July 17, 2026 • 37 mins
Who’s afraid of China’s open-source AI models? Nitasha Tiku (The Washington Post) breaks down the secret battle between Silicon Valley and Chinese AI labs, why Anthropic was surveilling China-based users, and whether "adversarial distillation" is a national security threat or just a term American companies invented to protect their bottom line.
Then: ‘hot surveillance summer’ is here. Taylor Lorenz (User Mag) on the changing public perception of smart glasses, why the Kylie Jenner partnership was so successful, and where this trend may lead.
Also, Apple is suing OpenAI for allegedly stealing trade secrets.
Additional Reading:
- Inside the secret AI war between Silicon Valley and China | The Washington Post
- Hot Surveillance Summer, AKA Why Smart Glasses Feel Different This Time | User Mag
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 (00:11):
Natasha Taylor. There was a fascinating story and Wired last
week about an internal memo Meta regarding a loose squirrel.
Speaker 2 (00:19):
I did see this, Yeah, I didn't understand.
Speaker 3 (00:22):
We had a bird get into the New York Times
office once and it did derail our afternoon.
Speaker 1 (00:26):
So this was this was a hard hard to pass.
But there was a squirrel mails to Meta in the Bangkok.
Speaker 3 (00:32):
Office, the poor squirrel.
Speaker 1 (00:34):
There was great hilarity because it's been a rough time
for Meta, but people were commenting on how this was
the most joyful moment the company had for a while,
despite the fact that Janet had got scratched on his finger.
Speaker 3 (00:45):
Well, I hope he got a Rabi's shot.
Speaker 1 (00:50):
Welcome to Tech Stuff. I'm was Vloscha and this is
the Week in Tech where I'm joined by the world's
most plugged in reporters to break down what's really happening
in tech right now. Today we're joined by Taylor, the
rends if use a mag and Natasha Tiku, tech reporter
at the Washington Post.
Speaker 3 (01:04):
Welcome both, Hey everybody, Hello, thanks for having us.
Speaker 1 (01:08):
It's good to see you.
Speaker 3 (01:09):
Natasha.
Speaker 1 (01:09):
You recently published an article titled Inside the Secret Ai
War between Silicon Valley and China. Great headline, but it
wasn't about ships or competing for talent, but actually about
the models themselves. Can you explain what's going on here?
Speaker 4 (01:25):
So this is something of a narrative war between the
US leading AI companies and the leading Chinese AI companies.
So what companies like Anthropic and Open Ai are alleging
is that these Chinese companies are accessing their models which
are not supposed to be available in China and Hong Kong.
(01:48):
They're accessing their models in order to do this process
called distillation, which basically involves kind of you know, distilling,
taking the learnings from a larger model and applying it
to us to improve a smaller model. And distillation is
a process that has been used in the AI industry
for a really long time. It's standard process. It's how
(02:08):
state of the art improvements are diffused throughout the system,
you know, through academia and open source. But Open Ai
and Anthropic have started using this term adversarial distillation and
they are like seeking government US government protection from this process,
and they're alleging, you know that a lot of the
(02:30):
gains that we're seeing in Chinese models in open source
AI in particular is coming at the you know, at
the expense of anthropic.
Speaker 1 (02:39):
Yeah, I have a serial distillation. It sounds like makes
me think of prohibition actually.
Speaker 4 (02:44):
But.
Speaker 2 (02:46):
Yeah, I was.
Speaker 4 (02:47):
I was at this National Security AI conference and the
person from open ai brought it up, and Irene's salimon
from hugging face like Brazer Hins. She was like, I've
never I've been in this indust for so long. She
was a super early open ai employee, and she was like,
I've never heard this term before, but now it's now.
It's kind of it is diffusing throughout the US because,
(03:10):
you know, they're extremely concerned that the Chinese open source
models are cheaper, more accessible, and are being adopted very
widely by US companies. Not that they the US companies
necessarily want to talk about it, but you know, particularly
around the restrictions around fable that happened with US government,
(03:32):
the inability to access mythos. You know, companies are like,
it's more reliable and more cost effective for me to
build on top of Chinese technology. And I think another
kind of distinction that's really been alided over is that
distillation might be a violation of your terms of service.
You know, if you use if you access the Anthropic API,
(03:54):
you are not supposed to try to ask questions twenty
seven million times and then and you know, use it
to improve your own model.
Speaker 1 (04:02):
What's the difference between what's happening right now in your
story and the deep Seek freak of early last year, Like,
is this kind of a continuation of the same story
or is it something slightly different?
Speaker 4 (04:13):
No, I would say it's a continuation of the same story.
