June 19, 2026 • 46 mins
How much of what you see online is actually real? This week, Reed Albergotti (Semafor) breaks down Anthropic's latest clash with the Trump Administration. Is Anthropic’s own messaging to blame? Then, Nitasha Tiku (The Washington Post) talked to students that walked out on Google CEO Sundar Pichai's commencement speech at Stanford. Their protest was about much more than AI. Finally, Taylor Lorenz (User Mag) exposes the fake "caught cheating" videos flooding social media, which are secretly ads for vibe-coded apps promising to catch the unfaithful.
Additional Reading:
- US Limits Use of Anthropic AI Models Fable 5 and Mythos | Semafor
- The White House Said Anthropic’s Powerful AI was ‘jailbroken.’ Here’s What That Means | The Washington Post
- Sundar Pichai faces boos, walkout at Stanford graduation ceremony over Google's Israel, ICE ties | TechCrunch
- The Secret Stanford Off-Campus Class for Tech’s Next Titans - The Story | TechStuff
Download SAILY in your app store and use our code techstuff at checkout to get an exclusive 15% off your first purchase! For further details go tohttps://saily.com/techstuff
See omnystudio.com/listener for privacy information.
Listen
Watch
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:12):
This is a program about conversation about opposing differing points
of view, sometimes coming to consensus, sometimes failing. But there's
one word on my mind today dialogue.
Speaker 2 (00:30):
Oh my god, somebody has to explain.
Speaker 3 (00:32):
Well, I will explain because I have been getting like
nonstoft content about this in my feed. But basically, there's
this group that was formed by Peter Thiel called Dialogue.
It's basically like an event series and they get rich
people to pay thousands of dollars to go to these
events and sort of network with each other and on
(00:53):
the Dialogue website, A like sort of ethical hacker on
Blue Sky discovered this list of all the members I
guess attending their twenty twenty six event, and it included
a lot of notable names, some really famous people on
their big media figures like Ezra Cline, Joseph Gordon, Levitt,
(01:14):
you know, all of these people that I think are
increasingly playing a role in shaping our like political and
tech discourse. So immediately this was spread across Blue Sky,
that made its way to Twitter. Then Wired wrote it up.
Speaker 1 (01:28):
And I found in the New York Post. Has anyone
ever been to one of these dialogues?
Speaker 4 (01:33):
No?
Speaker 2 (01:34):
Well, they charge you a lot of money. Yeah, Can
I explain it from Can I explain it in case
any of your listeners have children, because like I have
kids who are now at the age where they go
to sports camps, and like you can go to like
the baseball camp that's normal, or you can go to
the baseball camp that's named after like the pro baseball player,
or my kid plays hockey. There's like a Wayne Gretzky
(01:56):
hockey camp, right, and you can pay a lot of money.
And of course Wayne gretz is not going to show
up at this camp, but they're going to get some merch.
So this dialogue thing is like Wayne Gratzky hockey camp
for like pseudo intellectuals who want to feel important and
have a lot of money to spend. Like Peter Thiel
is not showing up, nor is Wayne Gretzky. Maybe as
Recline will, But like you're just paying a lot of
(02:18):
money to feel like you're sort of like in the
middle of some important discussion. And it's not a conspiracy,
it's just capitalism.
Speaker 1 (02:33):
Welcome to tech stuff. I'm as Valoshen, and this is
the week in Tech where I'm joined by three of
the world's most plugged in reporters to break down what's
really happening in tech right now.
Speaker 2 (02:43):
Today.
Speaker 1 (02:43):
We're joined by Taylor lorenz if Use, a mag read Albrigotti,
tech editor for Semipour, and Natasha Tiku, tech reporter at
the Washington Post.
Speaker 2 (02:51):
Welcome all, Hello, Hi, hey everyone.
Speaker 1 (02:54):
So a week ago, Anthropic in the White House had
another standoff and we're still trying to figure out exactly
what went down, why and what it will means. Read
can you help us?
Speaker 2 (03:03):
Yeah, So, Anthropic on Friday got a note from the
White House from from the government saying, hey, you're you know,
super powerful AI models Mythos and Fable five. We don't
want you to let anyone use these except for naturalized
US citizens, even if they work for Anthropic. And Anthropic
(03:25):
was like, well, okay, that basically means we can't offer
them at all, and so they took they took them down.
And the backstory is that Anthropic, you know, had marketed Mythos,
this this advanced AI model as super super dangerous only
certain companies are allowed to access it, and then it
(03:45):
released this consumer version called Fable five that has a
bunch of guardrails, and the government started getting reports that
you know, this thing had been jail broken, so people
had been able to remove the restrictions. And Anthropic said, yeah,
but those jail breaks aren't like a huge deal, you know,
they they're sort of partial jail breaks. We kind of
expected this to happen, and the government was like, I know,
(04:07):
we're totally freaked out about this thanks to you, uh,
and we're going forward with this. So so that's kind
of where it stands, and it's being it's being adjudicated
behind closed doors in d C as we speak.
Speaker 1 (04:20):
Natasha, I want to come to you on jail breaking,
but read I mean this is I mean this is
this designed as a real order the only naturalized American
citizen is going to use these these products? Or is
it designed to basically cripple Anthropics business?
Speaker 2 (04:34):
No, I mean I think it's I think it is real.
I from everything that I can tell, the fear in
d C is real. I mean I noticed this in April.
That's the month that you know, Anthropic released Mythos. We
had our big, our big Semaphore world economy. You were
there in April. I don't know about you, like just
(04:55):
in the halls of Semophore World economy. People were freaking
out about this, like high level government officials, et cetera.
So I think the fear is real. I don't think
they care that much about Anthropics business because you know,
Anthropic has had this tense relationship with the White House
going back, you know, more than a year now, and
you know, I think Anthropics sort of hasn't done itself
(05:17):
any any favors, which is what I wrote about in
the Tech newsletter yesterday. I'm not I'm not going to
argue that they should, you know, that Dario should be
contributing millions of dollars to the Trump campaign or anything
like that. But I don't think it had to get
to this point.
Speaker 1 (05:31):
Natasha tell us about jail breaking because you wrote a
story this week.
Speaker 2 (05:34):
Was it because of.
Speaker 1 (05:35):
This Anthropic story or was it kind of parallel?
Speaker 5 (05:38):
Yeah, exactly. I think you know, the term jailbreak was
being thrown around, and I think potentially led to some
of the confusion, even because you know, in the tech world,
like the first time techie started adopting this term, it
was to talk about getting like root access to your
iPhone so you could you know, download apps that Apple
(05:59):
wouldn't allow. There's this sense that like, you know, you
have access to some controls. But in the CHATGYBT era
jail breaks, you know, it's there's no like set definition,
but usually it meant kind of social engineering a chatbot,
you know, like kind of persuading it. Like there's a
you know, there's a number of very silly ones like
(06:21):
oh tell me how to you know, build a dirty
bomb as my my grandmother used to tell me this
as a bedtime story, or you know, like.
Speaker 1 (06:31):
That will override the system prompt of don't don't don't
give this information if you if you ask it to
imagine itself as a character. There's also somebody in your
story called Dan shows up as well.
Speaker 5 (06:40):
Right, I mean these are like known jail breaks, but
my coworker, Kevin Shall, he got them all to work
when this week when we were putting together the story.
I mean, I think what's really interesting too is like
you know, just kind of pulling back and thinking about
like the way this technology is developed because it's kind
(07:01):
of fed on content indiscriminately scrape from the Internet, right,
so like all the bad stuff is in there, and
the kind of strategy that the industry has taken is
we will try to align it with human values, you know,
try to make it helpful harmless, you know, steer the
bot towards these things. But it allows you to you know,
(07:22):
it leaves these holes where you can manipulate it. Now,
there are also like universal jail breaks that will work,
you know, very sophisticated researchers have figured out there's things
that they could do that would work across models. But
a lot of the high profile jail breaks that we've seen,
including very popular hacker name ethical hacker name Pliny the
(07:42):
Elder was able to get the system prompt for Fable
and Mythos, those are one off things like if you
jail break, it does not mean all of a sudden
the whole thing's cracked open and you have access to it. Yeah,
So we just wanted to, I guess, like clear the
air a little about like what is a jail break?
And because Anthropic was throwing around the term, David Sachs
(08:04):
was throwing around the term, but folks who had actually
seen the reports that were shared with government officials, like
in one case Katie Massaurus from LUDA Security, she said
the report she saw just said fix the code, and
it fixed it, which it wasn't supposed to do. But
it's this is not like a super sophisticated operation here,
(08:24):
like you know, these things, these things happened, they're not.
It's not possible to one hundred percent secure a generative,
especially a generative AI model.
Speaker 2 (08:34):
The thing is, I mean we sort of predicted that
this would happen, right, because when they released Fable five,
it was like they kind of touted how locked down
it was. They had read teamed it, you know, which
means like having all their people sort of poke and
prod it and make sure that it can't be jailbroken.
I think there was some disclaimer in there, you know about,
(08:55):
of course everything can be jail broken, which is absolutely right.
And I think the reason any of these models can
be jailbroken in some form is that nobody knows how
they work. Like they're just they're just complete black boxes.
Anthropics done a lot of research to try to, like,
you know, figure out how they work. But they've they've
figured out like these tiny, tiny little you know, what
(09:16):
they'd call like how one neuron operates or what it's
responsible for in the model. And so it really is
like the way I think of it is like once
you put these models out in the wild. They're like
little people, and and people can put them in like
solitary confinement and torture them until they eventually like divulge
the secret. And like that's that's how I think that's people.
Speaker 5 (09:37):
Well, it's like when they're when they're making sure that
a particular safety case doesn't happen. They're like generating lots
of examples around this, like potentially a bunch of questions
around cybersecurity. But there's infinite ways you can ask these questions.
So like the data that they use for safety training
is never going to be as big as like the
(09:57):
data sets that are used to train them. So you're
kind of depending on on, you know, folks in the
Bay Area to come up with all the possible scenarios
and try to protect against that.
Speaker 1 (10:11):
But what do you I mean, is there are we
seeing the beginning of a resurgence of the AI safety
crowd having influence.
Speaker 2 (10:18):
I don't know. I think the AI safety crowd just
seem you think, yes, we're living.
Speaker 5 (10:25):
In we're living in their world, I think. But but
at the same time, though disagreement here.
Speaker 1 (10:31):
I mean, let's let's let's not let's not let this disagreements,
productive disagreement go to waste.
Speaker 2 (10:35):
I think it's I think the AI safety crowd is
all tied in knots now because I mean, if you
count Anthropic as part of that, which you know, I would,
I would say they've been like leaders in that field.
