WEBVTT

00:00:00.000 --> 00:00:01.879
Hello and welcome to another Deep Dive. Today

00:00:01.879 --> 00:00:06.839
is Monday, February 2, 2026. And if you've been

00:00:06.839 --> 00:00:09.660
tracking the headlines, or honestly just trying

00:00:09.660 --> 00:00:11.300
to scroll through your phone over the last couple

00:00:11.300 --> 00:00:14.160
of weeks, you know things are different. Different

00:00:14.160 --> 00:00:17.320
is one word for it. The digital ground has fundamentally

00:00:17.320 --> 00:00:20.719
shifted under our feet. It really has. We're

00:00:20.719 --> 00:00:22.660
looking at what feels like a completely new internet

00:00:22.660 --> 00:00:25.059
architecture, and it all happened, what, just

00:00:25.059 --> 00:00:28.809
10 days ago. It's the end of an era. And the

00:00:28.809 --> 00:00:32.350
start of. Well, that's what we're here to figure

00:00:32.350 --> 00:00:34.210
out, isn't it? But you're right. We are sitting

00:00:34.210 --> 00:00:36.350
here just a couple of weeks past January 22,

00:00:36.649 --> 00:00:39.710
2026. And that's a date that is going to be in

00:00:39.710 --> 00:00:42.189
the history books for media studies, for political

00:00:42.189 --> 00:00:45.149
science, probably for computer engineering for

00:00:45.149 --> 00:00:47.869
decades to come. That was the day the ink finally

00:00:47.869 --> 00:00:51.119
dried. After years of speculation, the US operations

00:00:51.119 --> 00:00:53.840
of TikTok were officially sold to a US consortium.

00:00:53.859 --> 00:00:56.359
The saga is finally over. I mean, the ban threats,

00:00:56.679 --> 00:00:59.479
the court cases, the executive orders, it's been

00:00:59.479 --> 00:01:01.920
going on for what feels like forever. And it

00:01:01.920 --> 00:01:04.439
all just culminated in this one absolutely massive

00:01:04.439 --> 00:01:07.719
transaction. It is massive. And I think before

00:01:07.719 --> 00:01:09.439
we even get into the weeds of what's happening

00:01:09.439 --> 00:01:11.659
now, we have to just take a second to appreciate

00:01:11.659 --> 00:01:14.400
the sheer scale of what was just sold. Right.

00:01:14.480 --> 00:01:16.379
We're not just talking about some app where teenagers

00:01:16.379 --> 00:01:18.980
do dances. I mean, we are talking about a platform

00:01:18.980 --> 00:01:22.920
that way back in 2021 surpassed Google. Surpassed

00:01:22.920 --> 00:01:25.219
Google. As the most popular domain on the entire

00:01:25.219 --> 00:01:27.180
Internet. I remember when that stat dropped,

00:01:27.299 --> 00:01:29.319
it felt like a glitch in the matrix. You know,

00:01:29.359 --> 00:01:31.500
Google is the Internet for most people. It's

00:01:31.500 --> 00:01:34.840
the front door. Yeah. For a video feed to overtake

00:01:34.840 --> 00:01:37.730
search traffic. That was the signal that the

00:01:37.730 --> 00:01:40.230
entire hierarchy of information had just been

00:01:40.230 --> 00:01:42.909
turned on its head. Precisely. And the growth

00:01:42.909 --> 00:01:45.629
was just astronomical. It reached... A billion

00:01:45.629 --> 00:01:48.209
users faster than Facebook, faster than Instagram,

00:01:48.390 --> 00:01:51.489
faster than YouTube. A billion. By 2020, it had

00:01:51.489 --> 00:01:55.209
over 2 billion downloads and was operating in

00:01:55.209 --> 00:01:58.109
50 languages. This wasn't just a product. It

00:01:58.109 --> 00:02:01.390
became a global cultural engine. It effectively

00:02:01.390 --> 00:02:04.049
replaced the whole idea of the social graph from

00:02:04.049 --> 00:02:07.609
the 2010s with something entirely new. And that

00:02:07.609 --> 00:02:11.199
is exactly. the mission for this deep dive. We

00:02:11.199 --> 00:02:14.620
have a huge stack of sources here from technical

00:02:14.620 --> 00:02:17.599
breakdowns of the recommendation engine to supply

00:02:17.599 --> 00:02:20.680
chain reports on pharmaceutical shortages. Everything.

00:02:21.120 --> 00:02:23.199
We're not just going to rehash the political

00:02:23.199 --> 00:02:25.439
theater of the last month, though we obviously

00:02:25.439 --> 00:02:27.199
have to touch on it. We're going to try to unpack

00:02:27.199 --> 00:02:30.240
the black box of the app itself. We have to answer

00:02:30.240 --> 00:02:33.490
the fundamental question. What did this American

00:02:33.490 --> 00:02:36.610
consortium actually buy? Because, sure, the ownership

00:02:36.610 --> 00:02:39.289
changed, but the machinery is what matters. The

00:02:39.289 --> 00:02:42.360
engine. the engine how does that algorithm actually

00:02:42.360 --> 00:02:45.639
work how did it rewrite the rules for the music

00:02:45.639 --> 00:02:49.159
industry for fashion and i mean even how we buy

00:02:49.159 --> 00:02:52.340
medicine and crucially what does this new ownership

00:02:52.340 --> 00:02:55.599
mean for the future of that machinery we're already

00:02:55.599 --> 00:02:57.900
seeing the fallout we've got reports of censorship

00:02:57.900 --> 00:03:01.360
we've got the rise of competitor apps like upscrolled

00:03:01.360 --> 00:03:03.139
and this fracturing of the internet that people

00:03:03.139 --> 00:03:05.419
have been warning about for years it's happening

00:03:05.419 --> 00:03:07.740
it really is it's a lot to get through it is

00:03:07.740 --> 00:03:10.800
but we have to start with that ender, the secret

00:03:10.800 --> 00:03:13.180
sauce. Okay, let's get into it. Segment one,

00:03:13.360 --> 00:03:17.560
the algorithm. For years, the narrative around

00:03:17.560 --> 00:03:21.939
TikTok has been that it's just magic. You know,

00:03:21.979 --> 00:03:24.439
you open it and it just reads your mind. Right,

00:03:24.500 --> 00:03:27.060
the telepathic feed. But looking at the technical

00:03:27.060 --> 00:03:28.960
papers and some of the leaked audit documents

00:03:28.960 --> 00:03:32.219
we have, it's not magic. It's just extremely

00:03:32.219 --> 00:03:34.680
efficient collaborative filtering. It's math.