In that case, it was open Ai alleging that, you
know that Deep Deep Seek did the distillation, but it
also has to do with the fact that the US
was really caught off guard by how good the Deep
Seek model was. And you know, many of the gains
that you're seeing in the Chinese market are you know,
(04:36):
like open Ai and Anthropic would like you to believe
that it's only because they're taking from US models, but
that's not true. In fact, like many of the you know,
speaking to two experts in the field, they say that
like many of the gains that the Chinese models have
made in efficiency, in size, cost effectiveness, those have all
(04:57):
come from the constraints on the industry, like the fact
that we have export controls on chips. You know, they
don't have access to as many as.
Speaker 1 (05:06):
Resources Jnsui Aguen, right, which is a necessity is the
mother invention and actually restricts exactly.
Speaker 4 (05:11):
I mean, you know this is like the startup way.
And obviously you know that if you look at like
how much money has been raised by the major AI labs,
that's not been a constraint for them, right, And uh,
you know what's happened recently is that a one of
the Anthropic users in Hong Kong discovered that Anthropic had
(05:33):
was using this kind of covert way to identify Chinese users.
They were looking at what time zone they're in. They
were looking at the you know, where they're logging in from.
Is it associated? Is that domain associated with the Chinese
AI lab? And then it's sending the information back to
Anthropic but in a way it like hidden in today's date.
(05:55):
So like the apostrophe in today's date said like you know,
Chinese user associated with minimax logging in in this time zone,
et cetera. And people were really upset about this, like.
Speaker 1 (06:09):
This is sort of a nightmare scenario for AI companies,
for users to feel like they're being actively surveiled, right,
I mean, Taylor, what's your I know that we're going
to talk about surveillance in the second half of the episode,
but what's your what's your take on the idea that
you know, Anthropic have been caught essentially spying on users
to try and figure out where they are and if
they're Chinese users, illicitly trying to distill the models.
Speaker 3 (06:31):
I think this whole thing is honestly, I don't want
to say a losing game, but it seems ridiculous and
I'm kind of curious to Tasha, like why why isn't
the US developing open source models? Like why why are
you know, why can't we have our own open source
that competes with these Chinese open source models.
Speaker 4 (06:49):
Well, now there is a lot more investment going into it,
but yeah, it wasn't like fine as financially lucrative. And
you know, even the you hear a lot of rhetoric
from the VCS, right like we we can't stop AI,
like we have to support open source. You weren't seeing
commensurate investment in open source. There are like a lot
(07:11):
of Alan Institute for AI hugging face.
Speaker 2 (07:16):
There was this project called the Adam Project.
Speaker 4 (07:19):
People were really interested in pushing this and having you know,
the US be the center of open source. But yeah,
there wasn't as much investment in it, and the vcs
are all backing the closed sourced companies, and so yeah,
this is the situation we now find ourselves in where
we're playing ketchup.
Speaker 3 (07:37):
Yeah, I feel like we're playing ketchup. And also it's
a losing battle. Like I mean, I know Alex Karp
was talking about this recently, and obviously he's self interested.
I think volunteer uses, maybe the open source models, but
but like it just seems like this idea of this
like closed source situation and like tokens and I don't
(07:58):
I don't know, like everything for profit like that Anthropic
and OPA are built on is just again it's just
a losing game because ultimately open source is going to win.
Speaker 1 (08:08):
I guess if Reid was here, if our play read,
because he's not here this week he's in Boston. But
I mean, I think what he would say is yes,
except that if you are leading edge company or startup
or lab, you need to have access to the best
models otherwise you will fall behind competitively. So like the
future general you say, I may be more and more
open source, but there will always be demand for like
(08:31):
mythos or whatever it is that you have to pay for.
Is that I mean? I guess what Red would say, Well.
Speaker 4 (08:37):
You can, I mean, you could build a state of
the art model on top of open source. You don't
need to start with Anthropic. You know, Anthropic could be
building on top of deep seek. I think, like, you know,
what's really interesting is that the US position, at least
under David Sachs, was that you want to build on
top of the American stack, like you want to make
(09:00):
sure just in the same way that you know, China
has diffused its power by working with you know, companies
in the Middle East and Africa. They were saying, you know,
we want the US, the US stack to be like
the layer on open source. But then the Trump administration
policies are really going against that. And actually Anthropic had said,
(09:20):
you know, they tried to argue that if the US
enforces you know, a stronger stance against distillation, that maybe
the US labs could maintain their lead for twelve to
twenty four months longer. And yeah, it's just it's uh
has led to exactly like what you're you, Taylor. You
mentioned Alex Carp's rant recently, and it's completely related because
(09:45):
companies are extremely frustrated. They feel like, you know, do
they even own the intellectual property that comes from using
your model? Why are they paying for tokens? Like it's
really had people even kind of rethinking the potential business
model and what the value is for the companies if
like they don't own the technology there. Yeah, it's just
(10:07):
I think they feel like it's highly inefficient and not
a good way to to like start the new AI
era building on this technology that could be turned off
by the US government or anthropic could change something, or
they're spying on you.
