36:37
say, WhatsApp. So you'll be able to send messages
36:40
from WhatsApp to other kinds of services and
36:42
back again. Right. And what they've what they've
36:45
said is, in particular, you can't block. third
36:48
-party app stores and third -party payment processors
36:51
so google and apple make 30 cents on every dollar
36:55
that's collected through the app store not just
36:59
the money you spend buying something but the
37:00
money you spend buying an app but the money you
37:03
spend every month on your subscription the money
37:05
you spend every month when you buy something
37:07
when you buy digital asset when you buy an album
37:09
on bandcamp 30 cents out of every dollar is going
37:12
to apple and google to process that transaction
37:15
now um The actual cost of processing a transaction
37:18
in the European Union is between zero and one
37:20
percent. So depending on whether you pick the
37:23
zero number or the one number, it's either a
37:25
one infinity percent markup or a three thousand
37:28
percent markup. But it's a big goddamn markup.
37:31
It's worth a hundred billion dollars to Apple
37:33
alone. This is also something the major console
37:35
vendors do. So this is PlayStation and Xbox,
37:40
Nintendo. are also taking 30 cents out of every
37:43
dollar steam is taking 30 cents out of every
37:45
dollar uh and so this is just an economy -wide
37:48
tax that's being levied by these middlemen and
37:51
so what the dma says is you can't stop people
37:53
from adding their own app store to their device
37:57
and you can't stop vendors right people who sell
38:01
stuff through your app store from choosing their
38:04
own payment processor so if like you sign up
38:07
for patreon through the app store apple doesn't
38:10
get to process the patreon payments patreon does
38:13
and they can find someone who charges one percent
38:15
and not thirty percent to do that payment processing
38:18
and so apple is so pissed off about this that
38:21
they threatened last week to stop selling phones
38:23
in europe which is a hilarious threat to make
38:25
given that there's 500 million affluent consumers
38:29
in europe it's just the most transparently stupid
38:31
unrealistic threat and indeed like A day after
38:35
they made it, they were like, oh, no, you misunderstood
38:37
us. We're not going to stop selling phones in
38:39
Europe. That would be nuts. But they definitely
38:42
said they were going to stop selling phones in
38:43
Europe. They're really angry about this, which
38:45
is great. They're making the right people angry.
38:48
What do you think is driving this wave of antitrust
38:52
activity all of a sudden? Because we did speak
38:54
about state capture by billionaires before. So
38:58
this is pretty remarkable if it truly is happening.
39:02
Yeah, so I think some of it is that it is the
39:05
coalescence of a widespread rage at concentrated
39:08
power and wealth and the corrupting influence
39:10
it has on our society. It's not coherent yet.
39:14
People don't know that that's what they're angry
39:16
about. They think they're angry about the climate
39:18
emergency or xenophobia and racism or the falling
39:21
standard of living or, you know, naked corruption.
39:25
But those things are all downstream of concentrated
39:28
power and wealth. I think there's just a kind
39:31
of political tailwind for this stuff. I also
39:34
think that there's a fair bit of billionaire
39:37
on trillionaire violence. So, you know, you have
39:41
people like Daniel. Sure. Yeah. Yeah. You have
39:45
like like Daniel Lack, right, who started Spotify,
39:48
who's furious that Apple's taking 30 cents out
39:50
of every dollar. So he's bringing antitrust cases
39:54
in Europe. Right. You have the Epic Games make
39:58
Fortnite doing the same thing in America. Right.