00:03:35.180 --> 00:03:38.759
Very specific, very aggressive math. And to understand

00:03:38.759 --> 00:03:41.080
why it worked so well, you really have to contrast

00:03:41.080 --> 00:03:42.819
it with the incumbent at the time, which was

00:03:42.819 --> 00:03:45.860
Facebook. Facebook and Instagram, to a lesser

00:03:45.860 --> 00:03:47.979
extent, were built on what we called the social

00:03:47.979 --> 00:03:50.340
graph. The social graph being, I see content

00:03:50.340 --> 00:03:52.819
because I know you, we're friends, or I follow

00:03:52.819 --> 00:03:55.379
you, so your posts appear in my feed. The personal

00:03:55.379 --> 00:03:57.960
connection is the filter. Exactly. The constraint

00:03:57.960 --> 00:04:00.379
on your feed was your social circle. TikTok just

00:04:00.379 --> 00:04:02.419
flipped that completely on its head. They built

00:04:02.419 --> 00:04:05.919
an interest graph. The For You page. FYP. And

00:04:05.919 --> 00:04:08.280
that page didn't care who you knew. It didn't

00:04:08.280 --> 00:04:10.840
care if you had zero friends or zero followers

00:04:10.840 --> 00:04:13.060
on the platform. Which is so different from anything

00:04:13.060 --> 00:04:16.860
else. It only cared about one thing. The content's

00:04:16.860 --> 00:04:19.199
performance against your specific behavioral

00:04:19.199 --> 00:04:21.839
vector. And this is where the cold start problem

00:04:21.839 --> 00:04:25.120
usually kills new apps, right? Yeah. But TikTok

00:04:25.120 --> 00:04:28.589
somehow solved it. They did. You could join post

00:04:28.589 --> 00:04:31.189
one video. And if the signals were right, you

00:04:31.189 --> 00:04:34.269
could hit a million views in 24 hours. The barrier

00:04:34.269 --> 00:04:37.209
to entry was effectively zero. So let's talk

00:04:37.209 --> 00:04:39.110
about those signals, because I think most people

00:04:39.110 --> 00:04:42.269
still overestimate the value of a like. A like

00:04:42.269 --> 00:04:45.339
is a very weak signal. It's almost. You might

00:04:45.339 --> 00:04:47.139
like a video just to be nice or because your

00:04:47.139 --> 00:04:48.600
friend sent it to you, but it doesn't mean you

00:04:48.600 --> 00:04:51.500
really loved it. The algorithm, according to

00:04:51.500 --> 00:04:53.360
the leaked engineering documents we're looking

00:04:53.360 --> 00:04:56.939
at, prioritized one thing above all else. Right.

00:04:57.040 --> 00:04:59.399
Retention. So weighted completion rate. That's

00:04:59.399 --> 00:05:01.420
the industry term for it. Yeah. Yeah. Did you

00:05:01.420 --> 00:05:03.519
watch the video from the very start to the very

00:05:03.519 --> 00:05:06.720
finish? That is a massive positive signal to

00:05:06.720 --> 00:05:08.339
the algorithm. And if you watched it more than

00:05:08.339 --> 00:05:11.060
once. Oh, that's exponential. A rewatch is gold.

00:05:11.629 --> 00:05:14.389
But it goes so much deeper than that. It tracks

00:05:14.389 --> 00:05:16.970
the velocity of your scroll. The speed of your

00:05:16.970 --> 00:05:20.149
thumb. Exactly. If you linger on a video about,

00:05:20.189 --> 00:05:22.350
say, sourdough baking for just three seconds

00:05:22.350 --> 00:05:24.649
before swiping versus swiping instantly on a

00:05:24.649 --> 00:05:27.069
dance video, the system records that hesitation.

00:05:27.129 --> 00:05:30.129
It learns. So it's analyzing your boredom threshold

00:05:30.129 --> 00:05:33.910
in real time. It's building a profile based on

00:05:33.910 --> 00:05:37.029
these tiny micro interactions that you, the user.

00:05:37.420 --> 00:05:39.240
aren't even aware you're performing. And this

00:05:39.240 --> 00:05:40.920
is where that Wall Street Journal investigation

00:05:40.920 --> 00:05:44.040
from 2021 was so pivotal. You remember that?

00:05:44.100 --> 00:05:45.899
Vaguely, yeah. They made a bunch of fake accounts.

00:05:46.040 --> 00:05:48.139
They created all these bot accounts with very

00:05:48.139 --> 00:05:50.259
specific interests. One was programmed to be

00:05:50.259 --> 00:05:52.459
into depression content. One was into politics

00:05:52.459 --> 00:05:55.480
and so on. And they just tracked how long it

00:05:55.480 --> 00:05:57.480
took the algorithm to figure them out. And it

00:05:57.480 --> 00:06:00.100
was terrifyingly fast, wasn't it? I think. The

00:06:00.100 --> 00:06:01.699
number that stuck with me was something like

00:06:01.699 --> 00:06:05.079
40 minutes. 40 minutes for the algorithm to completely

00:06:05.079 --> 00:06:07.579
zero in on a user's psychological vulnerabilities

00:06:07.579 --> 00:06:10.459
and build a feed around them. Wow. Because it

00:06:10.459 --> 00:06:13.160
excels at understanding emotion. It's not just

00:06:13.160 --> 00:06:15.800
tagging videos with cats or cars. It's tagging

00:06:15.800 --> 00:06:20.040
them with soothing, chaotic, depressive, aspirational.

00:06:20.120 --> 00:06:23.579
So it creates a feed that feels telepathic. Because

00:06:23.579 --> 00:06:26.220
it's literally mapping your emotional state in

00:06:26.220 --> 00:06:28.240
real time. And it's serving you dopamine hits

00:06:28.240 --> 00:06:29.939
that are perfectly tailored to your psychological

00:06:29.939 --> 00:06:31.899
profile at that exact moment you're scrolling.

00:06:32.019 --> 00:06:34.579
Which brings us to the algorithmic meritocracy

00:06:34.579 --> 00:06:37.560
argument. The idea that the best content just

00:06:37.560 --> 00:06:40.740
rises to the top and the feed is this perfect

00:06:40.740 --> 00:06:44.079
unbiased mirror of what you want to see. The

00:06:44.079 --> 00:06:47.790
great myth. But. We know that's not entirely

00:06:47.790 --> 00:06:49.790
true, is it? We have to talk about heating. The

00:06:49.790 --> 00:06:51.829
heating button. This is the thing that completely

00:06:51.829 --> 00:06:54.310
undermines that whole meritocracy myth. This

00:06:54.310 --> 00:06:58.410
story broke in Forbes back in January 2023. And

00:06:58.410 --> 00:06:59.769
I remember reading it and just thinking, OK,

00:06:59.850 --> 00:07:02.529
so the game is rigged. It was. The report revealed

00:07:02.529 --> 00:07:05.009
that employees at both TikTok and its parent

00:07:05.009 --> 00:07:08.899
company, ByteDance, had a manual tool. I'm literally

00:07:08.899 --> 00:07:11.740
a boost button that could just push certain videos

00:07:11.740 --> 00:07:14.160
into the viral stratosphere. So it wasn't just

00:07:14.160 --> 00:07:15.980
the algorithm picking the winners. It was humans,

00:07:16.120 --> 00:07:19.699
employees sitting at desks. The report estimated

00:07:19.699 --> 00:07:22.220
that at any given time, one to two percent of

00:07:22.220 --> 00:07:25.319
total daily video views were heated. And one

00:07:25.319 --> 00:07:27.740
or two percent sounds small. It sounds tiny.