Speaker 3 (10:21):
Yeah, it seems like ridiculous. It seems like obviously we
should be moving towards an open source model. Obviously China
has the superior sort of structure to the AI environment. Also,
it's better for consumers, it's more affordable, it's better for
business consumers, better for I would say average consumers as well.
I think, I mean, I don't use deep Seek for
like tons of stuff, but I find it completely comparable
(10:42):
to the US you know, companies that are out there,
and I just I also wonder if if it would
help if we if we were investing more in open
source as well, if it would help sort of stem
some of the like curb some of the Antai backlash,
because I feel like so much of the Antai backlash
is not just from companies, like people like Alex Karp
(11:02):
going on rants, but consumers. I think consumers feel squeezed.
They feel like these companies are just doing everything to profit,
the token maxing and minning, the sort of wars. I
just feel like if we yeah had this different sort
of open source ecosystem, it would be a radically different
tech landscape and a better one in my opinion.
Speaker 1 (11:21):
If you're the US government, I mean you're looking at
you're looking at the spaces and anthropic and open AI
as your kind of national champions essentially, right, And so
if you take away the ability by letting too much
distillation happen of these companies to protect their value.
Speaker 3 (11:36):
Like, but let's be clear, let's be clear, open source
like models are not only developed by distilling the anthropic
and open ai like this is what they're alleging, but like,
these open source models are also progressing at a significant rate.
This is why they're so obsessed with competing with China,
Like these open source models are getting better and better,
(11:56):
and it's not just because they're distilling anthropic and open AI, but.
Speaker 1 (11:59):
In your store or Natasha. Chinese researchers did some research
on this and it turns out that Quen, which is
Ali Barbara's AI model, this is, according to Chinese researchers,
misidentified itself as clawed in nearly one third of all cases.
I mean, that's pretty stunning.
Speaker 2 (12:15):
Yeah, that's true.
Speaker 4 (12:16):
But I will just say that, you know, I have
spoken to researchers who say when they test Quad and
other US models in Chinese and they do the same
thing and they say, who are you? It's as deep seek.
So I mean, like, this is what I'm trying to
say about. You know, it's really challenging when you use
this terminology that means just like a process, a methodology,
(12:40):
a technique, and imply that it's necessarily, you know, a
violation in some way. It might be breaking in terms
of service, which users do all the time and which
you know gets you kicked off of the platform, but
in some cases, like there is something different going on.
It's more of like jail breaking in a way. But
when you kind of use this term so broadly, it's
(13:02):
really hard to tell what exactly is happening and whether
some like whether it's really an egregious violation or just
something that the companies don't really want.
Speaker 1 (13:13):
Tell me about this phrase secret war in the in
the headline of your story, like how is the secret
war playing out? Like who are the combatants? And what
are they doing to each other?
Speaker 4 (13:23):
Well, I would say anthropic is one of the primary ones.
I mean when Taylor is asking about like why isn't
there a lot of open source. Anthropic was a big
voice in DC starting in the Biden administration, talking about
how dangerous open source was because you wouldn't be able
to check the downstream uses as closely, like if somebody
(13:43):
is using it to create a bioweapon or what have you,
because you can just download open source and then kind of,
you know, fork it, use it for yourself and the
company the developers of the open source technology won't necessarily
know what's happened. But I think we've also seen that
we can don't really rely on at least, you know,
when it came to how young people were using this technology,
(14:08):
you know, in terms of like therapy or some of
the things that we've seen with with the mental health issues,
we haven't seen like a robust reports from the companies
about downstream usage or even when it comes to you know,
how is how is it affecting jobs? It's not like
it's transparent to the public or to regulators or to
advocates how these models are being used downstream. I mean,
(14:31):
they put out these reports, but it's like quite filtered.
I would say, it's really hard when you dig into
methodology to really understand how people are using it. So
only the companies have access to that information. But anthropics
argument was very persuasive because you know, they're really worried
about the use of these models for uh, you know,
(14:52):
cyber warfare in some way, or or bioweapons or you know,
developing novel chemicals in some way, and those like once use.
Speaker 1 (15:03):
This is what secure Miss Hassabis talked about this week
right the CEO of DeepMind at Google, when he called
for a US led AI standards group to regulate frontier models,
which got quite a lot of buzz. He also talked
about quote fostering international collaborational key safety issues. I assume
he was talking about China.
Speaker 4 (15:20):
Yes, that's that's the way I read it as well. Yeah,
I mean it's kind of becoming this like a uniform
voice now from because we saw Jack Clark from Aanthropics
say something similar, Dario open Ai. Think what they're pushing
for is like potentially a market basedally, like you'll have
a third party person testing. It won't be through the government,
(15:44):
it would be a startup or a company that has
some independence.
Speaker 2 (15:49):
And yeah, I could.
Speaker 4 (15:50):
See more movement on this because the situation with Mythos
and fable has been so tumultuous and not good for business.
Speaker 1 (15:58):
And do you think that's a good thing, I mean
for the average person in the world if this like
safety initiative happens, or do you think the fact that
it's led by the industry itself automatically creates kind of
perverse incentives.