40:00
They're really pissed off about this stuff. And
40:03
you have like the telcos being really angry at
40:07
the Internet companies because they're like,
40:08
well, we want to end net neutrality and charge
40:12
rent. if you know youtube wants to show a video
40:14
to someone we want them to to cut us in for a
40:16
piece of it not just the money google spends
40:18
every month on its lines but like a share of
40:21
all the revenue they make from the lines right
40:23
no honor among thieves so so this i think contributed
40:27
to the inertia um and i think that this weird
40:33
coalition of very rich people who are angry at
40:38
obscenely rich people and everyday people who
40:40
are angry at all of them has created a momentum
40:44
that's just very hard to stop. You know, antitrust
40:48
action has been especially vibrant in China,
40:50
which is not a place you think of as being angry
40:54
at its giant tech companies. You know, those
40:56
are typically seen as partners with the Chinese
40:58
state. Certainly, that's the story Facebook has
41:01
told for more than a decade. Wherever there was
41:03
any kind of rumblings of antitrust action against
41:05
Facebook, they'd say, are you crazy? We're, quote,
41:07
defending Chinese or European cyberspace from
41:10
Chinese communism. Right. But it's pretty clear
41:14
that China doesn't think that they're exporting
41:16
Chinese communism with tech companies. They think
41:19
those tech companies are a competitor of theirs
41:22
for power, not an arm of the Chinese state. Xi
41:27
Jinping is rounding up Chinese tech bosses and
41:29
sticking them in gulags. You know, there is this
41:32
maybe also the sense that the prerogatives of
41:36
lawmakers have been usurped by firms and they're
41:38
getting pretty sick of it, too. I mean, you got
41:41
to figure being a Congress jerk sucks. right
41:44
like you have to spend four hours a day calling
41:46
donors right because if you don't do it your
41:48
competitor who is doing it is going to win your
41:51
seat away from you and then these guys call you
41:54
up and demand that you take the most ridiculous
41:56
stands uh and you can't get anything done and
41:59
everyone hates you and you go to town halls and
42:01
they throw tomatoes at you i mean you know like
42:04
it i think that there's just generally A kind
42:08
of fed -up -edness. You know, there's a saying
42:11
in finance, Stein's Law, that anything that can't
42:15
go on forever eventually stops. And I think we've
42:18
reached kind of the Stein's Law singularity here.
42:21
What comes after total degradation, after the
42:25
last phase of encheatification then? Well, ideally,
42:29
the company would fail, right? That's what happened
42:31
to MySpace. From your mouth to God's ears for
42:34
a lot of these guys. They just they just died.
42:36
And that's what you would like to see happen.
42:38
I'm not like the world's biggest believer in
42:39
markets or anything. But, you know, to the extent
42:41
that markets do anything useful, they like do
42:44
what fire does. They clean out the deadwood and,
42:47
you know, they punish bad companies and elevate
42:49
good companies. And when the good companies decide
42:51
to get bad, they punish them, too. Right. So
42:54
the thing that stopped these firms from failing
42:57
is lock in. And this is the most underappreciated
43:00
aspect of what's going on with these platforms.
43:03
It is really hard to convince your friends to
43:06
leave a platform because they're not just there
43:08
for you. I'm still on Facebook and I fucking
43:09
hate that place. Like you said, it's just like
43:12
old aunts spreading anti -vax conspiracies. But
43:16
somehow I'm still there and I don't even know
43:18
why. If you're there and there's like 200 people
43:21
you're connected to, some of them are there because
43:24
they have a rare disease and that's the only
43:26
place they can talk to other people with the
43:27
same disease. Or they're there because they emigrated
43:30
from another country and that's how they stay
43:31
in touch with the people back home. Or they're
43:34
there because that's where their customers or
43:35
their audiences or it's how they organize the
43:39
carpool for their kids little league. Then, you
43:42
know, you don't just have to convince the friends
43:43
that you have to go. You have to convince all
43:45
those other people. to go. And they all have
43:47
their people who have to go. But if you switch
43:50
phone numbers, if you go from T -Mobile to Verizon,
43:53
you don't have to convince everyone who used
43:55
to call you on T -Mobile to get a Verizon SIM.
43:57
Your number just moves over. And that's because
44:00
of interoperability. We could mandate or allow
44:04
for reverse engineering to create systems of
44:08
portability between social media services. You
44:11
know, if you use the new federated open source
44:14
version of Twitter called Mastodon, you know,
44:16
it's a Twitter competitor. It's got a lot of
44:19
different servers, right? So it's like email
44:21
where there's like, you can sign up with Google.
44:23
You can sign up with Outlook. You can have a
44:25
friend who's got a mail server at a data center
44:28
in their garage. You can sign up with them and
44:30
everyone can send email to everyone and receive
44:32
email from everyone. That's how Mastodon works.