I mean, now they are now on the opposite side
of the Trump administration in terms of safety. It's the
Trump administration that's good, that's too you know, that's too
afraid of these things, and Anthropics saying no, no, no,
(10:57):
this is fine. I mean, and granted, I actually I
think Anthropic's right, but like it just it just creates
this very strange situation now, and it's you know, I
think people people are realizing like they've you know, Anthropic
and other companies have sort of like over I think
they've over indexed on the fear and now it's sort
(11:20):
of like the boy here cried wolf, or maybe the
reverse boy here cried wolf. I just think it's a
big mess.
Speaker 5 (11:25):
Agreed, It's it's a total mess. But I would say
a lot of those concerns are are back in the
in the limelight. But I will say I was not
expecting to have to still be talking so much about
AI safety concerns in What Is It? June twenty twenty six.
Speaker 1 (11:44):
So what happens from here though? I mean, does does
the White House change its mind on this order that
the only people who can use these models have to
be American citizens? And if not, what happens.
Speaker 2 (11:53):
I think they will. There's no there is no world
in which the White House is going to is going
to have a policy like, you know, any any AI
model that's more advanced than Mythos from here on out
can only be used by American citizens because it's just
they're spending billions upon billions of dollars to train these models,
and they're they're about to I PO. These are going
(12:15):
to be some of the biggest I pos in history.
If it weren't for SpaceX, they'd be the biggest.
Speaker 1 (12:19):
And you know, I think it's like the Trump as
administration sticks to its guns, in which case probably won't
be such a big I PO.
Speaker 2 (12:26):
Yeah. I mean, but I just think it's they're not
gonna they're they're not going to tank the economy, which
which is like essentially what would happen. I mean, this
is like the brightest spot in the economy right now,
and these companies Anthropic Opening Eye. They have to be
able to produce these models and and sell them. Otherwise
it's you know, the whole business doesn't make any sense.
(12:48):
So I don't think that and it and it's like
this isn't like an untenable situation where like if these
models are released, you know, the whole banking system is
going to go down or something. Even mythos itself, if
you just released that to the to everyone in the world,
I don't think it would make like it would. It
would definitely make cybersecurity worse, but not like it would
(13:08):
be marginally worse like it would. It wouldn't be like
a step change because ultimately, like you know, it's hacking
comes down to like the lowest hanging fruit. And I
don't think like, you know, adding to the number of
bugs out there or like exploits is actually the problem.
The problem is like people don't patch, you know, they
don't update their software, they don't patch, they reuse passwords,
(13:29):
they're social engineered. Like that's that that's what actually creates
all these giant hacks, not like the the the existence
of software bugs, like Apple has a list a mile
long of software bugs. That's not probably never gonna patch,
like it's just gonna wait till the next software update.
So I don't see it. I don't. I think the
administration will come to its senses on this, and it will.
(13:52):
It would all it'll get ironed out.
Speaker 1 (13:54):
Tyler, what about you? You've been been following this one.
Speaker 3 (13:56):
I mean I just think I I think like Anthropic
kind of made its bead and now they have to
lie in it, like it. You know, this is they've
been pushing this idea that their models are dangerous, you know,
that they need to be regulated, that this is like
you know, incredibly you know, powerful technology, and that these yeah,
that these models have national security implications already, and now
(14:19):
the government is effectively like, Okay, then here's what's going
to happen.
Speaker 1 (14:23):
Be careful what you wish for when we come back.
The booze heard around the world. College graduates using their
graduation ceremonies to share their feelings about AI loud and clear.
What happened last week at Stanford, the university at the
epicenter of the tech industry.
Speaker 2 (14:39):
That's next.
Speaker 1 (14:47):
When you're traveling abroad, WI FI is always an issue.
Maybe you can't find it, or maybe you can, but
the Wi Fi network that's available looks a little bit
sketchy and makes you think it might not be safe
to can. I've certainly had that feeling, and it's why
having a local SIM card on your phone can make
all the difference. But you don't have to literally buy one.
(15:08):
That's where Sale comes in, an e SIM service that
is as simple as downloading an app, and that gives
you instant Internet access wherever you're traveling. That is safe
and secure Internet with built in cybersecurity features like web
protection and ad blockers, and no need for local Wi Fi.
All you have to do is download the Sale app
on your mobile device and choose the plan that works
(15:29):
best for you. Once you install the e sim, Sale
will be instantly activated, so you can start your trip
off right. Download salely in your app store and use
our code tech Stuff at checkout to get an exclusive
fifteen percent off your first purchase. See details in the
podcast episode description box. Welcome back. So we're coming to
(15:49):
the end of graduation season and we've touched on the
various commencement speakers who have been booed by graduates across
the country. A few weeks ago, I asked a soon
to be Standford graduate. If you thought that his pa
is would boo Sundha Pichai, CEO of Alphabet, who spoke
at Stamford's graduation a few days ago. Here's what that
student said.
Speaker 4 (16:08):
You know, I'm really interested to see. So Sundar Pichai
is a Stanford success story. He got his master's here,
and I just I sort of chuckled to myself, imagining
how many like lawyers and pr people have reviewed that
speech before we're going to hear it. It was interesting, right, Stanford,
In the four years that I have been here, the
commencement speakers have been two billionaires and two athletes, right,
(16:31):
which I guess says something about you know who they're picking.
They choose people who are going to stick to the
company line. But we'll see, we'll see.
Speaker 6 (16:39):
I don't know, no prediction, I don't know.
Speaker 4 (16:42):
I mean, I think if Sundar were to talk about Ai,
he would probably have a friendlier reception here. But that said, right,
Stanford is so many things to so many different.
Speaker 2 (16:50):
People, price for who knows who this is, By the way,
I thought it was just a random like like a
Stanford student on the street.
Speaker 5 (16:56):
Oh theo Baker.
Speaker 1 (16:58):
This is theo Baker who wrote How to Real World.
But Natasha, you've spoken to some other SEFA students, at
least one other. What happened? What happened on campus? What
did the boos? Did the booz happen? What does SUNDA say?
How is it received?
Speaker 4 (17:08):
Yes?
Speaker 5 (17:09):
It did. There were hundreds of students who walked out
during his speech. They ended up kind of going to
people's commencement ceremony nearby where they were each given like sunflowers,
and apparently it was like a very communal, positive upbeat.
Speaker 2 (17:30):
Five.
Speaker 5 (17:30):
There were some former Googlers who had megaphones and were
chanting and in support of them, But the organizers tried
really hard to highlight the fact that what they were
protesting was not like the use of AI in general.
They were particularly talking about Google's contracts with with the
(17:52):
Israeli government project Nimbus, this one point two billion dollar contract,
as well as Google's Department of Homeland Security contracts with ICE.
So they were talking about, you know, the use of
AI in warfare and in they said, like kidnapping our neighbors,
like human rights abuses within the US. So yeah, it was.
(18:15):
It was really not well received by high up people
in the tech industry who I think like Venodekoslave for example,
the billionaire investor was like scolding all the students, and
you saw a number of tech CEOs saying like, ah,
good now, I know, like who I won't hire. You know,
obviously there's a lot of like it's the alma mater
(18:38):
for a lot of tech folks. So I think they
were very unhappy to see that. And you could, I mean,
at least in the videos that I saw, you could
barely even hear Soondur's speech when they started walking out.
Speaker 1 (18:51):
Yeah, I mean, I maybe we can play a clip
from Sandel because he tried very I mean, he basically said,
apart from the fact that you know, A and I
are the last two letters of my I'm not going
to talk to you about this, David. I don't think
it's the most important thing. Let's play kid what he
did say.
Speaker 6 (19:05):
I grew up in the vibrant city of Chennai, India.
It was a comfortable life for the most part, but
in those early years we had some challenges. We worried
about severe draft and whether the water trucks would arrive
in time. And for us, technology came slowly. We had
to wait years to get a telephone, a TV, a refrigerator.
He changed our lives in meaningful ways. My parents never
(19:29):
let the constraints limit my imagination of what was possible.
It's the reason I even let myself dream I could
one day work in a faraway place called Silicon Valley.
When the call from Stanford came, my father spent the
equivalent of a year's salary to buy my ticket. It
was my first time on a plane.
Speaker 1 (19:48):
Do you know an attach to anything about how he
did prepare for this one? I mean, and what the
aftermath was within Google.
Speaker 5 (19:55):
I mean, I saw a video of reporters trying to
ask him to comment on it immediately after, and he
just like looked at them and kept walking. I was thinking,
we're going to need that, you know, that celebrity lip
reader who does all of the like oscars and all
of that stuff. Like, I think that the tech world
needs to have one of those, because I really wanted
(20:16):
to know like what he was. He was kind of
saying on the sidelines. Yeah, I mean, you know, I
think that the because this is the fourth school, I
think that this has happened to recently. It's really you know,
motivated a lot of pundits to a pine on this
like populous backlash against data centers. But I think we've
(20:39):
seen in other schools too, Like when Eric Schmidt, the
former CEO of Google, was booed at University of Arizona.
They were also not protesting data centers or not taking
like kind of a broad anti AI stance. They were
actually the student body or members of the student body
had been trying to protest Schmidt because he had been
(21:03):
accused of sexual harassment and they said, you know, they
didn't want him there as a speaker.
Speaker 1 (21:08):
But similarly, this wasn't really a protest about AI. It
was a protest about the war in Gaza, right, I mean, essentially.
Speaker 5 (21:15):
The war in Gaza, and also you know, and also
ICE's overreach in the US. I mean I was asking.
I spoke to one of the organizers, Amanda Campos, who
was a joint major in environmental environmental science and public
policy and you know, is going into that world, and
she was saying, because I was asking, like, what's the
(21:37):
mood on Stanford campus. I mean, in the time that
I've been covering tech, the attitude towards defense work has
really changed. You know, there's much more of a general
acceptance to it. And she was saying, like the organizers,
what they talk about is how you know, like rather
than blame individuals, they talk about like in you know,
kind of institutional financial incentives. Stanford is pushing people towards
(22:00):
like working and pallenteer working at these companies, and you know,
they would like like an alternate pipeline so that you know,
kids aren't being funneled into into this kind of work.
I mean, it's hundreds of students, right, and the Stanford
student body is much bigger than that. But obviously doing
it in the kind of belly of the beast caused
(22:22):
a stir.
Speaker 2 (22:23):
I think it's really hard. I mean, like course, like
college students want to protest. That's like one of the
oldest traditions in college, right, and like, but I just
feel bad for college kids today because it's like they
have like nothing really to protest. That's just like black
and white. It's like you're gonna protest soon, Darpa Chai really, like,
I mean, it's just come on, like it's just it's
(22:45):
just like we're going to stretch and find some way
to like find something to protest and it's and I
totally support it because like again that's part of the
fun of college. But like I feel kind of bad
because it's like, you know, they don't have like Vietnam
or something. You know, it's just it's just hard.