Speaker 4 (16:11):
I mean, I would say all tech policy in the
US is led by industry. They have you know, they
they have donated to all the nonprofits, they have the
biggest teams able to brief members of Congress. You know,
we just kind of naturally default to the tech executives
to tell us how to develop the technology. And you know,
(16:35):
lots of areas of government that need this kind of
technical expertise have been gutted. I would say there's like
suspicion obviously towards the tech companies, but I would say
all I mean, in the fifteen years that I've been
covering tech, it's all been guided by the companies themselves,
like from the beginning stage to kind of see the ideas,
and then once the law is written, you know, have
(16:57):
their lobbyist tweak things line to line.
Speaker 3 (16:59):
I think so terrifying because like Anthropic kind of reminds
me of Meta in this way where they will they're
obsessed with safety. They want to frame everything as doing
for safety. It's for safety for safety, and often what
those regulations ultimately do is cement their power and consolidate
power in the most powerful tech companies. I mean, I
(17:23):
don't know. The way that Entropic and sort of the
co founders have spoken about freedom of speech and access
to information is like deeply concerning. These are people that
you know, I just I don't trust them at all
to regulate these tech policies in ways that preserve our rights.
And I think that because they go out and they
(17:44):
do these pr tours and they say things that people
want to hear like oh, safety safety, child safety, whatever, whatever.
People somehow think that they're responsible, and I think that
they're not responsible. I think there's I don't know. To me,
it very much feels like this METAP playbook of like
going out and just saying a bunch of stuff and
meanwhile you're just using the rules to tighten control over
(18:05):
our information ecosystem, titan control over the market. It's just
like regulatory capture.
Speaker 4 (18:11):
I will say, it is a step above like what
was you know, pitch during the Biden EO and what's
happening now, which is like voluntary disclosure. So it's the
companies themselves doing these tests and like if you look
very closely at the you know, every time a new
model comes out, they use a different benchmark, they use
different standards. So even just having you know, I don't
(18:35):
want to like stand behind it, because what if we
find out, you know, the companies are like it just
the financial incentives are so skewed towards the companies that
it would be really hard for a third party. However,
I mean even just having a standard across the board
would make it a lot easier to compare models, you know,
to like for outsiders for the public to push for
(18:57):
different kinds of standards, like right now, we have no
say in how they're testing, what they're testing for, you know,
and and like the kinds of language.
Speaker 2 (19:06):
That they use, the methods they use.
Speaker 4 (19:08):
So at least if it were a third party, you know,
we would have that approach. But that, you know, doesn't
mean that it's better if it's a market based approach
versus Some of the testing that we heard was done
through Casey, the NIST organization, which is very very very
small budget but was apparently did like some great testing
(19:29):
of mythos maybe and the UK the UK Ai Safety Institute,
which was one of the first to like jail break
a lot of US models. I think, like a third party, yes,
and then I guess the rest kind of depends on
the on the fine print.
Speaker 1 (19:46):
Just on the topic of AI and China, Taylor, I
saw another story this week they made me think of you,
which is that China has moved to ban Chinese AI
companies developing chatbots that people establish comp onionship with or
any kind of emotional connection whatsoever. I'm curious if you
saw the story and what it made you think.
Speaker 3 (20:06):
Yes, they're specifically sort of trying to forbid chatbots that
are designed for companionship. So, as you mentioned, it's like
the chatbots that are like kind of like the character
AI type ones here where they're specifically designed to provide
a companion like experience, and so they're sort of built
(20:26):
to you know, give more maybe emotionally resident answers. They're
you know, designed to kind of help you navigate your
life problems, somebody to talk to. You know. This is
this is obviously like a very sort of controversial topic.
I think what's so concerning is is that you know,
China has extremely authoritarian uh sort of control over the
(20:50):
content that comes out of the tech ecosystem there, Like
they have a lot of government oversight. I don't think
that we have any evidence like this is not a
very evidence based approach, Like they're saying that they want
to ban it so that you know, basically so that
people have more babies. And I just I don't think
that the reason people are not having more babies in
(21:12):
China is related to the AI chatlots at all. I
think it's just again this is as you normally hear
with them sort of these types of tech laws. It's
not really based in science, it's based more in vibes.
Speaker 1 (21:24):
It is interesting those and I mean it's it's like,
you know, we have, you know, in the US all
this like handering about technology out of control and their
regulation and et cetera, et cetera, and then you know,
you have this policy in China, and it's so hard
to you know, between the between the between those two
paths figure out like how how should societies you know, Well, I.
Speaker 3 (21:45):
Think what's really scary and concerning is that we're seeing
America begin to regulate our information ecosystem increasingly like China.