44:34
And it can be a little intimidating when you
44:36
join Mastodon because you're like, well, which
44:38
server should I join? There's hundreds of them,
44:40
but actually it doesn't matter because if you
44:43
don't like the server you're on, if you think
44:45
that it's badly managed, badly moderated, if
44:47
you're getting harassed there, if you're getting
44:49
a bad vibe, with one click, you can export a
44:54
file that has everyone who follows you and everyone
44:56
you follow. And you upload that file to a new
44:59
server and everyone just follows you over. The
45:02
same way you can change your phone number and
45:03
everyone follows you over. And it's kind of weird
45:06
to even note which server you're on. It's kind
45:09
of like calling up a friend and saying, hey,
45:11
buddy, you wouldn't believe whose SIM is in my
45:13
phone today. Like, who gives a damn, right? There's
45:16
already people working on interoperability between
45:19
Blue Sky and Mastodon. But the real Holy Grail
45:22
will be letting people leave Facebook, leave
45:25
Twitter, and still be in touch with the people
45:27
that they left behind so that they don't have
45:29
to face these switching costs. Same with Amazon,
45:32
right? It's really hard to get off Amazon because
45:34
Amazon has pushed all the merchants out of your
45:36
neighborhood. There's very few things you can
45:38
buy without going to Amazon. And of course, people
45:40
don't open stores because everyone in the neighborhood
45:41
shops on Amazon. But every item on Amazon has
45:44
a unique identifier called an ASIN, which you
45:49
see in the URL. If you look at an Amazon URL,
45:51
there's like a eight or nine digit number in
45:54
there. There's only like two or three massive
45:57
inventory point of sale systems in the country
45:59
that anyone uses. We could map those ASINs onto
46:03
standard UPC barcodes for all kinds of products.
46:07
And you could have neighborhood cooperative ventures
46:10
that just listed the inventory of all the stores
46:12
in your neighborhood. And if you were looking
46:14
at something on Amazon that was for sale at a
46:17
merchant in your neighborhood, it could replace
46:18
the buy now button with a buy now at the corner
46:21
store button. And, you know, in that way, you
46:25
wouldn't have to choose between not having something
46:29
and shopping locally. Every time there was a
46:31
chance to shop locally, you could buy it locally
46:33
and you could get neighborhood level aggregated
46:37
statistics that would tell merchants what people
46:39
were buying on Amazon that wasn't for sale in
46:41
their stores so that they could start to stock
46:43
it too. So this is a way that you can like incrementally
46:46
move, move people out of these platforms and
46:48
strip them of the power that they derive from
46:50
lock in the power to make things worse. uh with
46:53
uh the secure knowledge that you're not going
46:55
to go anywhere else or you know think of uber
46:57
right uber's got this powerful network effect
46:59
for every driver who's driving uber that's another
47:02
reason for a passenger to pick up the uber app
47:04
because there's more cars on the road with uber
47:06
so if you start a co -op of drivers um drivers
47:11
aren't going to install it because riders don't
47:12
have it and riders aren't going to install it
47:14
because drivers don't have it but imagine if
47:16
you had another app on your phone one that watched
47:19
the uber app right one that violated section
47:22
12 .1 of the DMCA. And when you set up a ride
47:25
with a driver, it would make a little fingerprint
47:29
of that ride and stick it on a server and check
47:32
to see whether a matching fingerprint had been
47:34
made by a driver anywhere in the system. And
47:36
if a driver and a rider could discover that they'd
47:38
both set up a ride with Uber, but they both also
47:42
have the co -op app, they could tear down the
47:43
Uber ride and rebuild it as a co -op ride. So
47:46
there'd never be a reason not to use Uber. And
47:48
there'd also never be a reason not to use the
47:50
co -op app. right that you would get the best
47:53
of all worlds in this in this way now all of
47:56
this is like not only technologically possible
47:58
it's a good business right someone could make
48:00
a shit ton of money selling this stuff uh the
48:03
only reason we don't do it is because there are
48:06
laws that prevent firms from competing with dominant
48:11
actors and again like i'm not the world's biggest
48:15
market true fan but if you believe in free markets
48:18
this is a thing that should really upset you
48:21
What do you think of the concept of techno feudalism
48:24
from Yanis Varoufakis? So some sort of a repeat
48:27
of the good old dark ages days only with the
48:29
big tech companies replacing feudal lords. And
48:31
instead of physical land, they control all of
48:33
our digital spaces, data and online markets.