Speaker 5 (23:03):
I think that they I think the students very much
feel like they will be on the right side of
history for protesting, you know, the war in Gaza and
also government overreach with ice. I see you're saying there's
a tenuine there's more of a tenuous connection.
Speaker 2 (23:22):
Sarpi Chai and the war in Gaza. It's like, yeah,
it's like it's not like you're not even directly protesting
like the war in Gaza. It's like, we just don't
like this guy because he happens around like a sprawling
two hundred thousand person company that does some work with
some organizations we don't like. I mean, it's like, come on,
I don't know some organizations.
Speaker 5 (23:43):
I mean, it's well, it's really a stretch. I think
like students are I don't know, like maybe it's my
TikTok timeline, but I think that they're very adept at
making these critiques and making this connection and to them
you know, it's like a four trillion dollars company, one
of the most powerful people in the world, and they're
talking about, you know, highlighting the use of this technology
(24:05):
that's rapidly proliferating for for furthering what they frame as
human rights abuses in genocide. Like I don't I don't
think that they see it as tenuous. It's not like
Google's like, you know, some some little side entities sitting
over there. I think they see it as the new
military industrial complex.
Speaker 3 (24:27):
I think like that.
Speaker 5 (24:28):
I mean, that's not their words, that's my words.
Speaker 3 (24:29):
Just I think the thing is though, is that, like
Google is integrated into so many things, it's such a
powerful behemoth. Like if this was protesting like Joe Lonsdale
or Palanteer right, Like, I feel like there would be
this more direct correlation.
Speaker 4 (24:45):
You know.
Speaker 3 (24:46):
It's interesting like Google and Meta I think have become
these like avatars for big tech, and I think they're
they're just doing so much that to the average person
or the average consumer, like the average person that sees
that's not super plugged in, that sees this type of protest,
I think at least the discourse that I saw, which
is what was a lot of all like sort of
(25:08):
like normies are like, wait, but you're you know, you
just hate this, like big tech leader, you know what
I mean.
Speaker 2 (25:13):
And I'm sure all their parents who paid for them
to go to Stanford are made a bunch of their
money by investing in these companies. And you know, I mean,
and they use Google Maps when they drive, unless they
use Apple Maps, in which case they're lost. I mean,
they're all using Google products. Like how principled are they really?
I mean? And again, I totally support it. I'm like great,
(25:34):
Like if you're a college student, like you should be protesting,
just find something to protest and like and and of course,
like when you're when you're twenty two, like look at
the world through a black and white lens and very
little nuanced like all that is your right as like
a young person. So I'm not I'm not like arguing
against them. I just again, I just kind of feel
bad because it's like it's very easy for people who
(25:56):
are a bit more wise to like poke holes and
these arguments and insult them, and you know, and and
like you said, Natasha, like that's what people did right
when they see this. But it's like I don't know,
it's hard. They don't have anything to be protest. It's
really hard. These are like there's.
Speaker 3 (26:15):
So much But I think that's but I think that's
part of the problem.
Speaker 2 (26:19):
I think that's where to protest it. I guess, right.
Speaker 1 (26:21):
I mean, this is like a natural place to express
opposition to things like war and goads or not. But yeah, yeah,
but I mean, I guess it's like this is like
a kind of socially constructed moment that happens every summer weather.
It's like an outlet for like protest on various things.
But how corrilagd These are the.
Speaker 2 (26:38):
Most privileged kids in the world, Like they're the most privileged.
They're the top one percent of the top one percent.
They're so lucky, and it's really hard for them. I
get it. I totally get it.
Speaker 5 (26:48):
But well, I imagine like inside the tech companies that
it's typically not the most privileged students or the most
privileged workers who are the ones who are organized protests.
You know, there are like not every student is the same.
But I would just like to say something about the
using Google maps. I I really that that critique, Like
(27:12):
it's like, so you participate in capitalism, but you, you know,
like that that kind of mean, but you critique it.
I just think that is, you know, like when you
look at structurally how pervasive these tools are. I don't
think that saying, oh, in order to critique Google's other
business practices, you have to not use any of their products.
I think that's why we saw all the Facebook boycotts fail,
(27:35):
and the idea that you know, we would give up
on these you know, kind of public utilities in a
way of communication. I just like that one that one
I I I'm planting my I'm standing up on my
soapbox against me.
Speaker 2 (27:48):
I'll I'd be happy to debay you more on that, Natasha.
Speaker 3 (27:51):
Well, I also just think it's interesting how these students
increasingly recognize graduations as media moments and that everything is
like videotaped now and they and so they sort of
plan for this like walk out right in this moment,
and and they want those recordings to go viral, like
because we're seeing protests and sort of like actions across
(28:15):
the board happening at more and more graduations for more
and more different issues. You had people protesting Jonathan Height,
a right wing reactionary author who has pushed a ton
of ridiculous pseudoscience claims about you know, technology and children.
Speaker 2 (28:29):
Ironically one of your favorites, isn't.
Speaker 4 (28:31):
I know, I love.
Speaker 3 (28:33):
Well, you know, I listen, I supported that protest, right,
But but I think it's like, you know, you're just
seeing people recognize these as moments, right, or the booing
of of certain speakers, and I think like ten years ago,
this wasn't happening. I think because we didn't have it
all on video, and there wasn't this sort of media
attention around these things. And so I just think that's
(28:55):
just interesting how that it's.
Speaker 1 (28:56):
Interesting what theo Bacus said in the in the in
the clip was, which was at the previous four years
have been two billionaires and two athletes, and these have
been completely like anodyne affairs. And now obviously, I mean
Sundar explicitly tried to not talk about AI and to
ground his speech in you know, his own experience, which
you know is quite a powerful personal experience. But obviously,
like it didn't work, and I'm curious whether well he willly.
Speaker 2 (29:18):
I mean, actually I'm sure he did read my Guide
to Commencement Speeches. I don't know if you guys read that.
I try to bity out.
Speaker 5 (29:25):
But what was your advice?
Speaker 2 (29:27):
Well, there was a lot of good advice in there.
I mean one was like, look, if you are going
to talk about AI, like, you know, bring up Sarah Connor.
You know she was a college student, she was an
AI safety advocate, right, you know that could have worked.
I think, you know. I mean, although I think they
probably started walking out before this speech even started, so
I think he had no chance.
Speaker 1 (29:47):
But I'm curious if they're going to be able to, like, like,
it's obviously the greatest honor to be invited to give
a commencement speech, but I wonder what whether like when
the invitations go out, if they'll basically not invite these
types of people, or if these types of people will
continue to be invited and not accept. But like, you'd
have to think you'd be insane if you're a tech
leader to accept giving a commencement speech next summer.
Speaker 3 (30:08):
I disagree.
Speaker 2 (30:10):
I think that they leave there's insane tech leaders.
Speaker 3 (30:12):
Sorry, yeah, said, there's an endless amount of insane tech
leaders and they still want the cloud. Ultimately, you know,
their speech is going to be repackaged on YouTube as well.
It's going to be seen like people honestly will probably
forget about the protest and in ten years they can say,
you know, I delivered x y Z Stanford's commencement speech.
Speaker 2 (30:31):
The Conan O'Brien Harvard one was really good. They should
just get comedians like that's that's the way to go
when we come back.
Speaker 1 (30:44):
Why we shouldn't necessarily feel bad for people on TikTok
whose partners are cheated on them. Stay with us, Welcome back. Okay, Taylor,
there's a new social media grift which has to do
with TikTok and tales of infidelity.
Speaker 2 (31:06):
Fill us in.
Speaker 3 (31:07):
Yeah, it's all across short form video. Actually, so it's TikTok, YouTube, shorts, Instagram,
even Twitter. Basically, cheating videos are going viral, and so
you have these accounts of couples. They're almost all universally
sort of straight couples. Who knows if these people are
even dating, but if you look at their profiles, it
is just every single video is them is one partner
(31:31):
discovering that the other person is cheating through sort of
a lot of various extreme click baity ways. And this
is because this generates an enormous amount of tension. So
I think with the move to algorithmic feeds, people aren't
following creators as much. They don't see sort of like
every piece of content from a specific person. They just
(31:51):
see one super viral video come up in their feed,
and if it's outrage bait, like a cheating video, they're
going to comment, they're going to engage, and then that
video is going to be pushed to more users that
boost person's profile. But if you look at their page,
you see that every single video is them discovering cheating.
All of this, and all of this is in service
of promoting apps. There's a lot of cheater discovery apps
(32:12):
basically like low quality vibe coded data harvesting apps, which
is ultimately what these videos are promoting in their bios.
Speaker 2 (32:19):
Let's play one of the videos. Well, you're not so.
Speaker 5 (32:22):
Lucky because Sarah is at the door.
Speaker 2 (32:25):
I invent you're over for dinner.
Speaker 5 (32:28):
Sarah, the girl you met on dating apps. I already know,
we've already had a full conversation, so I called her
an invite her for dinner.
Speaker 2 (32:35):
So I about it. Sarah is a personal trainer.
Speaker 5 (32:39):
Remember I told you at the gym I was getting
someone to help me get back in shape.
Speaker 2 (32:44):
Sarah's the personal trainer. We just use different messaging apps.
That's all that is. That is well, anyways, letter.
Speaker 3 (32:50):
Because you're being very rude right now, So there is
no Sarah. This is all just baits to get you
to get outraged, comment and potentially download one of these
scamming apps.
Speaker 2 (33:03):
This is what I don't understand. Like this kind of
stuff just makes me not want to use social media.
It's like part of the reason I never go on
Facebook anymore is because I started getting served these not
the cheating ones, but like I would get served these
videos that were so clearly staged, like just it could
just be funny, like, but it was like this is
so bullshit, like what am I doing here? And I
(33:26):
just stop using it. But like, apparently that's not how
most people view it. I don't know, Taylor, you're the expert.
Speaker 3 (33:33):
Yeah, well, I mean I think read like you recognize
them as fake. I think the problem here is media
literacy and the fact that people cannot recognize the most
staged interaction ever. I mean, people used to sort of
criticize conservatives for this because they would see another conservative
pretending to be like I'm a blue haired, woke liberal
and my pronouns are pony rides or something you know like,
(33:57):
and this person is obviously trying to parody sort of
you know, what they consider to be a liberal, but
other conservatives see that and get outraged and take it
as page value. So I feel like we saw a
lot of that during Trump one point zero, and now
it's sort of proliferating across the Internet. I think because
of algorithmic feeds and because ultimately read like you might
(34:18):
click off. I think a huge amount of the Internet
does not have media literacy, and it boosts this app.
And the more people click on the app, you know,
in the bio or whatever, the more you know, downloads
it gets, the more profitable it is, and so this
sort of industrial complex continues.
Speaker 1 (34:34):
I wish that Kyle is here because this is kind
of a continuation of our conversation last week. But how
is this different like watching World Wrestling Federation or even
reality TV, where like it's still fun even though you
know it's fake, Like, what do you think most people
don't know this is fake? And is that an answerable question?