So we continue to pass laws and actually you'll see
people that are ostensibly liberals praising China and extremely other
authoritarian regimes. And we have like Jonathan Heights, sort of
pseudoscience guy who's amassed a big audience here as a pundit,
(22:07):
praising countries like Malaysia and Indonesia. You know, these are
countries that don't have for free press, that arrest people
for their online speech, and often they're working with them.
I mean, we had even lawmakers praising the UAE and
their restrictions on sort of speech online and access to information.
So I think this is really worrying. There's also a
(22:27):
bunch of research that has come out around sort of
information and how sort of different llms criticize government. Basically,
in countries with lower levels of media freedom, when you
query models in the language of that country about politics
or the government or whatever, you are much more likely
to get a pro regime authoritarian answer than if you
(22:49):
ask in a different language. So these models, and this
came out in like a paper last month. Met As
Oversight Board is doing research into this. But basically what
we realize is that like training, data is shaped by
the sort of political ecosystem of that country. This has
consequences for LLM output. We need to make sure that
we are preserving our free and open information ecosystem here
(23:13):
in the US. And I don't think we're doing any
of that. And I think when we look at laws
like what China is doing and we sort of praise it,
it makes me very nervous.
Speaker 1 (23:28):
When we come back, Tayla tells us why all the
girls having a hot surveillance summer stay with us? Welcome back, Taylor.
Where does the phrase hot surveillance summer come from? What
(23:50):
does it mean?
Speaker 3 (23:52):
So it's actually from this podcaster Natalie, who hosts this
podcast Boys Club It's like a tech podcast. She wrote
this post about having a surveillance summer that was basically
pro Meta smart glasses. She came out right along when
sort of Meta began to launch this campaign with Kylie
Jenner and all the influencers are having, you know, their
(24:15):
sort of surveillance summer moment. Basically what she said is
like masurveillances everywhere, so why not participate in it?
Speaker 2 (24:22):
Which is I.
Speaker 3 (24:23):
Trust, it's horrifying. I hate it. But a lot of influencers,
if you look at the replies to her posts, were like, yes,
let's go. And you know, I was in New York
just last week and I saw the first pair of
Meta smart glasses in the wild on this girl. I
don't know if she was an influencer, but she was
really pretty, and I imagine that she was an influencer
because she had the early access to these glasses. But they're, yeah,
(24:46):
they're getting popular and people just kind of are treating
it as this like fun thing and they love the
Kylie glasses and yeah, they're talking about sort of proactively
having a hot surveillanceummer in like a positive way.
Speaker 1 (24:58):
Tally can you kind of lay out the history smart glasses.
I remember when I was a TV producer back in
like the twenty eleven period, Sebastian Thrun came on the
show I worked, only had this like Bionic Man prototype
of the Google glass that didn't quite catch on. Then
they were the first generation of Snapchat spectacles. Then metas
(25:18):
ray Band partnership kind of brought these somewhat into the mainstream.
Obviously Kylie is now has her own Meta ray Ban collab,
and then later this year or early next year, snap
had to going to release these new spectacles are apparently
going to cost two thousand dollars. I mean, what's the
kind of what's the technological history of these glasses? And
are they are there things I've actually I'm not sure
i've ever seen anyone actually wearing them in the wild
(25:40):
as well as I know.
Speaker 3 (25:41):
I've seen them unfortunately never Google glass. But but but yeah,
I mean these, I mean Meta RayBan glasses are popular.
I mean people have them. I've seen them in you
in LA. I think what's interesting to me, I mean
you just sort of like laid out the general progression
of a smart glas is I think the idea of
having something on your face that is constantly recording. That
(26:04):
can you know, act as this sort of second brain
for you, remind you of someone you can't remember, someone
you know, Like that's always been the sort of like
sci fi fantasy for years. We've had recording capabilities. So
I actually had the Snapchat spectacles back in twenty sixteen,
and you know, you could record very easily, you could
post a snapchat, so and that is sort of what
(26:24):
pioneered I think the popularity, the niche popularity, I guess,
of the original sort of meta rap ban glasses. What
makes this moment interesting to me and why I think
we're reaching a turning point in the normalization of these
products is that you now have sort of plausible deniability.
Previously they were only for recording, right, And so it's
kind of weird, like your people know that that's why
(26:46):
you're wearing them. It's like, what do you creep? Why
do you have to record? Why you know? Why can't
you just take out your phone like a normal person.
But two things have happened. One, we've actually normalized taking
out our phone and recording at all times. That's basically
been completely destigmatized ten years ago. That was actually still stigmatized.
Now there was like houn its like influencers in the
wild that would like make fun of people recording in public.
Now everyone records in public. Second of all is that
(27:09):
these glasses have AI integrated into them, and so they're
being sold as this like smart AI assistant. And so
the people that I spoke to in the replies of
this tweet that I was like, why would you do this,
They're like, well, I want AI. I want somebody to
see that everything that I see and help me. And
that is I think when I sort of see the
discussion around these consumer products, they're like, well, yeah, people
(27:31):
that record all day around them? Are you know, if
you're sort of like recording women with your glasses, you're
a greep. But I actually use them for help for AI.