48:36
Yeah, it's a valid concept. I really do. Yeah,
48:40
I think, you know, there are multiple people
48:42
have used the term techno feudalism to describe
48:43
different things. It's such an easy coinage.
48:46
So Bruce Schneier uses it in a slightly different
48:48
sense. But I think Yanis' sense is really good.
48:51
He is drawing a distinction between two kinds
48:56
of income, profits and rents. So profits are
48:59
things you get from making things that people
49:01
buy. And rents are things that you get from owning
49:04
something that people need to make a profit.
49:06
So if you own the building the coffee shop is
49:10
in, you get rent. Meanwhile, the person who owns
49:14
the coffee shop. is getting a profit by combining
49:16
capital like espresso machines with uh alienated
49:20
labor from uh baristas right and um the interesting
49:26
thing about profits and rents is that rents are
49:28
to a large extent insulated from competitive
49:31
risk so if you own the building the coffee shop
49:35
is in and the coffee shop goes under because
49:37
there's a brand new coffee shop across the street
49:39
and it's poached all the best baristas and all
49:41
the all the most ardent customers And now you've
49:44
got an empty storefront. If you're the landlord,
49:46
that's great news because now you've got an empty
49:49
storefront on a block with a really hot coffee
49:51
shop and you can rent it out again. Meanwhile,
49:54
the entrepreneur, right? The profit seeker who
49:57
owned the coffee shop, they're out of luck. So
50:00
every capitalist aspires to being a landlord,
50:03
right? They don't want to make profits. They
50:05
want to make rents because they want to be insulated.
50:06
They want to, Doug Rushkoff calls it going meta.
50:09
So, you know, don't drive a cab, own a medallion.
50:12
Don't own a medallion, make an app. Don't make
50:15
an app, found an app company. Don't found an
50:18
app company, invest in an app company. Don't
50:21
invest in an app company, buy options in an app
50:24
company. So get as far as you can, as insulated
50:28
as you can. And it's not just a fetish for abstraction.
50:31
Each layer of this is insulated from risk. Obviously,
50:35
the cab driver is exposed to the most risk. But
50:38
as we saw with Uber, medallion owners are exposed
50:40
to risk too. You know, as we saw with a bunch
50:44
of companies that competed with Uber that went
50:45
under, the app company can be exposed to risk.
50:49
So at each level, you are cabining off your risk,
50:52
right? And so this is a way of shoving risk off
50:55
onto other participants in the market. And usually
51:00
shit flows downhill. So you're shoving risk onto
51:04
workers and people who are... more disadvantaged
51:07
you know the great uh uh poets of capitalism
51:10
people like adam smith and ricardo they were
51:13
not uh when they said free markets they were
51:16
not advocating for markets free of regulation
51:17
they were arguing for markets free of rents there
51:21
they said rents were the biggest risk to profit
51:24
and they acknowledged that all profit makers
51:29
would rather be rent takers And they saw the
51:31
role of the state in intervening. You know, Smith
51:33
wrote, seldom do men of commerce gather that
51:36
the conversation does not turn to a conspiracy
51:38
to raise prices. Right. So, you know, these are
51:42
these, you know, you tell this to people who
51:46
consider themselves capitalists today. You read
51:48
quotes from Adam Smith at them, including people
51:50
from the Adam Smith Institute. And they'll look
51:53
at you like you're a Trotskyist. They can be
51:56
happy about that. Yeah. Yeah. And so. So Varoufakis'
52:01
point is that many of the factors of production
52:03
are now owned by tech companies. They own the
52:06
data centers. They own the marketplaces. They
52:08
own the platforms. And they make most of their
52:10
money from rent, not from profit. So Amazon takes
52:14
45 to 51 cents out of every dollar that its sellers
52:17
make. They make so much money charging sellers
52:20
to access Prime that Prime costs Amazon nothing,
52:24
right? That when they ship a parcel, like when
52:27
Amazon sells you something that is in Amazon's
52:29
inventory and not a third -party merchant's inventory,
52:32
it costs them nothing to ship it. It's been fully
52:35
subsidized by companies that are its competitors,
52:38
right? So this is just pure rents. And I think
52:43
what Varoufakis is trying to do here, in addition
52:46
to being quite usefully analytical, is he's trying
52:49
to find a coalition, right? He's trying to say
52:52
like, look, you know, if you are a capitalist,
52:54
you should hate these rentiers. right? Because
52:56
they are getting in the way of your ability to
52:59
make a living. You can be Daniel Ek, right? You
53:01
can be a music monopolist who is screwing every
53:03
musician on earth and you are still getting fucked
53:06
by Apple and Google, right? So how do you avoid
53:09
this rent seeking? Well, you have to end feudalism.