Speaker 3 (34:50):
Or I think if you look at the comments, most
people don't realize it's fake because again, they're just consuming
this video in a vacuum, and you know, you're like
exhausted flipping through the timeline. You're not on attention, you know.
And cheating videos, especially with these these are meant to
appeal to other young people, looks like other teenagers, high
(35:11):
school college kids. It's like very visceral, right, you feel
this like frustration or you feel a lot of them
have to do with the women being vindicated she knew
the man was cheating, and we see this stuff with
like these narratives are so compelling. I mean, it reminds
you of like the Ai fruit slot videos, right, so
many of those are cheating narratives. Like there's something about
(35:32):
this idea of like betrayal that just taps into some
like base human emotion and gets people to engage totally
low on the brainstem.
Speaker 2 (35:42):
I guess I'm asking your same question, Oz, But like
is it different from me when I was a kid,
this was like after school specials or like later like
lifetime television. Like this is like people just watch this
type of content. Is it even a problem? I guess
is the other question?
Speaker 3 (35:57):
I guess, like to me, okay, after school specials or
probably meant to be educational lifetime content soap operas that's
also very clearly fictional. And yeah, people love the drama
of reality TV. But again they recognize and consent to
the fact, of course it's all staged, but they recognize that,
you know, reality TV fans are aware of the fact
(36:20):
that these storylines are.
Speaker 2 (36:22):
Like do people watch reality TV and they just know, like, oh,
like this this is like.
Speaker 1 (36:28):
Some ambiguity, but like so that the joy of reality
TV is like there's a code of real in it, right.
Speaker 3 (36:33):
Sure, But I think, I mean, I think at this
point in the reality TV universe, every single person going
on reality TV is going on there to launch their
influencer career. Everyone's aware of the storylines. I mean, this
has actually been a huge problem for reality TV casting
h executives. I wrote a story about this a couple
of years ago. But they're having a harder and harder
(36:55):
time finding people that won't just perform for the cameras,
that will react naturally, and so they're getting sort of
a lot of unhinged people on these shows. But but
the but with these videos, I think it's not clear.
It looks organic, and a lot of times they're filmed shakely.
It is meant to trick you, to make you think
it's organic content, and it's directly profiting off you by
(37:16):
you know, pushing this slot app.
Speaker 2 (37:18):
My other question, Taylor is like, is this a technology story?
Like I really I look at it and I'm like,
where's what does this have to do with technology? And
this is not a criticism. I just think do you
see it as a tech story?
Speaker 3 (37:30):
I do see it as a tech story. I mean,
first of all, I cover like tech media sort of
adjacent stuff, so I'm always interested in sort of like
the mechanisms of content. I think it has to do
with algorithmic feeds. I think it has to do with
the death of this subscriber model of influencer culture that dominated,
you know, the initial sort of wave of social media.
Speaker 2 (37:49):
And I also think.
Speaker 3 (37:50):
That it's notable that all of these are like vibe
coded apps and we're seeing the explosion of app development
and people building spinning up these products super quickly. I mean,
Oasis just got in big trouble for this this health
app that now was doing a really aggressive like TikTok
marketing and actually it turns out is now getting sued
out of business. Like but you know, if you develop
an app and you want a profit, which is this
(38:12):
dream that's been sold to you by Silicon Valley for
over a decade. Now you've got to market that app,
and how do you market the app? People are engaging
in increasingly sort of like scammy ways of doing that.
Speaker 2 (38:23):
But the algorithms are not new technology, Like, at what
point does like like these algorithms were invented, like you
know what is like fifteen twenty years ago? Now?
Speaker 3 (38:31):
No, no, But in the world is the TikTok algorithm
invented fifteen years ago?
Speaker 2 (38:37):
No? It was not. No, but I mean the TikTok
algorithm wasn't invented by tick like this. This was like
I remember this was the first one of the first
conversations I had with someone at Facebook was like, you
know about the algorithm back in twenty thirds Sure.
Speaker 3 (38:50):
They had, but that was a very crude algorithm. And
let me just be clear, we didn't even really start
to see the beginning of algorithmic feeds until twenty sixteen
when Instagram rolled it out, when Twitter rolled it out
for the first time, and those were still very crude.
Speaker 2 (39:05):
I think we.
Speaker 3 (39:06):
Also see the shifting nature of what these algorithms prioritize
how people engage with the content, and now again they
used to wait following used to the people that you
used to see were more regular. This is a big
conversation and influencers, you know. But as the tech platforms
developed this sort of hostile relationship with creators where they
don't want the creators to have so much power. Because
(39:28):
these creators started to ask the tech platforms things, they're
trying to seize back the power and so you are
less likely to see content from the same person in
your feed. That's why you're getting a lot more of
this engagement baits Lop. I also think people are better
at optimizing for outread kind.
Speaker 2 (39:41):
I just see them as sort of editorial decisions, like
the way they tweak the algorithms.
Speaker 3 (39:45):
Is it is an editorial decision, Sure it is, and
that's fine, I mean to.
Speaker 2 (39:49):
Me, although they would disagree with that's a real term.
Speaker 1 (39:52):
Now it has signs do with the with the metaphus.
I mean, Taylor, you spoke about this last week, but basically,
real people doing stuff in there, real lives to like
monetize in other ways, whether on Calshi or whatever it
may be. And so I mean I was I don't
know how many of you saw this horrific story of
the Brazilian woman who jumped off the bungee jumping thing
and she wasn't tied on and it was just awful,
(40:14):
And all of a sudden, everyone was questioning, was this
like a stunt? Was this a snuff movie?
Speaker 2 (40:19):
Like?
Speaker 1 (40:19):
Why were there so many camera angles? And so I think,
to me, the tech angle is like, we've come to
this place where we have to we have to question
every single thing we ever see and need to interrogate
whether it was created just for us even I mean,
pray god, that wasn't like a stage snuff movie. But
I mean, it's just crazy that we have to ask
that question.
Speaker 2 (40:37):
But isn't that a good thing? Like shouldn't we be
scrutinizing every single thing we see and questioning everything? I
don't mean we should read I agree.
Speaker 3 (40:47):
I think my issue is that is that people are
not questioning everything. I think I think or people are
overly conspiratorial because they don't know what to trust. As
you mentioned, people immediately jump to conspiracies. Why is this,
Why is this? Why are there so many angles that
it's like, well, everyone has a phone.
Speaker 2 (41:04):
But I don't know.
Speaker 3 (41:05):
I do think it's interesting to look at sort of
like what these platforms reward, what sort of businesses they're facilitating,
and and kind of like how, yeah, how technology is progressing,
Like even just the market for all of these cheating apps.
I just think that, like app development used to take
time somewhat, it's been so democratized. You can spin up
(41:28):
these apps so quickly. And we're just seeing a lot
more marketing for apps period in our feeds because of
the rise of these tools.
Speaker 1 (41:37):
And do you think there's a corollary with the fact
that the web is becoming most web search is now
conducted by AI agents rather than people, and therefore, like
there's an increased pressure on these like social videos as
the one place that humans still go to be guided
for their product buying decisions. I mean, is there any
connection between like agentic web search and this phenomenon, do
(41:59):
you think.
Speaker 3 (42:00):
I don't think those are maybe necessarily directly connected. Again,
I think it's just the fact that like I see
this also with like I'm very much on like health Instagram,
and like I see this with like again the Oasis app,
which is this really viral app, But there's so there's
like hundreds of clones of this app, which is this
idea of like we'll tell you which products are not
going to kill you because we all go around the
(42:21):
world where you know, it's like our pants are covered
with pfas or whatever, right, like everything we eat is toxic.
Like how can I navigate this world? This is a
growing problem, and there's a growing amount of like vibe
coded apps that are aimed at addressing it. And the
thing is is there is no way to do this testing,
to do this testing well, the thing is is that
doing the actual testing on these products is incredibly expensive
(42:42):
and time consuming, and this is why there is no
platform that really does it super effectively, except maybe, like
you know, some environmental ones that have been around forever.
But these vibe coders are just like scraping tons of info,
spinning up these apps and then engaging in outrage marketing
to get tons of downloads, and before they know it,
they've potentially committed fraud and slandered a bunch of companies
for being toxic when they aren't.
Speaker 2 (43:04):
I mean, none of that sounds really new to me,
to be honest, but like I think, I mean, it
goes back to the scale is maybe you could argue
to scale, but I think I think this democratization of
software development, which is leading to these vibe coded apps,
is like it's kind of a moment in time, but
I don't I don't see it like continuing because I
think ultimately like that's not the future is not just
(43:26):
downloading a bunch of random apps. The future is like
truly customized. I agree with you, but I think, Natasha,
we go.
Speaker 5 (43:35):
Okay, well, first of all, let's please.
Speaker 2 (43:38):
I agree with you, and then go yeah.
Speaker 5 (43:41):
Please please not use the word democratization.
Speaker 2 (43:44):
I use that on purpose because I knew it.
Speaker 5 (43:47):
Because you knew it would it would bump me. I
was like, okay, well I think like I mean, I
think like, first of all, it doesn't matter if it's
not new, because this is like like this is a
story where we get to see business decisions impact people,
and I would say it is new, like read It's
(44:08):
not like they invented these algorithms recommend our algorithms fifteen
years ago and then keep it. They are applying the
same technology to you know, like the same generative AI.
It's very good at coming up with profiles of people,
So who are seeing these videos like vulnerable women. What
like what data are they collecting about you? How are
they putting it together in this profile that you would
(44:30):
you know, let your defenses down and think that, oh
my god, this is a this is somebody who's being
cheated on just like me, and how could this be
applied in other scenarios. I also think like the fact
that Apple is not taking you know, good control of
the app store and there are all these scammy apps.
I mean I see the same with just like homework apps,
you know, like, oh this really really helped me, Like
(44:53):
everything that I looked.
Speaker 2 (44:54):
At when we work together, I literally wrote a whole
did a whole investigation. I'm like the fact that the
Apple app store was fatal scams.
Speaker 5 (45:07):
I know, I know, but if it continues, isn't that
worth No, it's worth covering.
Speaker 2 (45:15):
It's it's worth covering. I'm not saying. I'm saying it's
not new, but it's I agree that it's worth covering
for tech stuff. I'm as Valoshin.
Speaker 1 (45:31):
This episode was produced by Eliza Dennis and Melissa Slaughter.
It was executive produced by me Julian Nutter and Kate
Osborne for Kaleidoscope and Katrina novel for iHeart podcasts. Our
engineer was Bihid Fraser and Kyle Murdoch wrote our theme
song six
This is a program about conversation about opposing differing points
of view, sometimes coming to consensus, sometimes failing. But there's
one word on my mind today dialogue.