00:07:27.860 --> 00:07:30.680
But on a platform with billions of views a day,

00:07:30.879 --> 00:07:34.079
that's tens of millions of artificially generated

00:07:34.079 --> 00:07:38.810
impressions every single day. And the criteria

00:07:38.810 --> 00:07:41.290
for heating a video, it wasn't just, oh, this

00:07:41.290 --> 00:07:44.170
is a good video. It was strategic. Highly strategic.

00:07:44.269 --> 00:07:46.709
I mean, originally the stated purpose was innocent

00:07:46.709 --> 00:07:48.970
enough. It was about diversity. Right. Because

00:07:48.970 --> 00:07:51.449
if the algorithm is too efficient, it just rabbit

00:07:51.449 --> 00:07:53.829
holes you into one type of content. Exactly.

00:07:53.829 --> 00:07:56.769
And you get bored. So they would heat videos

00:07:56.769 --> 00:07:59.889
from different genres to force them into your

00:07:59.889 --> 00:08:03.050
feed to sort of broadening your horizons and,

00:08:03.129 --> 00:08:05.269
you know, keep you on the app longer. But they

00:08:05.269 --> 00:08:07.490
also use it for business. Of course they did.

00:08:07.569 --> 00:08:10.250
They used it to court influencers and major brands.

00:08:10.370 --> 00:08:12.990
If you're trying to get, say, Taylor Swift to

00:08:12.990 --> 00:08:15.550
join the platform or the NBA or the FIFA World

00:08:15.550 --> 00:08:17.550
Cup. You can't leave their view counts up to

00:08:17.550 --> 00:08:20.029
chance. You can't. You heat their first few videos.

00:08:20.149 --> 00:08:21.769
They see these millions of views. They get that

00:08:21.769 --> 00:08:24.149
same dopamine hit we were talking about and suddenly

00:08:24.149 --> 00:08:25.949
they're hooked. It's basically a bait and switch.

00:08:26.189 --> 00:08:30.350
It's just corporate vanity metrics. But the Forbes

00:08:30.350 --> 00:08:32.429
report went darker than that, didn't it? It wasn't

00:08:32.429 --> 00:08:35.460
just corporate strategy. No. There were instances

00:08:35.460 --> 00:08:39.419
of employees using it for personal reasons. That's

00:08:39.419 --> 00:08:42.059
the real risk of any kind of manual intervention.

00:08:42.519 --> 00:08:45.259
So what, boosting their own videos? You had reports

00:08:45.259 --> 00:08:48.080
of employees heeding their own personal accounts

00:08:48.080 --> 00:08:50.799
or their spouse's account or their friend's accounts.

00:08:51.039 --> 00:08:53.299
It just shattered the illusion that if you just

00:08:53.299 --> 00:08:57.500
made good content, you could be famous. In reality,

00:08:57.620 --> 00:09:00.039
there was always this editorial layer, a human

00:09:00.039 --> 00:09:02.379
layer, deciding what the culture should see.

00:09:02.840 --> 00:09:04.480
Which brings us to a really important distinction,

00:09:04.720 --> 00:09:07.039
the one between what we saw here in the West

00:09:07.039 --> 00:09:10.139
and what exists in China. Because for a long

00:09:10.139 --> 00:09:11.919
time, there was this misconception that TikTok

00:09:11.919 --> 00:09:14.720
was the same app everywhere. Right. But we need

00:09:14.720 --> 00:09:16.960
to talk about Douyin. Douyin is the original.

00:09:17.100 --> 00:09:19.600
TikTok is effectively the export version of the

00:09:19.600 --> 00:09:21.899
app. And while they share the same underlying

00:09:21.899 --> 00:09:24.740
code base, the same DNA, they really evolved

00:09:24.740 --> 00:09:27.539
into two completely different species. Douyin

00:09:27.539 --> 00:09:30.460
is often described as a super app in China. It's

00:09:30.460 --> 00:09:33.059
not just for entertainment. No, not at all. It's

00:09:33.059 --> 00:09:35.460
integrated into the economy in a way that TikTok

00:09:35.460 --> 00:09:38.120
is only just now starting to attempt in the West.

00:09:38.700 --> 00:09:41.779
Years ago, Davin had these advanced features.

00:09:42.100 --> 00:09:45.580
Like what? Like you could search a video by a

00:09:45.580 --> 00:09:48.179
person's face to find other content they appeared

00:09:48.179 --> 00:09:51.120
in. You could book a hotel directly through a

00:09:51.120 --> 00:09:53.700
travel video. If you saw a restaurant review

00:09:53.700 --> 00:09:56.059
you liked, you could tap a button and reserve

00:09:56.059 --> 00:09:58.879
a table. Wow. It was an e -commerce engine first

00:09:58.879 --> 00:10:01.690
and an entertainment app second. But the biggest

00:10:01.690 --> 00:10:03.210
difference, the one that always came up in the

00:10:03.210 --> 00:10:05.710
political debates, was the regulation, the so

00:10:05.710 --> 00:10:08.289
-called teenage mode. This is what critics of

00:10:08.289 --> 00:10:11.450
ByteDance always, always pointed to. In China,

00:10:11.509 --> 00:10:13.549
the government enforced these incredibly strict

00:10:13.549 --> 00:10:16.169
controls on Douyin for minors. If you were under

00:10:16.169 --> 00:10:18.490
14 years old, you were automatically placed in

00:10:18.490 --> 00:10:20.409
teenage mode. And that wasn't just a suggestion

00:10:20.409 --> 00:10:22.409
you could click past. It was a hard lock. A hard

00:10:22.409 --> 00:10:25.190
lock. You were limited to 40 minutes a day. The

00:10:25.190 --> 00:10:27.490
app was completely inaccessible between 10 p

00:10:27.490 --> 00:10:31.129
.m. and 6 a .m. And most importantly, the content

00:10:31.129 --> 00:10:34.149
itself was filtered. So no viral dances or pranks.

00:10:34.600 --> 00:10:37.019
No. You were seeing science experiments, virtual

00:10:37.019 --> 00:10:40.019
museum tours, history lessons, patriotic content.

00:10:40.320 --> 00:10:42.679
It was all educational. It was essentially a

00:10:42.679 --> 00:10:45.700
state -run digital classroom. It was, whereas

00:10:45.700 --> 00:10:50.059
the export version, or TikTok, had none of those

00:10:50.059 --> 00:10:52.419
guardrails, at least not initially. It was the

00:10:52.419 --> 00:10:55.440
Wild West. And critics called it an opium for

00:10:55.440 --> 00:10:58.500
the masses strategy. The argument was keep the

00:10:58.500 --> 00:11:00.580
domestic population educated and disciplined

00:11:00.580 --> 00:11:03.320
and export the digital fentanyl to the West.

00:11:04.460 --> 00:11:06.620
Digital fentanyl, as you call it, didn't just

00:11:06.620 --> 00:11:09.159
keep people scrolling. It fundamentally reshaped

00:11:09.159 --> 00:11:11.700
entire industries. Absolutely. Which brings us

00:11:11.700 --> 00:11:14.299
perfectly to the next section, the cultural ripple

00:11:14.299 --> 00:11:16.960
effect. Let's start with music, because I think

00:11:16.960 --> 00:11:18.860
that's where the impact was most immediate and

00:11:18.860 --> 00:11:21.720
most obvious. TikTok became the de facto A &amp;R

00:11:21.720 --> 00:11:24.139
department for the entire global music industry.