53:11
You have to end techno feudalism. Corey, last
53:14
question very, very quickly. Is a new, better
53:17
internet possible or is that just a pipe dream
53:19
in 2025? And if we're being pragmatic, what can
53:21
we do to accelerate this right now? Because I
53:24
imagine holding a placard saying Peter Thiel
53:27
sucks won't really do much. Well, I think a new,
53:30
better internet is definitely possible. Overcoming
53:33
structural impediments to better infrastructure
53:35
is something that we've been doing since the
53:37
Gilded Age. Our ancestors broke up Standard Oil
53:41
and, you know, uh defeated john d rockefeller
53:45
they weren't smarter than we were they weren't
53:47
more powerful than we were they don't they didn't
53:49
have access to lost knowledge it's not like we're
53:51
trying to recover the secret of building pyramids
53:53
without power tools right we know how to do this
53:56
stuff and right now all over the world there
53:59
is far more interest in doing this than there
54:01
was even a year ago because people want to get
54:03
the hell away from trump and so you know trump
54:06
has been a a a weird mixed blessing here you
54:09
know life gives you stars you make sarsaparilla
54:12
But I really do think that we're having a moment
54:16
everywhere that's not America for thinking about
54:20
what the Internet looks like if you don't allow
54:22
American tech companies to screw you. And if
54:25
you incubate your own domestic tech sector and
54:28
launch it into a stable orbit by raiding the
54:30
massive rents acquired by these American tech
54:34
giants. So I don't think it's a pipe dream at
54:37
all. I think that it is absolutely within our
54:40
grasp and I think it's quite necessary. Because,
54:43
you know, the Internet is the platform we're
54:44
going to organize to resist genocide and the
54:47
climate emergency, xenophobia and fascism on.
54:50
Right. I grew up as a young activist riding a
54:53
bicycle around the streets of Toronto, wheat
54:55
pasting posters to telephone poles. If you think
54:57
you can get people out for a mass demonstration
54:59
without social media, without email, without
55:02
other organizing tools, you need to spend a couple
55:05
of nights out in subzero weather sticking these
55:07
posters up and then tell me what you think. Corey,
55:11
thank you so much for your time. This was amazing.
55:13
It's my pleasure. Thank you very much for listening.
55:15
Please subscribe to the podcast on YouTube, follow
55:16
and rate on Spotify and Apple Podcasts. You can
55:19
also hype videos on YouTube now. Marketing tutorials
55:22
insist. It's really important that I keep harping
55:25
on and hassle you about that, so I'm doing it.
55:27
Also, if you'd like to, please support me on
55:29
Patreon. Again, this is a one -person podcast.
55:32
I do everything by my little old self. There's
55:34
no team behind me, but you can help by hopping
55:37
on Patreon and finding the smart cookies. podcast
55:40
on there and then last but certainly not least
55:43
i have to thank my existing patrons without you
55:47
there would be no smart cookies which makes you
55:50
actual producers of this show so eternal thanks
55:53
goes to That's a long list. And I'm super, super
56:20
grateful that it's getting longer. I'm deeply
56:23
touched. Thanks again for all your support. Bye
56:25
-bye.