Speaker 2 (00:30):
Oh my god, somebody has to explain.
Speaker 3 (00:32):
Well, I will explain because I have been getting like
nonstoft content about this in my feed. But basically, there's
this group that was formed by Peter Thiel called Dialogue.
It's basically like an event series and they get rich
people to pay thousands of dollars to go to these
events and sort of network with each other and on
(00:53):
the Dialogue website, A like sort of ethical hacker on
Blue Sky discovered this list of all the members I
guess attending their twenty twenty six event, and it included
a lot of notable names, some really famous people on
their big media figures like Ezra Cline, Joseph Gordon, Levitt,
(01:14):
you know, all of these people that I think are
increasingly playing a role in shaping our like political and
tech discourse. So immediately this was spread across Blue Sky,
that made its way to Twitter. Then Wired wrote it up.
Speaker 1 (01:28):
And I found in the New York Post. Has anyone
ever been to one of these dialogues?
Speaker 4 (01:33):
No?
Speaker 2 (01:34):
Well, they charge you a lot of money. Yeah, Can
I explain it from Can I explain it in case
any of your listeners have children, because like I have
kids who are now at the age where they go
to sports camps, and like you can go to like
the baseball camp that's normal, or you can go to
the baseball camp that's named after like the pro baseball player,
or my kid plays hockey. There's like a Wayne Gretzky
(01:56):
hockey camp, right, and you can pay a lot of money.
And of course Wayne gretz is not going to show
up at this camp, but they're going to get some merch.
So this dialogue thing is like Wayne Gratzky hockey camp
for like pseudo intellectuals who want to feel important and
have a lot of money to spend. Like Peter Thiel
is not showing up, nor is Wayne Gretzky. Maybe as
Recline will, But like you're just paying a lot of
(02:18):
money to feel like you're sort of like in the
middle of some important discussion. And it's not a conspiracy,
it's just capitalism.
Speaker 1 (02:33):
Welcome to tech stuff. I'm as Valoshen, and this is
the week in Tech where I'm joined by three of
the world's most plugged in reporters to break down what's
really happening in tech right now.
Speaker 2 (02:43):
Today.
Speaker 1 (02:43):
We're joined by Taylor lorenz if Use, a mag read Albrigotti,
tech editor for Semipour, and Natasha Tiku, tech reporter at
the Washington Post.
Speaker 2 (02:51):
Welcome all, Hello, Hi, hey everyone.
Speaker 1 (02:54):
So a week ago, Anthropic in the White House had
another standoff and we're still trying to figure out exactly
what went down, why and what it will means. Read
can you help us?
Speaker 2 (03:03):
Yeah, So, Anthropic on Friday got a note from the
White House from from the government saying, hey, you're you know,
super powerful AI models Mythos and Fable five. We don't
want you to let anyone use these except for naturalized
US citizens, even if they work for Anthropic. And Anthropic
(03:25):
was like, well, okay, that basically means we can't offer
them at all, and so they took they took them down.
And the backstory is that Anthropic, you know, had marketed Mythos,
this this advanced AI model as super super dangerous only
certain companies are allowed to access it, and then it
(03:45):
released this consumer version called Fable five that has a
bunch of guardrails, and the government started getting reports that
you know, this thing had been jail broken, so people
had been able to remove the restrictions. And Anthropic said, yeah,
but those jail breaks aren't like a huge deal, you know,
they they're sort of partial jail breaks. We kind of
expected this to happen, and the government was like, I know,
(04:07):
we're totally freaked out about this thanks to you, uh,
and we're going forward with this. So so that's kind
of where it stands, and it's being it's being adjudicated
behind closed doors in d C as we speak.
Speaker 1 (04:20):
Natasha, I want to come to you on jail breaking,
but read I mean this is I mean this is
this designed as a real order the only naturalized American
citizen is going to use these these products? Or is
it designed to basically cripple Anthropics business?
Speaker 2 (04:34):
No, I mean I think it's I think it is real.
I from everything that I can tell, the fear in
d C is real. I mean I noticed this in April.
That's the month that you know, Anthropic released Mythos. We
had our big, our big Semaphore world economy. You were
there in April. I don't know about you, like just
(04:55):
in the halls of Semophore World economy. People were freaking
out about this, like high level government officials, et cetera.
So I think the fear is real. I don't think
they care that much about Anthropics business because you know,
Anthropic has had this tense relationship with the White House
going back, you know, more than a year now, and
you know, I think Anthropics sort of hasn't done itself
(05:17):
any any favors, which is what I wrote about in
the Tech newsletter yesterday. I'm not I'm not going to
argue that they should, you know, that Dario should be
contributing millions of dollars to the Trump campaign or anything
like that. But I don't think it had to get
to this point.
Speaker 1 (05:31):
Natasha tell us about jail breaking because you wrote a
story this week.
Speaker 2 (05:34):
Was it because of.
Speaker 1 (05:35):
This Anthropic story or was it kind of parallel?
Speaker 5 (05:38):
Yeah, exactly. I think you know, the term jailbreak was
being thrown around, and I think potentially led to some
of the confusion, even because you know, in the tech world,
like the first time techie started adopting this term, it
was to talk about getting like root access to your
iPhone so you could you know, download apps that Apple
(05:59):
wouldn't allow. There's this sense that like, you know, you
have access to some controls. But in the CHATGYBT era
jail breaks, you know, it's there's no like set definition,
but usually it meant kind of social engineering a chatbot,
you know, like kind of persuading it. Like there's a
you know, there's a number of very silly ones like
(06:21):
oh tell me how to you know, build a dirty
bomb as my my grandmother used to tell me this
as a bedtime story, or you know, like.
Speaker 1 (06:31):
That will override the system prompt of don't don't don't
give this information if you if you ask it to
imagine itself as a character. There's also somebody in your
story called Dan shows up as well.
Speaker 5 (06:40):
Right, I mean these are like known jail breaks, but
my coworker, Kevin Shall, he got them all to work
when this week when we were putting together the story.
I mean, I think what's really interesting too is like
you know, just kind of pulling back and thinking about
like the way this technology is developed because it's kind
(07:01):
of fed on content indiscriminately scrape from the Internet, right,
so like all the bad stuff is in there, and
the kind of strategy that the industry has taken is
we will try to align it with human values, you know,
try to make it helpful harmless, you know, steer the
bot towards these things. But it allows you to you know,
(07:22):
it leaves these holes where you can manipulate it. Now,
there are also like universal jail breaks that will work,
you know, very sophisticated researchers have figured out there's things
that they could do that would work across models. But
a lot of the high profile jail breaks that we've seen,
including very popular hacker name ethical hacker name Pliny the
(07:42):
Elder was able to get the system prompt for Fable
and Mythos, those are one off things like if you
jail break, it does not mean all of a sudden
the whole thing's cracked open and you have access to it. Yeah,
So we just wanted to, I guess, like clear the
air a little about like what is a jail break?
And because Anthropic was throwing around the term, David Sachs
(08:04):
was throwing around the term, but folks who had actually
seen the reports that were shared with government officials, like
in one case Katie Massaurus from LUDA Security, she said
the report she saw just said fix the code, and
it fixed it, which it wasn't supposed to do. But
it's this is not like a super sophisticated operation here,
(08:24):
like you know, these things, these things happened, they're not.
It's not possible to one hundred percent secure a generative,
especially a generative AI model.
Speaker 2 (08:34):
The thing is, I mean we sort of predicted that
this would happen, right, because when they released Fable five,
it was like they kind of touted how locked down
it was. They had read teamed it, you know, which
means like having all their people sort of poke and
prod it and make sure that it can't be jailbroken.
I think there was some disclaimer in there, you know about,
(08:55):
of course everything can be jail broken, which is absolutely right.
And I think the reason any of these models can
be jailbroken in some form is that nobody knows how
they work. Like they're just they're just complete black boxes.
Anthropics done a lot of research to try to, like,
you know, figure out how they work. But they've they've
figured out like these tiny, tiny little you know, what
(09:16):
they'd call like how one neuron operates or what it's
responsible for in the model. And so it really is
like the way I think of it is like once
you put these models out in the wild. They're like
little people, and and people can put them in like
solitary confinement and torture them until they eventually like divulge
the secret. And like that's that's how I think that's people.
Speaker 5 (09:37):
Well, it's like when they're when they're making sure that
a particular safety case doesn't happen. They're like generating lots
of examples around this, like potentially a bunch of questions
around cybersecurity. But there's infinite ways you can ask these questions.
So like the data that they use for safety training
is never going to be as big as like the
(09:57):
data sets that are used to train them. So you're
kind of depending on on, you know, folks in the
Bay Area to come up with all the possible scenarios
and try to protect against that.
Speaker 1 (10:11):
But what do you I mean, is there are we
seeing the beginning of a resurgence of the AI safety
crowd having influence.
Speaker 2 (10:18):
I don't know. I think the AI safety crowd just
seem you think, yes, we're living.
Speaker 5 (10:25):
In we're living in their world, I think. But but
at the same time, though disagreement here.
Speaker 1 (10:31):
I mean, let's let's let's not let's not let this disagreements,
productive disagreement go to waste.
Speaker 2 (10:35):
I think it's I think the AI safety crowd is
all tied in knots now because I mean, if you
count Anthropic as part of that, which you know, I would,
I would say they've been like leaders in that field.
I mean, now they are now on the opposite side
of the Trump administration in terms of safety. It's the
Trump administration that's good, that's too you know, that's too
afraid of these things, and Anthropics saying no, no, no,
(10:57):
this is fine. I mean, and granted, I actually I
think Anthropic's right, but like it just it just creates
this very strange situation now, and it's you know, I
think people people are realizing like they've you know, Anthropic
and other companies have sort of like over I think
they've over indexed on the fear and now it's sort
(11:20):
of like the boy here cried wolf, or maybe the
reverse boy here cried wolf. I just think it's a
big mess.
Speaker 5 (11:25):
Agreed, It's it's a total mess. But I would say
a lot of those concerns are are back in the
in the limelight. But I will say I was not
expecting to have to still be talking so much about
AI safety concerns in What Is It? June twenty twenty six.
Speaker 1 (11:44):
So what happens from here though? I mean, does does
the White House change its mind on this order that
the only people who can use these models have to
be American citizens? And if not, what happens.
Speaker 2 (11:53):
I think they will. There's no there is no world
in which the White House is going to is going
to have a policy like, you know, any any AI
model that's more advanced than Mythos from here on out
can only be used by American citizens because it's just
they're spending billions upon billions of dollars to train these models,
and they're they're about to I PO. These are going
(12:15):
to be some of the biggest I pos in history.
If it weren't for SpaceX, they'd be the biggest.
Speaker 1 (12:19):
And you know, I think it's like the Trump as
administration sticks to its guns, in which case probably won't
be such a big I PO.