00:11:24.320 --> 00:11:25.980
I mean, if you look at the Billboard Hot 100

00:11:25.980 --> 00:11:30.600
from, say, 2020 to 2025, there is a direct undeniable

00:11:30.600 --> 00:11:33.179
correlation between TikTok virality and chart

00:11:33.179 --> 00:11:36.580
success. It became the kingmaker. Lil Nas X is

00:11:36.580 --> 00:11:38.679
the classic case study, right? Old Town Road

00:11:38.679 --> 00:11:40.929
didn't break because of radio play. It broke

00:11:40.929 --> 00:11:43.809
because of memes. But I think the more interesting

00:11:43.809 --> 00:11:46.669
phenomenon really was the sleeper hit. The resurrection

00:11:46.669 --> 00:11:49.389
of the back catalog. This was both a nightmare

00:11:49.389 --> 00:11:52.230
and an absolute goldmine for the record labels.

00:11:52.429 --> 00:11:55.779
Right. Suddenly, Fleetwood Mac's Dreams is charting

00:11:55.779 --> 00:11:58.039
because of a guy on a skateboard drinking ocean

00:11:58.039 --> 00:12:01.919
spray. Or Kate Bush's Running Up That Hill hitting

00:12:01.919 --> 00:12:04.899
number one decades after it was released, all

00:12:04.899 --> 00:12:07.779
because it was in Stranger Things and then exploded

00:12:07.779 --> 00:12:11.200
in a TikTok trend. It completely broke the traditional

00:12:11.200 --> 00:12:14.019
music release cycle. Labels stopped focusing

00:12:14.019 --> 00:12:17.379
only on what was new and started like data mining

00:12:17.379 --> 00:12:20.080
their own archives for 15 second soundbites that

00:12:20.080 --> 00:12:22.740
might go viral. But that power dynamic created

00:12:22.740 --> 00:12:25.360
massive friction. We have to talk about the Universal

00:12:25.360 --> 00:12:28.200
Music Group standoff in early 2024. That was

00:12:28.200 --> 00:12:30.960
a huge moment. UMG, the biggest music company

00:12:30.960 --> 00:12:33.039
in the world, basically looked at the numbers

00:12:33.039 --> 00:12:35.220
and said, wait a minute, you're building a multi

00:12:35.220 --> 00:12:37.320
-billion dollar business on our back and you're

00:12:37.320 --> 00:12:39.320
paying us pennies for it. So they pulled the

00:12:39.320 --> 00:12:41.799
plug. They removed their entire catalog from

00:12:41.799 --> 00:12:44.559
the platform. Taylor Swift, Drake, The Weeknd

00:12:44.559 --> 00:12:46.620
just gone overnight. The app felt like a ghost

00:12:46.620 --> 00:12:48.200
town for a few months. I remember it was just

00:12:48.200 --> 00:12:51.100
like... royalty -free stock music and really

00:12:51.100 --> 00:12:53.220
bad acapella covers. It was a game of chicken.

00:12:53.679 --> 00:12:57.019
And UMG was also citing concerns about AI, which

00:12:57.019 --> 00:12:59.639
was a huge deal at the time. They were worried

00:12:59.639 --> 00:13:02.059
that TikTok was becoming a breeding ground for

00:13:02.059 --> 00:13:05.019
AI -generated tracks that were diluting the royalty

00:13:05.019 --> 00:13:08.340
pool for real human artists. But they came back

00:13:08.340 --> 00:13:11.230
eventually. They did. Taylor Swift's music returned

00:13:11.230 --> 00:13:14.929
in April 2024. But the whole episode proved that

00:13:14.929 --> 00:13:17.090
even the biggest artists in the world were now

00:13:17.090 --> 00:13:19.570
dependent on this one platform for discovery

00:13:19.570 --> 00:13:22.190
and distribution. But the ripple effect went

00:13:22.190 --> 00:13:24.929
way, way beyond entertainment. We saw it in global

00:13:24.929 --> 00:13:27.269
supply chains. Yeah. The Feta Pasta incident

00:13:27.269 --> 00:13:29.909
is kind of funny, but it's also a serious case

00:13:29.909 --> 00:13:32.480
study in logistics failure. It sounds ridiculous,

00:13:32.639 --> 00:13:35.299
but it was absolutely real. A single viral recipe

00:13:35.299 --> 00:13:38.240
video caused a global run on block -fedded cheese.

00:13:38.500 --> 00:13:41.320
A global run. Grocery stores in Finland and the

00:13:41.320 --> 00:13:43.279
U .S. and the U .K., they just ran out of stock

00:13:43.279 --> 00:13:46.039
for weeks. Dairy supply chains are not built

00:13:46.039 --> 00:13:48.299
for that kind of viral spike. You can't just,

00:13:48.379 --> 00:13:50.919
you know, make more cheese overnight. It takes

00:13:50.919 --> 00:13:53.240
months. That is the Amazon effect on steroids.

00:13:53.860 --> 00:13:56.519
But, I mean, while running out of cheese is annoying,

00:13:56.720 --> 00:13:58.500
running out of medicine is life -threatening.

00:13:58.659 --> 00:14:02.080
Yes. The Ozempic shortage is... probably the

00:14:02.080 --> 00:14:05.539
single most serious real -world consequence of

00:14:05.539 --> 00:14:07.860
the entire TikTok era. This is where the ethics

00:14:07.860 --> 00:14:11.200
get really, really murky. You had a drug, semaglutide,

00:14:11.419 --> 00:14:14.960
sold as Ozempic or Wegovy, that was designed

00:14:14.960 --> 00:14:17.679
for people with type 2 diabetes. Right. And it

00:14:17.679 --> 00:14:20.139
had a known side effect of weight loss. Suddenly,

00:14:20.200 --> 00:14:23.980
you have a hashtag, hashtag Ozempic journal with

00:14:23.980 --> 00:14:28.120
over 360 million views. You have influencers

00:14:28.120 --> 00:14:30.460
who were not diabetic at all documenting their

00:14:30.460 --> 00:14:32.539
dramatic weight loss. And they were getting prescriptions

00:14:32.539 --> 00:14:35.139
off -label. Exactly. The demand just skyrocketed

00:14:35.139 --> 00:14:37.269
overnight. Pharmacies could... not keep up. And

00:14:37.269 --> 00:14:39.450
the result was that actual diabetics people who

00:14:39.450 --> 00:14:41.649
needed that drug to regulate the insulin levels

00:14:41.649 --> 00:14:44.470
could not fill their prescription. We have reports

00:14:44.470 --> 00:14:46.750
from the Australian Medical Regulatory Agency

00:14:46.750 --> 00:14:49.850
explicitly linking the national shortage directly

00:14:49.850 --> 00:14:52.750
to these social media trends. It just highlights

00:14:52.750 --> 00:14:55.389
this massive failure in our information ecosystem.