And this is like part of the sort of girl
bossification of AI, right, Like I use my metal smart
glasses all day long and they help me figure out
what to buy at arrow on or you know, I
don't know, remember to shop stop by Alo or something.
(27:53):
I don't know. But it's yeah, which is I think
notable because when you look at who they've been giving
these glasses to, it is a lot of the like
West Village girlies, like lifestyle influencers, these aspirational women.
Speaker 1 (28:03):
Natasha, you've been flowing this story, yes.
Speaker 4 (28:05):
I mean like I of course went through the whole
Kylie photodump of the of the Meta glasses. And I
think when I moved to San Francisco was just when,
like maybe slightly after, people were being accosted in bars
for wearing the Google glasses, which are you know, like
extremely it looked you can tell when somebody was wearing
(28:27):
Google glasses, So I did see them here. I have
only seen the Metal ones in La one time, and
people were giving the guy a hard time. He was
on it was like on Abbot Kinney and you know,
it was just like what are you doing?
Speaker 2 (28:39):
U creep? But this was maybe in January, so a
while ago.
Speaker 4 (28:45):
Yeah, I mean, I understand the appeal of potentially having
something like as an always on assistant. I for some reason,
I always think about that scene in Devilwaar's Prada where
they're like whispering the name of the person to you.
(29:06):
I understand why we would want to feel like we're
the boss and we have you know, minions helping us.
Speaker 2 (29:13):
But I don't know. I mean at least when I
put on.
Speaker 4 (29:15):
The Google glasses, I remember too that it was just
it was not the seamless minority report experience that that
Sci Fi promised us.
Speaker 2 (29:23):
So I don't know.
Speaker 4 (29:26):
I mean, what do you think Taylor about, like how
how savvy they've been with targeting the right influencers. Do
you feel like going through women has been like, has
led to more adoption than they would have if they
tried to go like the Google glass route.
Speaker 3 (29:43):
Undeniably. I think, like number one, they've destigmatized the idea
that these are creep glasses used to surveil women, which
is what people sort of thought of metaglasses before. And
I think they've really made it aspirational. I think the
HIGHI partnership was brilliant. I think the sort of rollout
has been great, and I think you're seeing a lot
(30:03):
of women get reflexively defensive and be like, well I
want I mean I talked to my friend the other
day who wants to buy them because she has her kids,
and she's like, oh, it would be so great, Like
my hands are always full, I could just talk to
the glasses and also then I could record moments with
my children. I don't have to have my phone out
and like, I totally get it. I totally get it.
Speaker 4 (30:22):
Like I.
Speaker 3 (30:24):
Hate the mass surveillance aspect. I think like, if we
are going to normalize this level of masurveillance, let's have
like serious conversations around policy and privacy and privacy protections.
I think it's interesting as well. There was that Wired
report about that system called face Card that Meta was
going to ask you about this.
Speaker 1 (30:38):
This is interesting idea that the lasses were cross reference
with the database of all faces, so you could kind
of live identify people. But then they discontinued that, or
they kind of sum set it for the time being
off the wide article or what happened.
Speaker 3 (30:51):
Well sort of, And you know, Boz was on this
podcast with Nicholas Thompson, the editor of The Atlantic, a
Meta executive, and he's like, you know, I can't believe
there was so much backlash. Like, first of all, he
confirmed the story after Andy Stone, the comms person for Meta,
was like on Twitter all day denying it. This is nonsense,
this is nonsense. Then you have bas an executive going
(31:12):
on a podcast and being like, actually, we're really proud
of this work and it's actually really amazing. My favorite
part of that interview is that he's like, oh, people
think that there's just going to be some master database
and like it's just going to be pulling everyone for
the face databases. No, it's going to be pulling every
it's going to be pulling from a database that shows
you your connections, if you met that person once or whatever.
It's just like, Okay, that is such a meaningless distinction.
(31:33):
The point is that you are creating a database of
biometric data and facial recognition data and tying that to people,
and that is what we are mad about, you know.
Speaker 1 (31:43):
So it's amazing. I remember when Facebook first came out,
and I was at high school. Maybe not when it
first came out, but when it first started diffuse and
I was supposed sixteen or seventeen, I remember looking at
him thinking, oh my god, like there was soon not
in the not too in futury, possible to overlay everyone voluntary,
voluntarily kind of uploading themselves to this platform and just
having no expectation of privacy and people being able to
(32:04):
know who you are all the time. And it's like, wh.
Speaker 3 (32:07):
Boy, there, which, by the way, we normalize that, I mean,
I like, I mean, I did the same thing with Facebook,
like I would stock down everyone in my class, Like
I think I added everyone in my college dorm, you
know when I got into that. And so it's like
I think that, like people used to have a huge
amount of privacy, we have watched privacy be obliterated over
(32:30):
the past couple decades with the rise of social media.