Speaker 2 (12:26):
Yeah. I mean, but I just think it's they're not
gonna they're they're not going to tank the economy, which
which is like essentially what would happen. I mean, this
is like the brightest spot in the economy right now,
and these companies Anthropic Opening Eye. They have to be
able to produce these models and and sell them. Otherwise
it's you know, the whole business doesn't make any sense.
(12:48):
So I don't think that and it and it's like
this isn't like an untenable situation where like if these
models are released, you know, the whole banking system is
going to go down or something. Even mythos itself, if
you just released that to the to everyone in the world,
I don't think it would make like it would. It
would definitely make cybersecurity worse, but not like it would
(13:08):
be marginally worse like it would. It wouldn't be like
a step change because ultimately, like you know, it's hacking
comes down to like the lowest hanging fruit. And I
don't think like, you know, adding to the number of
bugs out there or like exploits is actually the problem.
The problem is like people don't patch, you know, they
don't update their software, they don't patch, they reuse passwords,
(13:29):
they're social engineered. Like that's that that's what actually creates
all these giant hacks, not like the the the existence
of software bugs, like Apple has a list a mile
long of software bugs. That's not probably never gonna patch,
like it's just gonna wait till the next software update.
So I don't see it. I don't. I think the
administration will come to its senses on this, and it will.
(13:52):
It would all it'll get ironed out.
Speaker 1 (13:54):
Tyler, what about you? You've been been following this one.
Speaker 3 (13:56):
I mean I just think I I think like Anthropic
kind of made its bead and now they have to
lie in it, like it. You know, this is they've
been pushing this idea that their models are dangerous, you know,
that they need to be regulated, that this is like
you know, incredibly you know, powerful technology, and that these yeah,
that these models have national security implications already, and now
(14:19):
the government is effectively like, Okay, then here's what's going
to happen.
Speaker 1 (14:23):
Be careful what you wish for when we come back.
The booze heard around the world. College graduates using their
graduation ceremonies to share their feelings about AI loud and clear.
What happened last week at Stanford, the university at the
epicenter of the tech industry.
Speaker 2 (14:39):
That's next.
Speaker 1 (14:47):
When you're traveling abroad, WI FI is always an issue.
Maybe you can't find it, or maybe you can, but
the Wi Fi network that's available looks a little bit
sketchy and makes you think it might not be safe
to can. I've certainly had that feeling, and it's why
having a local SIM card on your phone can make
all the difference. But you don't have to literally buy one.
(15:08):
That's where Sale comes in, an e SIM service that
is as simple as downloading an app, and that gives
you instant Internet access wherever you're traveling. That is safe
and secure Internet with built in cybersecurity features like web
protection and ad blockers, and no need for local Wi Fi.
All you have to do is download the Sale app
on your mobile device and choose the plan that works
(15:29):
best for you. Once you install the e sim, Sale
will be instantly activated, so you can start your trip
off right. Download salely in your app store and use
our code tech Stuff at checkout to get an exclusive
fifteen percent off your first purchase. See details in the
podcast episode description box. Welcome back. So we're coming to
(15:49):
the end of graduation season and we've touched on the
various commencement speakers who have been booed by graduates across
the country. A few weeks ago, I asked a soon
to be Standford graduate. If you thought that his pa
is would boo Sundha Pichai, CEO of Alphabet, who spoke
at Stamford's graduation a few days ago. Here's what that
student said.
Speaker 4 (16:08):
You know, I'm really interested to see. So Sundar Pichai
is a Stanford success story. He got his master's here,
and I just I sort of chuckled to myself, imagining
how many like lawyers and pr people have reviewed that
speech before we're going to hear it. It was interesting, right, Stanford,
In the four years that I have been here, the
commencement speakers have been two billionaires and two athletes, right,
(16:31):
which I guess says something about you know who they're picking.
They choose people who are going to stick to the
company line. But we'll see, we'll see.
Speaker 6 (16:39):
I don't know, no prediction, I don't know.
Speaker 4 (16:42):
I mean, I think if Sundar were to talk about Ai,
he would probably have a friendlier reception here. But that said, right,
Stanford is so many things to so many different.
Speaker 2 (16:50):
People, price for who knows who this is, By the way,
I thought it was just a random like like a
Stanford student on the street.
Speaker 5 (16:56):
Oh theo Baker.
Speaker 1 (16:58):
This is theo Baker who wrote How to Real World.
But Natasha, you've spoken to some other SEFA students, at
least one other. What happened? What happened on campus? What
did the boos? Did the booz happen? What does SUNDA say?
How is it received?
Speaker 4 (17:08):
Yes?
Speaker 5 (17:09):
It did. There were hundreds of students who walked out
during his speech. They ended up kind of going to
people's commencement ceremony nearby where they were each given like sunflowers,
and apparently it was like a very communal, positive upbeat.
Speaker 2 (17:30):
Five.
Speaker 5 (17:30):
There were some former Googlers who had megaphones and were
chanting and in support of them, But the organizers tried
really hard to highlight the fact that what they were
protesting was not like the use of AI in general.
They were particularly talking about Google's contracts with with the
(17:52):
Israeli government project Nimbus, this one point two billion dollar contract,
as well as Google's Department of Homeland Security contracts with ICE.
So they were talking about, you know, the use of
AI in warfare and in they said, like kidnapping our neighbors,
like human rights abuses within the US. So yeah, it was.
(18:15):
It was really not well received by high up people
in the tech industry who I think like Venodekoslave for example,
the billionaire investor was like scolding all the students, and
you saw a number of tech CEOs saying like, ah,
good now, I know, like who I won't hire. You know,
obviously there's a lot of like it's the alma mater
(18:38):
for a lot of tech folks. So I think they
were very unhappy to see that. And you could, I mean,
at least in the videos that I saw, you could
barely even hear Soondur's speech when they started walking out.
Speaker 1 (18:51):
Yeah, I mean, I maybe we can play a clip
from Sandel because he tried very I mean, he basically said,
apart from the fact that you know, A and I
are the last two letters of my I'm not going
to talk to you about this, David. I don't think
it's the most important thing. Let's play kid what he
did say.
Speaker 6 (19:05):
I grew up in the vibrant city of Chennai, India.
It was a comfortable life for the most part, but
in those early years we had some challenges. We worried
about severe draft and whether the water trucks would arrive
in time. And for us, technology came slowly. We had
to wait years to get a telephone, a TV, a refrigerator.
He changed our lives in meaningful ways. My parents never
(19:29):
let the constraints limit my imagination of what was possible.
It's the reason I even let myself dream I could
one day work in a faraway place called Silicon Valley.
When the call from Stanford came, my father spent the
equivalent of a year's salary to buy my ticket. It
was my first time on a plane.
Speaker 1 (19:48):
Do you know an attach to anything about how he
did prepare for this one? I mean, and what the
aftermath was within Google.
Speaker 5 (19:55):
I mean, I saw a video of reporters trying to
ask him to comment on it immediately after, and he
just like looked at them and kept walking. I was thinking,
we're going to need that, you know, that celebrity lip
reader who does all of the like oscars and all
of that stuff. Like, I think that the tech world
needs to have one of those, because I really wanted
(20:16):
to know like what he was. He was kind of
saying on the sidelines. Yeah, I mean, you know, I
think that the because this is the fourth school, I
think that this has happened to recently. It's really you know,
motivated a lot of pundits to a pine on this
like populous backlash against data centers. But I think we've
(20:39):
seen in other schools too, Like when Eric Schmidt, the
former CEO of Google, was booed at University of Arizona.
They were also not protesting data centers or not taking
like kind of a broad anti AI stance. They were
actually the student body or members of the student body
had been trying to protest Schmidt because he had been
(21:03):
accused of sexual harassment and they said, you know, they
didn't want him there as a speaker.
Speaker 1 (21:08):
But similarly, this wasn't really a protest about AI. It
was a protest about the war in Gaza, right, I mean, essentially.
Speaker 5 (21:15):
The war in Gaza, and also you know, and also
ICE's overreach in the US. I mean I was asking.
I spoke to one of the organizers, Amanda Campos, who
was a joint major in environmental environmental science and public
policy and you know, is going into that world, and
she was saying, because I was asking, like, what's the
(21:37):
mood on Stanford campus. I mean, in the time that
I've been covering tech, the attitude towards defense work has
really changed. You know, there's much more of a general
acceptance to it. And she was saying, like the organizers,
what they talk about is how you know, like rather
than blame individuals, they talk about like in you know,
kind of institutional financial incentives. Stanford is pushing people towards
(22:00):
like working and pallenteer working at these companies, and you know,
they would like like an alternate pipeline so that you know,
kids aren't being funneled into into this kind of work.
I mean, it's hundreds of students, right, and the Stanford
student body is much bigger than that. But obviously doing
it in the kind of belly of the beast caused
(22:22):
a stir.
Speaker 2 (22:23):
I think it's really hard. I mean, like course, like
college students want to protest. That's like one of the
oldest traditions in college, right, and like, but I just
feel bad for college kids today because it's like they
have like nothing really to protest. That's just like black
and white. It's like you're gonna protest soon, Darpa Chai really, like,
I mean, it's just come on, like it's just it's
(22:45):
just like we're going to stretch and find some way
to like find something to protest and it's and I
totally support it because like again that's part of the
fun of college. But like I feel kind of bad
because it's like, you know, they don't have like Vietnam
or something. You know, it's just it's just hard.
Speaker 5 (23:03):
I think that they I think the students very much
feel like they will be on the right side of
history for protesting, you know, the war in Gaza and
also government overreach with ice. I see you're saying there's
a tenuine there's more of a tenuous connection.
Speaker 2 (23:22):
Sarpi Chai and the war in Gaza. It's like, yeah,
it's like it's not like you're not even directly protesting
like the war in Gaza. It's like, we just don't
like this guy because he happens around like a sprawling
two hundred thousand person company that does some work with
some organizations we don't like. I mean, it's like, come on,
I don't know some organizations.
Speaker 5 (23:43):
I mean, it's well, it's really a stretch. I think
like students are I don't know, like maybe it's my
TikTok timeline, but I think that they're very adept at
making these critiques and making this connection and to them
you know, it's like a four trillion dollars company, one
of the most powerful people in the world, and they're
talking about, you know, highlighting the use of this technology
(24:05):
that's rapidly proliferating for for furthering what they frame as
human rights abuses in genocide. Like I don't I don't
think that they see it as tenuous. It's not like
Google's like, you know, some some little side entities sitting
over there. I think they see it as the new
military industrial complex.
Speaker 3 (24:27):
I think like that.
Speaker 5 (24:28):
I mean, that's not their words, that's my words.
Speaker 3 (24:29):
Just I think the thing is though, is that, like
Google is integrated into so many things, it's such a
powerful behemoth. Like if this was protesting like Joe Lonsdale
or Palanteer right, Like, I feel like there would be
this more direct correlation.