00:14:56.190 --> 00:14:58.769
Medical advice was being dispensed by influencers

00:14:58.769 --> 00:15:02.149
with zero qualifications, and the algorithm didn't

00:15:02.149 --> 00:15:04.490
distinguish between a board -certified endocrinologist

00:15:04.490 --> 00:15:07.529
and a 19 -year -old beauty vlogger. No, if the

00:15:07.529 --> 00:15:09.649
vlogger had better lighting and a catchier hook,

00:15:09.870 --> 00:15:12.269
they got the views. And the market responded

00:15:12.269 --> 00:15:14.909
to the views. It forced pharmaceutical companies

00:15:14.909 --> 00:15:17.389
to completely rethink their manufacturing and

00:15:17.389 --> 00:15:19.950
supply chains, all because of an algorithmic

00:15:19.950 --> 00:15:23.000
trend. And we saw the same thing happen in publishing

00:15:23.000 --> 00:15:25.460
with BookTok. BookTok was fascinating because

00:15:25.460 --> 00:15:28.519
it was one of the few like purely positive supply

00:15:28.519 --> 00:15:31.600
chain shocks. It single handedly revived the

00:15:31.600 --> 00:15:34.200
physical book market, especially for young adults.

00:15:34.299 --> 00:15:36.519
It's true. You go into a Barnes &amp; Noble now.

00:15:36.879 --> 00:15:39.580
And there are whole tables dedicated to as seen

00:15:39.580 --> 00:15:42.360
on BookTok. It drove millions of copies of sales

00:15:42.360 --> 00:15:45.179
for authors like Colleen Hoover. It proved that

00:15:45.179 --> 00:15:47.299
long form reading wasn't dead. It just needed

00:15:47.299 --> 00:15:49.919
a new discovery engine. And fashion, too. The

00:15:49.919 --> 00:15:52.279
midsize movement was born there. That was a really

00:15:52.279 --> 00:15:55.120
important corrective movement. For years, fashion

00:15:55.120 --> 00:15:58.860
was so binary. You were either a size zero or

00:15:58.860 --> 00:16:01.990
you were plus size. TikTok creators started showing

00:16:01.990 --> 00:16:05.190
what a size 10 or 12 or 14 actually looked like.

00:16:05.330 --> 00:16:07.750
It forced brands to acknowledge the existence

00:16:07.750 --> 00:16:10.870
of the average body. But we also saw the darker

00:16:10.870 --> 00:16:13.750
side of all that consumerism with the de -influencing

00:16:13.750 --> 00:16:17.009
trend. Oh man. I love the concept of de -influencing

00:16:17.009 --> 00:16:18.970
at first. Don't buy this. It's overhyped. It's

00:16:18.970 --> 00:16:21.600
trash. It felt like a genuine rebellion. It started

00:16:21.600 --> 00:16:23.799
that way. It was a backlash against the constant,

00:16:23.879 --> 00:16:26.919
unrelenting pressure to consume, to buy, to haul.

00:16:27.279 --> 00:16:29.799
But look at how quickly it was co -opted. Almost

00:16:29.799 --> 00:16:31.779
immediately. Within weeks, the script completely

00:16:31.779 --> 00:16:34.419
flipped. It became, don't buy this expensive

00:16:34.419 --> 00:16:37.919
$40 mascara. Buy this $10 mascara instead. And

00:16:37.919 --> 00:16:39.799
by the way, here's my personal affiliate link

00:16:39.799 --> 00:16:42.460
for the $10 mascara. So it wasn't anti -consumption

00:16:42.460 --> 00:16:45.139
at all. It was just... alternative consumption.

00:16:45.460 --> 00:16:47.799
The system just absorbs the critique. You cannot

00:16:47.799 --> 00:16:50.860
exist on that platform without feeding the commerce

00:16:50.860 --> 00:16:53.539
engine. Even the very act of saying don't buy

00:16:53.539 --> 00:16:56.000
became a new kind of shopping recommendation.

00:16:56.320 --> 00:16:59.039
That feels like a perfect, if slightly depressing,

00:16:59.460 --> 00:17:03.159
segue to the darker psychological impacts. Because

00:17:03.159 --> 00:17:04.859
if the app can trick you into buying mascara,

00:17:05.140 --> 00:17:07.140
it can trick you into much, much worse. Segment

00:17:07.140 --> 00:17:10.759
three, the dark side. We have to talk about the

00:17:10.759 --> 00:17:13.519
rabbit hole. This is the core safety flaw of

00:17:13.720 --> 00:17:17.259
any interest -based graph. If a user shows even

00:17:17.259 --> 00:17:20.160
a momentary fleeting interest in something harmful,

00:17:20.420 --> 00:17:24.490
say extreme dieting, or self -harm imagery. The

00:17:24.490 --> 00:17:27.150
algorithm doesn't see harmful. It just sees engagement.

00:17:27.349 --> 00:17:29.029
It just sees engagement. Yeah. And it serves

00:17:29.029 --> 00:17:31.269
more of it. The feedback loop tightens incredibly

00:17:31.269 --> 00:17:34.150
quickly. We have ample documentation of teenagers

00:17:34.150 --> 00:17:37.430
being just bombarded with pro -anorexia content

00:17:37.430 --> 00:17:40.069
or suicide ideation within minutes of searching

00:17:40.069 --> 00:17:42.089
for something as innocent as weight loss tips.

00:17:42.349 --> 00:17:44.289
And the users were always one step ahead of the

00:17:44.289 --> 00:17:46.349
moderators. We saw the rise of what people called

00:17:46.349 --> 00:17:48.710
algo -speak. Right. Users would write suicide

00:17:48.710 --> 00:17:52.759
as unalive or porn as They created this whole

00:17:52.759 --> 00:17:55.900
new lexicon to bypass the automated content filters.

00:17:56.079 --> 00:17:57.779
But the human brain, especially the adolescent

00:17:57.779 --> 00:18:00.980
brain, decoded it instantly. The filter didn't

00:18:00.980 --> 00:18:03.579
work on people. Not at all. And the most tragic

00:18:03.579 --> 00:18:06.460
outcomes of this were the challenges, the blackout

00:18:06.460 --> 00:18:09.299
challenge in particular. This was horrific. It

00:18:09.299 --> 00:18:12.500
encouraged users, mostly kids, to choke themselves

00:18:12.500 --> 00:18:15.160
until they passed out, supposedly to achieve

00:18:15.160 --> 00:18:17.839
some kind of high. And it resulted in the deaths

00:18:17.839 --> 00:18:20.990
of multiple children globally. There were major

00:18:20.990 --> 00:18:24.109
lawsuits filed in Italy, in the U .S. In Venezuela,

00:18:24.250 --> 00:18:26.430
the Supreme Court actually ordered a crackdown

00:18:26.430 --> 00:18:29.869
and fined TikTok tens of millions of dollars

00:18:29.869 --> 00:18:32.930
in late 2024 after two children died there. It

00:18:32.930 --> 00:18:34.730
just raises this fundamental question of liability.