Now we have AI which is super charging privacy violations.
I think that we are not having any conversations about
privacy like we should be having. We should be passing
privacy laws. Instead, we're passing the opposite. We're actually mandating.
Over twenty five states have mandated identity verification laws to
(32:52):
use Internet and then or to use parts of the Internet,
I should say. And then I think there's forty seven
states that are that are that have total either pass
these laws or are considering past them. And then we
have the Federal Kids Act discussions. So we're about to
mandate that Meta start harvesting biometric data, including on children
while they're building this technology. So if we want, if
(33:12):
we're really against surveillance, then we have to have a
bigger discussion about privacy.
Speaker 1 (33:18):
Just to close, Natasha, I want to ask you about
and There's two interesting stories I saw this week that
are kind of adjacent to this. One is Apple suing
open Ai for allegedly getting interviewing Apple execs and basically
using the interviews to steal Apple trade secrets around device manufacturing.
And the other was a report in Bloomberg about open
(33:39):
ai bringing its first device to market, which is not
a phone and not glasses. It sounds more like an
Alexa or a Sonos. But can you give us some
contexts on those two stories.
Speaker 4 (33:50):
Yeah, the Apple lawsuit was super interesting because it was
chock full of details about, you know, even having access
to people's texts and messages where they're like, lol, I
guess I still have access to this or I didn't
know you could, you know, bring a piece of an
Apple device out of the out of the company because
(34:13):
this long time Apple executive who moved to open ai
suggested that when you come to open ai, you bring
like an actual physical thing for a little show and
tell session. But one thing I thought was also really
interesting is Tony Fidel, the CEO of Nest. He reached
out after this newsletter writer Ben Thompson Etstra Tetri wrote
(34:39):
about it, and he said, this is how Apple scares
employees like this is I think what he was doing
was trying to frame it as not necessarily going after
open Ai, like you know, being sure that they'll they'll
be able to you know, get damages or prevent something
from open Ai, but rather as a message to Apple employees,
of which open ai has hired four hundreds.
Speaker 1 (35:00):
But these were texts onto Apple employees on their work phones,
which therefore Apple could read. Or how did Apple start
to like see this, Corsemond.
Speaker 4 (35:07):
Yeah, yeah, what like the a few of the texts
that they already had were on their Apple devices remaining
on their Apple devices, and they believe that, you know,
if they take it to discovery, they'll be able to
access even more. And they also mentioned this, you know,
they mentioned the Johnny I've partnership, the Johnny Ive company
that was that was acquired by open Ai, and they
(35:29):
were saying that you know, Johnny went to it was
something in particular about the metal finishing, like a proprietary
metal finishing that only Apple knows, and Johnny I've it
sounded like basically he's going to their suppliers and using
like Apple terminology, and so they had the impression this
was for an Apple product, And yeah, I guess we'll
(35:49):
see when this new Alexa comes out. I thought it
was supposed to be that tiny thing that we saw
like a few executives carrying around and get photographed with
like a compact almost with like in a dark metal finish,
with I think a little open AI logo. It looked
like a little clamshaw kind of thing and people assumed
(36:10):
that that was the device.
Speaker 2 (36:11):
But yeah, it sounds much less exciting.
Speaker 1 (36:16):
That's all we have time for this week. Taylor, Natasha,
thank you.
Speaker 2 (36:19):
Good to see you.
Speaker 3 (36:20):
Thanks for having us.
Speaker 1 (36:43):
For tech stuff. I'm os Voloshin. This episode was produced
by Eliza Dennis. It was executive produced by me and
Julian Nutta for Kaleidoscope and Katrina norvelf iHeart Podcasts. Jack
Insley makes this episode and Kyle Murdoch wrote our theme song.
A special thank you to Taylor Lorenz and Natasha Tiku.
Please check out all the work they put out into
(37:03):
the world. We're lucky to call them friends of the Pod.
Natasha Taylor. There was a fascinating story and Wired last
week about an internal memo Meta regarding a loose squirrel.
Speaker 2 (00:19):
I did see this, Yeah, I didn't understand.
Speaker 3 (00:22):
We had a bird get into the New York Times
office once and it did derail our afternoon.
Speaker 1 (00:26):
So this was this was a hard hard to pass.
But there was a squirrel mails to Meta in the Bangkok.
Speaker 3 (00:32):
Office, the poor squirrel.
Speaker 1 (00:34):
There was great hilarity because it's been a rough time
for Meta, but people were commenting on how this was
the most joyful moment the company had for a while,
despite the fact that Janet had got scratched on his finger.
Speaker 3 (00:45):
Well, I hope he got a Rabi's shot.
Speaker 1 (00:50):
Welcome to Tech Stuff. I'm was Vloscha and this is
the Week in Tech where I'm joined by the world's
most plugged in reporters to break down what's really happening
in tech right now. Today we're joined by Taylor, the
rends if use a mag and Natasha Tiku, tech reporter
at the Washington Post.