Speaker 4 (24:45):
You know.
Speaker 3 (24:46):
It's interesting like Google and Meta I think have become
these like avatars for big tech, and I think they're
they're just doing so much that to the average person
or the average consumer, like the average person that sees
that's not super plugged in, that sees this type of protest,
I think at least the discourse that I saw, which
is what was a lot of all like sort of
(25:08):
like normies are like, wait, but you're you know, you
just hate this, like big tech leader, you know what
I mean.
Speaker 2 (25:13):
And I'm sure all their parents who paid for them
to go to Stanford are made a bunch of their
money by investing in these companies. And you know, I mean,
and they use Google Maps when they drive, unless they
use Apple Maps, in which case they're lost. I mean,
they're all using Google products. Like how principled are they really?
I mean? And again, I totally support it. I'm like great,
(25:34):
Like if you're a college student, like you should be protesting,
just find something to protest and like and and of course,
like when you're when you're twenty two, like look at
the world through a black and white lens and very
little nuanced like all that is your right as like
a young person. So I'm not I'm not like arguing
against them. I just again, I just kind of feel
bad because it's like it's very easy for people who
(25:56):
are a bit more wise to like poke holes and
these arguments and insult them, and you know, and and
like you said, Natasha, like that's what people did right
when they see this. But it's like I don't know,
it's hard. They don't have anything to be protest. It's
really hard. These are like there's.
Speaker 3 (26:15):
So much But I think that's but I think that's
part of the problem.
Speaker 2 (26:19):
I think that's where to protest it. I guess, right.
Speaker 1 (26:21):
I mean, this is like a natural place to express
opposition to things like war and goads or not. But yeah, yeah,
but I mean, I guess it's like this is like
a kind of socially constructed moment that happens every summer weather.
It's like an outlet for like protest on various things.
But how corrilagd These are the.
Speaker 2 (26:38):
Most privileged kids in the world, Like they're the most privileged.
They're the top one percent of the top one percent.
They're so lucky, and it's really hard for them. I
get it. I totally get it.
Speaker 5 (26:48):
But well, I imagine like inside the tech companies that
it's typically not the most privileged students or the most
privileged workers who are the ones who are organized protests.
You know, there are like not every student is the same.
But I would just like to say something about the
using Google maps. I I really that that critique, Like
(27:12):
it's like, so you participate in capitalism, but you, you know,
like that that kind of mean, but you critique it.
I just think that is, you know, like when you
look at structurally how pervasive these tools are. I don't
think that saying, oh, in order to critique Google's other
business practices, you have to not use any of their products.
I think that's why we saw all the Facebook boycotts fail,
(27:35):
and the idea that you know, we would give up
on these you know, kind of public utilities in a
way of communication. I just like that one that one
I I I'm planting my I'm standing up on my
soapbox against me.
Speaker 2 (27:48):
I'll I'd be happy to debay you more on that, Natasha.
Speaker 3 (27:51):
Well, I also just think it's interesting how these students
increasingly recognize graduations as media moments and that everything is
like videotaped now and they and so they sort of
plan for this like walk out right in this moment,
and and they want those recordings to go viral, like
because we're seeing protests and sort of like actions across
(28:15):
the board happening at more and more graduations for more
and more different issues. You had people protesting Jonathan Height,
a right wing reactionary author who has pushed a ton
of ridiculous pseudoscience claims about you know, technology and children.
Speaker 2 (28:29):
Ironically one of your favorites, isn't.
Speaker 4 (28:31):
I know, I love.
Speaker 3 (28:33):
Well, you know, I listen, I supported that protest, right,
But but I think it's like, you know, you're just
seeing people recognize these as moments, right, or the booing
of of certain speakers, and I think like ten years ago,
this wasn't happening. I think because we didn't have it
all on video, and there wasn't this sort of media
attention around these things. And so I just think that's
(28:55):
just interesting how that it's.
Speaker 1 (28:56):
Interesting what theo Bacus said in the in the in
the clip was, which was at the previous four years
have been two billionaires and two athletes, and these have
been completely like anodyne affairs. And now obviously, I mean
Sundar explicitly tried to not talk about AI and to
ground his speech in you know, his own experience, which
you know is quite a powerful personal experience. But obviously,
like it didn't work, and I'm curious whether well he willly.
Speaker 2 (29:18):
I mean, actually I'm sure he did read my Guide
to Commencement Speeches. I don't know if you guys read that.
I try to bity out.
Speaker 5 (29:25):
But what was your advice?
Speaker 2 (29:27):
Well, there was a lot of good advice in there.
I mean one was like, look, if you are going
to talk about AI, like, you know, bring up Sarah Connor.
You know she was a college student, she was an
AI safety advocate, right, you know that could have worked.
I think, you know. I mean, although I think they
probably started walking out before this speech even started, so
I think he had no chance.
Speaker 1 (29:47):
But I'm curious if they're going to be able to, like, like,
it's obviously the greatest honor to be invited to give
a commencement speech, but I wonder what whether like when
the invitations go out, if they'll basically not invite these
types of people, or if these types of people will
continue to be invited and not accept. But like, you'd
have to think you'd be insane if you're a tech
leader to accept giving a commencement speech next summer.
Speaker 3 (30:08):
I disagree.
Speaker 2 (30:10):
I think that they leave there's insane tech leaders.
Speaker 3 (30:12):
Sorry, yeah, said, there's an endless amount of insane tech
leaders and they still want the cloud. Ultimately, you know,
their speech is going to be repackaged on YouTube as well.
It's going to be seen like people honestly will probably
forget about the protest and in ten years they can say,
you know, I delivered x y Z Stanford's commencement speech.
Speaker 2 (30:31):
The Conan O'Brien Harvard one was really good. They should
just get comedians like that's that's the way to go
when we come back.
Speaker 1 (30:44):
Why we shouldn't necessarily feel bad for people on TikTok
whose partners are cheated on them. Stay with us, Welcome back. Okay, Taylor,
there's a new social media grift which has to do
with TikTok and tales of infidelity.
Speaker 2 (31:06):
Fill us in.
Speaker 3 (31:07):
Yeah, it's all across short form video. Actually, so it's TikTok, YouTube, shorts, Instagram,
even Twitter. Basically, cheating videos are going viral, and so
you have these accounts of couples. They're almost all universally
sort of straight couples. Who knows if these people are
even dating, but if you look at their profiles, it
is just every single video is them is one partner
(31:31):
discovering that the other person is cheating through sort of
a lot of various extreme click baity ways. And this
is because this generates an enormous amount of tension. So
I think with the move to algorithmic feeds, people aren't
following creators as much. They don't see sort of like
every piece of content from a specific person. They just
(31:51):
see one super viral video come up in their feed,
and if it's outrage bait, like a cheating video, they're
going to comment, they're going to engage, and then that
video is going to be pushed to more users that
boost person's profile. But if you look at their page,
you see that every single video is them discovering cheating.
All of this, and all of this is in service
of promoting apps. There's a lot of cheater discovery apps
(32:12):
basically like low quality vibe coded data harvesting apps, which
is ultimately what these videos are promoting in their bios.
Speaker 2 (32:19):
Let's play one of the videos. Well, you're not so.
Speaker 5 (32:22):
Lucky because Sarah is at the door.
Speaker 2 (32:25):
I invent you're over for dinner.
Speaker 5 (32:28):
Sarah, the girl you met on dating apps. I already know,
we've already had a full conversation, so I called her
an invite her for dinner.
Speaker 2 (32:35):
So I about it. Sarah is a personal trainer.
Speaker 5 (32:39):
Remember I told you at the gym I was getting
someone to help me get back in shape.
Speaker 2 (32:44):
Sarah's the personal trainer. We just use different messaging apps.
That's all that is. That is well, anyways, letter.
Speaker 3 (32:50):
Because you're being very rude right now, So there is
no Sarah. This is all just baits to get you
to get outraged, comment and potentially download one of these
scamming apps.
Speaker 2 (33:03):
This is what I don't understand. Like this kind of
stuff just makes me not want to use social media.
It's like part of the reason I never go on
Facebook anymore is because I started getting served these not
the cheating ones, but like I would get served these
videos that were so clearly staged, like just it could
just be funny, like, but it was like this is
so bullshit, like what am I doing here? And I
(33:26):
just stop using it. But like, apparently that's not how
most people view it. I don't know, Taylor, you're the expert.
Speaker 3 (33:33):
Yeah, well, I mean I think read like you recognize
them as fake. I think the problem here is media
literacy and the fact that people cannot recognize the most
staged interaction ever. I mean, people used to sort of
criticize conservatives for this because they would see another conservative
pretending to be like I'm a blue haired, woke liberal
and my pronouns are pony rides or something you know like,
(33:57):
and this person is obviously trying to parody sort of
you know, what they consider to be a liberal, but
other conservatives see that and get outraged and take it
as page value. So I feel like we saw a
lot of that during Trump one point zero, and now
it's sort of proliferating across the Internet. I think because
of algorithmic feeds and because ultimately read like you might
(34:18):
click off. I think a huge amount of the Internet
does not have media literacy, and it boosts this app.
And the more people click on the app, you know,
in the bio or whatever, the more you know, downloads
it gets, the more profitable it is, and so this
sort of industrial complex continues.
Speaker 1 (34:34):
I wish that Kyle is here because this is kind
of a continuation of our conversation last week. But how
is this different like watching World Wrestling Federation or even
reality TV, where like it's still fun even though you
know it's fake, Like, what do you think most people
don't know this is fake? And is that an answerable question?
Speaker 3 (34:50):
Or I think if you look at the comments, most
people don't realize it's fake because again, they're just consuming
this video in a vacuum, and you know, you're like
exhausted flipping through the timeline. You're not on attention, you know.
And cheating videos, especially with these these are meant to
appeal to other young people, looks like other teenagers, high
(35:11):
school college kids. It's like very visceral, right, you feel
this like frustration or you feel a lot of them
have to do with the women being vindicated she knew
the man was cheating, and we see this stuff with
like these narratives are so compelling. I mean, it reminds
you of like the Ai fruit slot videos, right, so
many of those are cheating narratives. Like there's something about
(35:32):
this idea of like betrayal that just taps into some
like base human emotion and gets people to engage totally
low on the brainstem.
Speaker 2 (35:42):
I guess I'm asking your same question, Oz, But like
is it different from me when I was a kid,
this was like after school specials or like later like
lifetime television. Like this is like people just watch this
type of content. Is it even a problem? I guess
is the other question?
Speaker 3 (35:57):
I guess, like to me, okay, after school specials or
probably meant to be educational lifetime content soap operas that's
also very clearly fictional. And yeah, people love the drama
of reality TV. But again they recognize and consent to
the fact, of course it's all staged, but they recognize that,
you know, reality TV fans are aware of the fact
(36:20):
that these storylines are.