00:18:35.289 --> 00:18:38.490
Is the platform a publisher or is it just a mirror

00:18:38.490 --> 00:18:41.930
reflecting society? Well, the TikTok ticks phenomenon

00:18:41.930 --> 00:18:44.369
suggests it's much more than just a mirror. It

00:18:44.369 --> 00:18:46.970
can be a biological influence. This was a genuine

00:18:46.970 --> 00:18:49.220
medical mystery for a moment, wasn't it? Doctors

00:18:49.220 --> 00:18:51.559
in the US, the UK, Australia, they all started

00:18:51.559 --> 00:18:54.660
seeing this surge in teenage girls presenting

00:18:54.660 --> 00:18:57.480
with rapid onset tics. And they were very similar

00:18:57.480 --> 00:18:59.299
to Tourette's syndrome. They were shouting the

00:18:59.299 --> 00:19:01.940
same specific phrases, making the same specific

00:19:01.940 --> 00:19:03.819
physical movements. And they traced it all back

00:19:03.819 --> 00:19:06.630
to a few specific creators on the platform. They

00:19:06.630 --> 00:19:08.990
did. These patients were all watching popular

00:19:08.990 --> 00:19:11.910
creators who genuinely had Tourette's syndrome.

00:19:12.329 --> 00:19:15.109
And the doctors determined it was a form of mass

00:19:15.109 --> 00:19:17.509
psychogenic illness. So what does that mean?

00:19:17.710 --> 00:19:20.369
It means the teenagers' mirror neurons were firing

00:19:20.369 --> 00:19:22.650
so intensely from watching these videos over

00:19:22.650 --> 00:19:25.309
and over that they physically adopted the symptoms

00:19:25.309 --> 00:19:27.589
they were watching. It shows that the connection

00:19:27.589 --> 00:19:30.069
between the screen and the brain is far more

00:19:30.069 --> 00:19:32.250
porous, far more direct than we ever thought.

00:19:32.410 --> 00:19:35.299
Wow. And then on top of all that, there's the

00:19:35.299 --> 00:19:37.619
whole issue of child safety regarding predators.

00:19:37.839 --> 00:19:40.460
The fines really tell the story there. The UK

00:19:40.460 --> 00:19:43.339
fined them $12 .7 million for misusing children's

00:19:43.339 --> 00:19:47.420
data. The EU hit them with a massive $345 million

00:19:47.420 --> 00:19:51.079
fine. The U .S. had that early $5 .7 million

00:19:51.079 --> 00:19:53.460
fine. But the investigation into the lives feature

00:19:53.460 --> 00:19:55.720
was arguably the most disturbing part of it all.

00:19:55.819 --> 00:19:58.710
The refugee camp scams. Yeah. A BBC investigation

00:19:58.710 --> 00:20:01.210
found that families in Syrian refugee camps were

00:20:01.210 --> 00:20:03.750
going on TikTok live and begging for digital

00:20:03.750 --> 00:20:05.849
gifts from viewers. And the viewers thought they

00:20:05.849 --> 00:20:07.690
were donating directly to help these desperate

00:20:07.690 --> 00:20:10.670
families. Right. But when the money was actually

00:20:10.670 --> 00:20:13.569
processed, TikTok was taking up to 70 percent,

00:20:13.630 --> 00:20:17.359
seven zero of the proceeds as a commission. So

00:20:17.359 --> 00:20:20.039
they were profiting off digital begging in a

00:20:20.039 --> 00:20:22.579
literal war zone. It was an absolute low point

00:20:22.579 --> 00:20:25.460
for the company. And then you had the sexual

00:20:25.460 --> 00:20:28.019
exploitation scandals in Kenya that were exposed

00:20:28.019 --> 00:20:30.900
in 2025, where the live streaming feature was

00:20:30.900 --> 00:20:33.059
being used essentially as cam girl sites for

00:20:33.059 --> 00:20:35.000
minors. It just highlighted that at the scale

00:20:35.000 --> 00:20:37.599
of a billion users. Moderation is basically a

00:20:37.599 --> 00:20:39.859
myth. You cannot police that much live video

00:20:39.859 --> 00:20:42.599
effectively. It's impossible. And all of these

00:20:42.599 --> 00:20:44.779
safety failures, they provided the ammunition

00:20:44.779 --> 00:20:47.119
for the politicians. But the gun was already

00:20:47.119 --> 00:20:50.480
loaded by geopolitics. Exactly. Segment four,

00:20:50.619 --> 00:20:52.950
the road to the sale. You simply cannot separate

00:20:52.950 --> 00:20:55.450
the story of this app from the broader anxieties

00:20:55.450 --> 00:20:57.609
of the U .S.-China relationship. It all comes

00:20:57.609 --> 00:21:00.190
down to the fact that TikTok was owned by ByteDance,

00:21:00.309 --> 00:21:03.210
a company headquartered in Beijing. In the current

00:21:03.210 --> 00:21:05.950
geopolitical climate, that was just a non -starter

00:21:05.950 --> 00:21:08.549
for Washington. It centers on two key things,

00:21:08.789 --> 00:21:11.130
the golden share and the national intelligence

00:21:11.130 --> 00:21:13.509
law of 2017. Let's break those down. OK, the

00:21:13.509 --> 00:21:16.170
golden share. The golden share is a specific

00:21:16.170 --> 00:21:19.500
Chinese corporate mechanism. The Chinese government,

00:21:19.740 --> 00:21:22.059
usually through state -backed funds, takes a

00:21:22.059 --> 00:21:25.240
1 % stake in the key domestic subsidiaries of

00:21:25.240 --> 00:21:27.920
big tech companies like ByteDance. And 1 % sounds

00:21:27.920 --> 00:21:30.859
tiny, but it's not about the money. It's not

00:21:30.859 --> 00:21:32.400
about the money at all. It comes with special

00:21:32.400 --> 00:21:34.720
rights, usually a board seat and, crucially,

00:21:34.839 --> 00:21:38.380
veto power over major content decisions. So the

00:21:38.380 --> 00:21:40.339
argument that ByteDance is just a private company

00:21:40.339 --> 00:21:42.680
doesn't hold water in the same way it would for

00:21:42.680 --> 00:21:45.039
a U .S. company. Exactly. And then you layer

00:21:45.039 --> 00:21:47.140
on top of that the 2017 national intelligence

00:21:47.140 --> 00:21:50.740
law. This law explicitly states that any Chinese

00:21:50.740 --> 00:21:53.140
organization or citizen shall, and I'm quoting

00:21:53.140 --> 00:21:59.000
here, And the interpretation of that in the West

00:21:59.000 --> 00:22:01.680
was pretty clear. If Beijing asks for the data,

00:22:01.839 --> 00:22:04.420
ByteDance cannot legally say no. That was the

00:22:04.420 --> 00:22:07.680
core fear. And to be fair, ByteDance tried to

00:22:07.680 --> 00:22:10.539
fight this perception for years. They spent billions

00:22:10.539 --> 00:22:13.119
on lobbying. They swore up and down that the

00:22:13.119 --> 00:22:15.769
data was firewalled. But then came the leaks.

00:22:15.930 --> 00:22:18.250
The BuzzFeed news tapes, that quote, everything

00:22:18.250 --> 00:22:20.990
is seen in China. That audio leaked in 2022.