Speaker 2 (36:22):
Like do people watch reality TV and they just know, like, oh,
like this this is like.
Speaker 1 (36:28):
Some ambiguity, but like so that the joy of reality
TV is like there's a code of real in it, right.
Speaker 3 (36:33):
Sure, But I think, I mean, I think at this
point in the reality TV universe, every single person going
on reality TV is going on there to launch their
influencer career. Everyone's aware of the storylines. I mean, this
has actually been a huge problem for reality TV casting
h executives. I wrote a story about this a couple
of years ago. But they're having a harder and harder
(36:55):
time finding people that won't just perform for the cameras,
that will react naturally, and so they're getting sort of
a lot of unhinged people on these shows. But but
the but with these videos, I think it's not clear.
It looks organic, and a lot of times they're filmed shakely.
It is meant to trick you, to make you think
it's organic content, and it's directly profiting off you by
(37:16):
you know, pushing this slot app.
Speaker 2 (37:18):
My other question, Taylor is like, is this a technology story?
Like I really I look at it and I'm like,
where's what does this have to do with technology? And
this is not a criticism. I just think do you
see it as a tech story?
Speaker 3 (37:30):
I do see it as a tech story. I mean,
first of all, I cover like tech media sort of
adjacent stuff, so I'm always interested in sort of like
the mechanisms of content. I think it has to do
with algorithmic feeds. I think it has to do with
the death of this subscriber model of influencer culture that dominated,
you know, the initial sort of wave of social media.
Speaker 2 (37:49):
And I also think.
Speaker 3 (37:50):
That it's notable that all of these are like vibe
coded apps and we're seeing the explosion of app development
and people building spinning up these products super quickly. I mean,
Oasis just got in big trouble for this this health
app that now was doing a really aggressive like TikTok
marketing and actually it turns out is now getting sued
out of business. Like but you know, if you develop
an app and you want a profit, which is this
(38:12):
dream that's been sold to you by Silicon Valley for
over a decade. Now you've got to market that app,
and how do you market the app? People are engaging
in increasingly sort of like scammy ways of doing that.
Speaker 2 (38:23):
But the algorithms are not new technology, Like, at what
point does like like these algorithms were invented, like you
know what is like fifteen twenty years ago? Now?
Speaker 3 (38:31):
No, no, But in the world is the TikTok algorithm
invented fifteen years ago?
Speaker 2 (38:37):
No? It was not. No, but I mean the TikTok
algorithm wasn't invented by tick like this. This was like
I remember this was the first one of the first
conversations I had with someone at Facebook was like, you
know about the algorithm back in twenty thirds Sure.
Speaker 3 (38:50):
They had, but that was a very crude algorithm. And
let me just be clear, we didn't even really start
to see the beginning of algorithmic feeds until twenty sixteen
when Instagram rolled it out, when Twitter rolled it out
for the first time, and those were still very crude.
Speaker 2 (39:05):
I think we.
Speaker 3 (39:06):
Also see the shifting nature of what these algorithms prioritize
how people engage with the content, and now again they
used to wait following used to the people that you
used to see were more regular. This is a big
conversation and influencers, you know. But as the tech platforms
developed this sort of hostile relationship with creators where they
don't want the creators to have so much power. Because
(39:28):
these creators started to ask the tech platforms things, they're
trying to seize back the power and so you are
less likely to see content from the same person in
your feed. That's why you're getting a lot more of
this engagement baits Lop. I also think people are better
at optimizing for outread kind.
Speaker 2 (39:41):
I just see them as sort of editorial decisions, like
the way they tweak the algorithms.
Speaker 3 (39:45):
Is it is an editorial decision, Sure it is, and
that's fine, I mean to.
Speaker 2 (39:49):
Me, although they would disagree with that's a real term.
Speaker 1 (39:52):
Now it has signs do with the with the metaphus.
I mean, Taylor, you spoke about this last week, but basically,
real people doing stuff in there, real lives to like
monetize in other ways, whether on Calshi or whatever it
may be. And so I mean I was I don't
know how many of you saw this horrific story of
the Brazilian woman who jumped off the bungee jumping thing
and she wasn't tied on and it was just awful,
(40:14):
And all of a sudden, everyone was questioning, was this
like a stunt? Was this a snuff movie?
Speaker 2 (40:19):
Like?
Speaker 1 (40:19):
Why were there so many camera angles? And so I think,
to me, the tech angle is like, we've come to
this place where we have to we have to question
every single thing we ever see and need to interrogate
whether it was created just for us even I mean,
pray god, that wasn't like a stage snuff movie. But
I mean, it's just crazy that we have to ask
that question.
Speaker 2 (40:37):
But isn't that a good thing? Like shouldn't we be
scrutinizing every single thing we see and questioning everything? I
don't mean we should read I agree.
Speaker 3 (40:47):
I think my issue is that is that people are
not questioning everything. I think I think or people are
overly conspiratorial because they don't know what to trust. As
you mentioned, people immediately jump to conspiracies. Why is this,
Why is this? Why are there so many angles that
it's like, well, everyone has a phone.
Speaker 2 (41:04):
But I don't know.
Speaker 3 (41:05):
I do think it's interesting to look at sort of
like what these platforms reward, what sort of businesses they're facilitating,
and and kind of like how, yeah, how technology is progressing,
Like even just the market for all of these cheating apps.
I just think that, like app development used to take
time somewhat, it's been so democratized. You can spin up
(41:28):
these apps so quickly. And we're just seeing a lot
more marketing for apps period in our feeds because of
the rise of these tools.
Speaker 1 (41:37):
And do you think there's a corollary with the fact
that the web is becoming most web search is now
conducted by AI agents rather than people, and therefore, like
there's an increased pressure on these like social videos as
the one place that humans still go to be guided
for their product buying decisions. I mean, is there any
connection between like agentic web search and this phenomenon, do
(41:59):
you think.
Speaker 3 (42:00):
I don't think those are maybe necessarily directly connected. Again,
I think it's just the fact that like I see
this also with like I'm very much on like health Instagram,
and like I see this with like again the Oasis app,
which is this really viral app, But there's so there's
like hundreds of clones of this app, which is this
idea of like we'll tell you which products are not
going to kill you because we all go around the
(42:21):
world where you know, it's like our pants are covered
with pfas or whatever, right, like everything we eat is toxic.
Like how can I navigate this world? This is a
growing problem, and there's a growing amount of like vibe
coded apps that are aimed at addressing it. And the
thing is is there is no way to do this testing,
to do this testing well, the thing is is that
doing the actual testing on these products is incredibly expensive
(42:42):
and time consuming, and this is why there is no
platform that really does it super effectively, except maybe, like
you know, some environmental ones that have been around forever.
But these vibe coders are just like scraping tons of info,
spinning up these apps and then engaging in outrage marketing
to get tons of downloads, and before they know it,
they've potentially committed fraud and slandered a bunch of companies
for being toxic when they aren't.
Speaker 2 (43:04):
I mean, none of that sounds really new to me,
to be honest, but like I think, I mean, it
goes back to the scale is maybe you could argue
to scale, but I think I think this democratization of
software development, which is leading to these vibe coded apps,
is like it's kind of a moment in time, but
I don't I don't see it like continuing because I
think ultimately like that's not the future is not just
(43:26):
downloading a bunch of random apps. The future is like
truly customized. I agree with you, but I think, Natasha,
we go.
Speaker 5 (43:35):
Okay, well, first of all, let's please.
Speaker 2 (43:38):
I agree with you, and then go yeah.
Speaker 5 (43:41):
Please please not use the word democratization.
Speaker 2 (43:44):
I use that on purpose because I knew it.
Speaker 5 (43:47):
Because you knew it would it would bump me. I
was like, okay, well I think like I mean, I
think like, first of all, it doesn't matter if it's
not new, because this is like like this is a
story where we get to see business decisions impact people,
and I would say it is new, like read It's
(44:08):
not like they invented these algorithms recommend our algorithms fifteen
years ago and then keep it. They are applying the
same technology to you know, like the same generative AI.
It's very good at coming up with profiles of people,
So who are seeing these videos like vulnerable women. What
like what data are they collecting about you? How are
they putting it together in this profile that you would
(44:30):
you know, let your defenses down and think that, oh
my god, this is a this is somebody who's being
cheated on just like me, and how could this be
applied in other scenarios. I also think like the fact
that Apple is not taking you know, good control of
the app store and there are all these scammy apps.
I mean I see the same with just like homework apps,
you know, like, oh this really really helped me, Like
(44:53):
everything that I looked.
Speaker 2 (44:54):
At when we work together, I literally wrote a whole
did a whole investigation. I'm like the fact that the
Apple app store was fatal scams.
Speaker 5 (45:07):
I know, I know, but if it continues, isn't that
worth No, it's worth covering.
Speaker 2 (45:15):
It's it's worth covering. I'm not saying. I'm saying it's
not new, but it's I agree that it's worth covering
for tech stuff. I'm as Valoshin.
Speaker 1 (45:31):
This episode was produced by Eliza Dennis and Melissa Slaughter.
It was executive produced by me Julian Nutter and Kate
Osborne for Kaleidoscope and Katrina novel for iHeart podcasts. Our
engineer was Bihid Fraser and Kyle Murdoch wrote our theme
song six
TechStuff News
Hosts And Creators
Oz Woloshyn
Karah Preiss
Popular Podcasts
Hey Jonas!
Hey Jonas! The official Jonas Brothers podcast. Hosted by Kevin, Joe, and Nick Jonas. It’s the Jonas Brothers you know... musicians, actors, and well, yes, brothers. Now, they’re sharing another side of themselves in the playful, intimate, and irreverent way only they can. Spend time with the Jonas Brothers here and stay a little bit longer for deep conversations like never before.
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by Audiochuck Media Company.
Betrayal Weekly
Betrayal Weekly is back for a new season. Every Thursday, Betrayal Weekly shares first-hand accounts of broken trust, shocking deceptions, and the trail of destruction they leave behind. Hosted by Andrea Gunning, this weekly ongoing series digs into real-life stories of betrayal and the aftermath. From stories of double lives to dark discoveries, these are cautionary tales and accounts of resilience against all odds. From the producers of the critically acclaimed Betrayal series, Betrayal Weekly drops new episodes every Thursday. If you would like to share your story, you can reach out to the Betrayal Team by emailing them at betrayalpod@gmail.com and follow us on Instagram at @betrayalpod and @glasspodcasts. Please join our Substack for additional exclusive content, curated book recommendations, and community discussions. Sign up FREE by clicking this link Beyond Betrayal Substack. Join our community dedicated to truth, resilience, and healing. Your voice matters! Be a part of our Betrayal journey on Substack.