00:22:21.630 --> 00:22:24.670
It was from internal meetings where U .S.-based

00:22:24.670 --> 00:22:27.170
employees were admitting that the master admin

00:22:27.170 --> 00:22:30.130
accessed the keys to the kingdom still resided

00:22:30.130 --> 00:22:32.890
in Beijing. It completely undercut their entire

00:22:32.890 --> 00:22:35.190
public narrative. And then came the spying scandal

00:22:35.190 --> 00:22:38.609
in December 2022. This wasn't theoretical anymore.

00:22:38.730 --> 00:22:41.700
No, this was active surveillance. ByteDance's

00:22:41.700 --> 00:22:43.900
own internal audit team was trying to find the

00:22:43.900 --> 00:22:47.019
source of leaks to the press. To do that, they

00:22:47.019 --> 00:22:49.319
accessed the personal TikTok data of journalists

00:22:49.319 --> 00:22:52.359
from Forbes and the Financial Times. Oh. They

00:22:52.359 --> 00:22:54.339
tracked their IP addresses and their location

00:22:54.339 --> 00:22:56.500
data to see if they were ever in the same physical

00:22:56.500 --> 00:22:59.339
proximity as ByteDance employees who they suspected

00:22:59.339 --> 00:23:02.180
were the leakers. They used the app to hunt down

00:23:02.180 --> 00:23:04.859
whistleblowers. It was the smoking gun. It proved

00:23:04.859 --> 00:23:07.220
that the tool could be used for espionage. Because

00:23:07.220 --> 00:23:09.819
it was being used for espionage, albeit for corporate

00:23:09.819 --> 00:23:12.490
security. rather than state intelligence. But

00:23:12.490 --> 00:23:14.869
for lawmakers, the damage was done. So TikTok

00:23:14.869 --> 00:23:17.990
had to pivot hard. They announced Project Texas.

00:23:18.029 --> 00:23:20.829
This was their Hail Mary pass. The plan was to

00:23:20.829 --> 00:23:24.769
create a new subsidiary, TikTokUSDS, and store

00:23:24.769 --> 00:23:28.170
all U .S. user data on servers that were physically

00:23:28.170 --> 00:23:32.210
in the U .S. and owned by Oracle. The American

00:23:32.210 --> 00:23:34.710
tech giant. And Oracle was supposed to audit

00:23:34.710 --> 00:23:36.390
the code, right? That was part of the deal. That

00:23:36.390 --> 00:23:38.849
was the big pitch. Oracle would review every

00:23:38.849 --> 00:23:40.849
single line of code that went into the app to

00:23:40.849 --> 00:23:43.230
ensure there were no back doors to China. They

00:23:43.230 --> 00:23:45.029
did a similar thing in Europe called Project

00:23:45.029 --> 00:23:47.250
Clover. So why didn't it work? Why didn't Congress

00:23:47.250 --> 00:23:50.470
buy it? Because of the very nature of modern

00:23:50.470 --> 00:23:53.750
software. The TikTok algorithm isn't a static

00:23:53.750 --> 00:23:55.930
document you can read once and approve. It's

00:23:55.930 --> 00:23:58.890
updated constantly. It's machine learning. It

00:23:58.890 --> 00:24:01.670
evolves by the second. Critics, including a lot

00:24:01.670 --> 00:24:03.569
of technical experts, pointed out that you cannot

00:24:03.569 --> 00:24:05.869
effectively audit an algorithm that changes every

00:24:05.869 --> 00:24:08.329
second unless Oracle controlled the source of

00:24:08.329 --> 00:24:10.210
the code, not just reviewed it. The whole thing

00:24:10.210 --> 00:24:11.890
was kind of meaningless. So the trust was just

00:24:11.890 --> 00:24:15.890
gone. And that brings us to segment five, the

00:24:15.890 --> 00:24:19.710
showdown. The years 2024 to 2026. What a roller

00:24:19.710 --> 00:24:23.069
coaster. It really was. It started with the legislative

00:24:23.069 --> 00:24:26.579
hammer. In March of 2024, the U .S. House passed

00:24:26.579 --> 00:24:30.400
H .R. 7521. President Biden signed it into law

00:24:30.400 --> 00:24:32.660
in April. The clock officially started ticking.

00:24:32.880 --> 00:24:36.839
The deadline was January 19th, 2025. Divest or

00:24:36.839 --> 00:24:39.240
be banned. And people forget how close we came

00:24:39.240 --> 00:24:42.759
to the cliff. On January 18, 2025, the day before

00:24:42.759 --> 00:24:45.460
that deadline, TikTok actually went dark in the

00:24:45.460 --> 00:24:48.880
U .S. for about 12 hours. The near -death experience.

00:24:49.079 --> 00:24:51.039
I remember it so clearly. My feed just had that

00:24:51.039 --> 00:24:53.140
spinning circle. No connection. It was total

00:24:53.140 --> 00:24:55.619
chaos. For millions of creators and small businesses,

00:24:55.900 --> 00:24:58.660
their entire livelihood just vanished overnight.

00:24:59.180 --> 00:25:01.640
But then the political winds shifted. Donald

00:25:01.640 --> 00:25:04.519
Trump had just won the 2024 election. And despite

00:25:04.519 --> 00:25:06.559
having tried to ban it himself during his first

00:25:06.559 --> 00:25:09.700
term, he did a complete 180. He flipped. He realized

00:25:09.700 --> 00:25:11.680
the immense political capital of the platform.

00:25:11.859 --> 00:25:13.359
He didn't want to be the guy who killed the most

00:25:13.359 --> 00:25:16.180
popular app in America on day one of his inauguration.

00:25:16.200 --> 00:25:18.470
So he delayed it. He issued an executive order

00:25:18.470 --> 00:25:20.930
delaying the enforcement of the ban. He essentially

00:25:20.930 --> 00:25:23.390
bought them time to find a buyer. But he couldn't

00:25:23.390 --> 00:25:25.950
kill the law entirely. The legal pressure remained.

00:25:26.549 --> 00:25:28.849
ByteDance had to sell. And the rumors were just

00:25:28.849 --> 00:25:31.329
wild for a while. For a moment, it really looked

00:25:31.329 --> 00:25:33.410
like Elon Musk might buy it and try to merge

00:25:33.410 --> 00:25:36.609
it with X. I remember that. Then Steve Mnuchin

00:25:36.609 --> 00:25:38.849
was putting together a group. But the deal that

00:25:38.849 --> 00:25:42.630
finally closed on January 22, 2026 was much more

00:25:42.630 --> 00:25:45.009
establishment. Let's look at the new owners.

00:25:45.450 --> 00:25:47.650
Who actually holds the keys to the kingdom now?

00:25:47.849 --> 00:25:50.309
It's officially called the USDS joint venture.

00:25:50.690 --> 00:25:54.490
Now, ByteDance actually got to keep 19 .9%. They

00:25:54.490 --> 00:25:56.029
aren't fully out of the picture, which was a

00:25:56.029 --> 00:25:58.029
big concession. But the controlling stake? The

00:25:58.029 --> 00:26:01.630
controlling 80 .1 % is a mix of Oracle, the private

00:26:01.630 --> 00:26:04.630
equity firm Silverlake, an AI company called

00:26:04.630 --> 00:26:07.990
MGX, and a consortium led by Michael Dell. So

00:26:07.990 --> 00:26:10.630
it's big tech and private equity. It is as corporate

00:26:10.630 --> 00:26:13.630
America as you can possibly get. Adam Presser

00:26:13.630 --> 00:26:16.390
is the new CEO. Sho Chu is still on the board

00:26:16.390 --> 00:26:18.950
as a director. But the power has clearly shifted.

00:26:19.170 --> 00:26:21.970
The servers are fully under U .S. control. But

00:26:21.970 --> 00:26:23.809
there is one line in that deal announcement that

00:26:23.809 --> 00:26:26.210
caught everyone's attention. The source notes

00:26:26.210 --> 00:26:28.049
we have mentioned that the algorithm will be

00:26:28.049 --> 00:26:30.589
adjusted over time to favor American topics.

00:26:30.849 --> 00:26:33.549
That is the phrase that pays. Favor American

00:26:33.549 --> 00:26:36.349
topics. What does that mean technically? Does

00:26:36.349 --> 00:26:39.230
it mean boosting NFL clips over soccer clips?

00:26:39.470 --> 00:26:42.130
Or does it mean suppressing foreign news? Or

00:26:42.130 --> 00:26:44.250
does it mean something more ideological? And

00:26:44.250 --> 00:26:46.410
we're already finding out. Which brings us to

00:26:46.410 --> 00:26:49.049
right now, segment six, the new controversy.

00:26:49.970 --> 00:26:52.750
Because we all thought the sale would save the

00:26:52.750 --> 00:26:54.849
app, but it seems like we've just traded one

00:26:54.849 --> 00:26:57.450
set of problems for another. We basically traded

00:26:57.450 --> 00:27:00.730
China anxiety for corporate censorship anxiety.

00:27:01.210 --> 00:27:04.009
The ink on the deal is barely dry, and we are

00:27:04.009 --> 00:27:06.890
already seeing widespread reports of videos disappearing.

00:27:07.109 --> 00:27:09.779
Specifically political content. Yes. There are

00:27:09.779 --> 00:27:12.059
very serious allegations. And Gavin Newsom, the

00:27:12.059 --> 00:27:13.980
governor of California, is already launching

00:27:13.980 --> 00:27:16.039
an investigation into this, that the new algorithm

00:27:16.039 --> 00:27:18.660
is suppressing anti -Trump content, criticism

00:27:18.660 --> 00:27:22.680
of ICE, and very curiously, any mentions of Jeffrey

00:27:22.680 --> 00:27:24.940
Epstein. So the fear for years was that the Chinese

00:27:24.940 --> 00:27:27.839
government would use the app to manipulate American

00:27:27.839 --> 00:27:30.720
public opinion. Now, the accusation is that the

00:27:30.720 --> 00:27:33.380
new American owners are using the app to manipulate

00:27:33.380 --> 00:27:35.859
American public opinion. It really raises the

00:27:35.859 --> 00:27:39.269
question. Is a platform this powerful ever neutral?

00:27:39.589 --> 00:27:43.190
And the answer seems to be a resounding no. Whoever

00:27:43.190 --> 00:27:46.230
owns the server owns the reality. And the users

00:27:46.230 --> 00:27:48.329
are reacting. We are seeing a mass migration.

00:27:48.670 --> 00:27:50.569
It's the SplinterNet happening in real time.

00:27:50.710 --> 00:27:52.730
Just look at the App Store charts this week.

00:27:52.809 --> 00:27:55.089
An app called Upscrolled is sitting at number

00:27:55.089 --> 00:27:58.609
two. It's a direct TikTok clone and its whole

00:27:58.609 --> 00:28:01.309
marketing pitch is unfiltered feeds. But the

00:28:01.309 --> 00:28:04.150
irony, and I love this detail. is the rise of

00:28:04.150 --> 00:28:07.289
RedNote. This is the ultimate plot twist. RedNote

00:28:07.289 --> 00:28:10.029
is the international name for Xiaohongshu, which

00:28:10.029 --> 00:28:11.930
is another Chinese app. It's sort of like a mix

00:28:11.930 --> 00:28:14.150
of Instagram and Reddit. So let me get this straight.

00:28:14.329 --> 00:28:17.289
Users are fleeing the newly Americanized TikTok.

00:28:17.849 --> 00:28:20.690
because of censorship fears to join an actual

00:28:20.690 --> 00:28:22.630
Chinese app. Because they perceive it as being

00:28:22.630 --> 00:28:25.309
more authentic and crucially outside the reach

00:28:25.309 --> 00:28:27.809
of U .S. political tinkering. It just shows you

00:28:27.809 --> 00:28:30.069
how completely fractured the landscape has become.

00:28:30.369 --> 00:28:32.009
We used to have something resembling a global

00:28:32.009 --> 00:28:34.849
village. Everyone everywhere saw the same memes.

00:28:35.009 --> 00:28:39.089
Now. Now we have walls, digital walls. TikTok

00:28:39.089 --> 00:28:42.769
is banned in India, Nepal, Albania, banned it

00:28:42.769 --> 00:28:45.750
in 2025. It's restricted on government devices

00:28:45.750 --> 00:28:48.880
all across the EU and the U .S. And now the U

00:28:48.880 --> 00:28:52.140
.S. version is its own distinct, curated garden.

00:28:52.299 --> 00:28:55.319
The global era of the Internet is over. We are

00:28:55.319 --> 00:28:57.440
now entering the era of the sovereign Internet,

00:28:57.559 --> 00:29:00.440
where every region has its own truth, its own

00:29:00.440 --> 00:29:03.299
trends, and its own algorithmic bias baked right

00:29:03.299 --> 00:29:05.720
in. That is a heavy realization. We started this

00:29:05.720 --> 00:29:07.740
deep drive looking at a simple dancing app. And

00:29:07.740 --> 00:29:10.200
we ended up staring at the fragmentation of global

00:29:10.200 --> 00:29:12.779
reality itself. That's the story of the 2020s,

00:29:12.779 --> 00:29:15.269
I guess. The tools we built to connect us ended

00:29:15.269 --> 00:29:17.089
up just sorting us into our own little silos.

00:29:17.950 --> 00:29:19.910
So here's the final thought for you listening.

00:29:20.150 --> 00:29:22.710
The algorithm has new masters. It is now officially

00:29:22.710 --> 00:29:25.329
tuned, in their own words, to favor American

00:29:25.329 --> 00:29:27.730
topics. But does that actually make you feel

00:29:27.730 --> 00:29:30.289
safer? Or does it just make you wonder what you

00:29:30.289 --> 00:29:32.210
aren't seeing anymore? The black box is still

00:29:32.210 --> 00:29:34.109
a black box. We just painted it red, white, and

00:29:34.109 --> 00:29:36.890
blue. And if upscrolled is the new refuge from

00:29:36.890 --> 00:29:39.430
all this, does the cycle just start all over

00:29:39.430 --> 00:29:41.849
again? History would suggest it does. On that

00:29:41.849 --> 00:29:44.920
note. We're going to log off. Stay curious. Check

00:29:44.920 --> 00:29:48.480
your sources. And maybe go read a physical book.

00:29:48.559 --> 00:29:49.099
See you next time.