Speaker 3 (01:04):
Welcome both, Hey everybody, Hello, thanks for having us.
Speaker 1 (01:08):
It's good to see you.
Speaker 3 (01:09):
Natasha.
Speaker 1 (01:09):
You recently published an article titled Inside the Secret Ai
War between Silicon Valley and China. Great headline, but it
wasn't about ships or competing for talent, but actually about
the models themselves. Can you explain what's going on here?
Speaker 4 (01:25):
So this is something of a narrative war between the
US leading AI companies and the leading Chinese AI companies.
So what companies like Anthropic and Open Ai are alleging
is that these Chinese companies are accessing their models which
are not supposed to be available in China and Hong Kong.
(01:48):
They're accessing their models in order to do this process
called distillation, which basically involves kind of you know, distilling,
taking the learnings from a larger model and applying it
to us to improve a smaller model. And distillation is
a process that has been used in the AI industry
for a really long time. It's standard process. It's how
(02:08):
state of the art improvements are diffused throughout the system,
you know, through academia and open source. But Open Ai
and Anthropic have started using this term adversarial distillation and
they are like seeking government US government protection from this process,
and they're alleging, you know that a lot of the
(02:30):
gains that we're seeing in Chinese models in open source
AI in particular is coming at the you know, at
the expense of anthropic.
Speaker 1 (02:39):
Yeah, I have a serial distillation. It sounds like makes
me think of prohibition actually.
Speaker 4 (02:44):
But.
Speaker 2 (02:46):
Yeah, I was.
Speaker 4 (02:47):
I was at this National Security AI conference and the
person from open ai brought it up, and Irene's salimon
from hugging face like Brazer Hins. She was like, I've
never I've been in this indust for so long. She
was a super early open ai employee, and she was like,
I've never heard this term before, but now it's now.
It's kind of it is diffusing throughout the US because,
(03:10):
you know, they're extremely concerned that the Chinese open source
models are cheaper, more accessible, and are being adopted very
widely by US companies. Not that they the US companies
necessarily want to talk about it, but you know, particularly
around the restrictions around fable that happened with US government,
(03:32):
the inability to access mythos. You know, companies are like,
it's more reliable and more cost effective for me to
build on top of Chinese technology. And I think another
kind of distinction that's really been alided over is that
distillation might be a violation of your terms of service.
You know, if you use if you access the Anthropic API,
(03:54):
you are not supposed to try to ask questions twenty
seven million times and then and you know, use it
to improve your own model.
Speaker 1 (04:02):
What's the difference between what's happening right now in your
story and the deep Seek freak of early last year, Like,
is this kind of a continuation of the same story
or is it something slightly different?
Speaker 4 (04:13):
No, I would say it's a continuation of the same story.
In that case, it was open Ai alleging that, you
know that Deep Deep Seek did the distillation, but it
also has to do with the fact that the US
was really caught off guard by how good the Deep
Seek model was. And you know, many of the gains
that you're seeing in the Chinese market are you know,
(04:36):
like open Ai and Anthropic would like you to believe
that it's only because they're taking from US models, but
that's not true. In fact, like many of the you know,
speaking to two experts in the field, they say that
like many of the gains that the Chinese models have
made in efficiency, in size, cost effectiveness, those have all
(04:57):
come from the constraints on the industry, like the fact
that we have export controls on chips. You know, they
don't have access to as many as.
Speaker 1 (05:06):
Resources Jnsui Aguen, right, which is a necessity is the
mother invention and actually restricts exactly.
Speaker 4 (05:11):
I mean, you know this is like the startup way.
And obviously you know that if you look at like
how much money has been raised by the major AI labs,
that's not been a constraint for them, right, And uh,
you know what's happened recently is that a one of
the Anthropic users in Hong Kong discovered that Anthropic had
(05:33):
was using this kind of covert way to identify Chinese users.
They were looking at what time zone they're in. They
were looking at the you know, where they're logging in from.
Is it associated? Is that domain associated with the Chinese
AI lab? And then it's sending the information back to
Anthropic but in a way it like hidden in today's date.
(05:55):
So like the apostrophe in today's date said like you know,
Chinese user associated with minimax logging in in this time zone,
et cetera. And people were really upset about this, like.
Speaker 1 (06:09):
This is sort of a nightmare scenario for AI companies,
for users to feel like they're being actively surveiled, right,
I mean, Taylor, what's your I know that we're going
to talk about surveillance in the second half of the episode,
but what's your what's your take on the idea that
you know, Anthropic have been caught essentially spying on users
to try and figure out where they are and if
they're Chinese users, illicitly trying to distill the models.
Speaker 3 (06:31):
I think this whole thing is honestly, I don't want
to say a losing game, but it seems ridiculous and
I'm kind of curious to Tasha, like why why isn't
the US developing open source models? Like why why are
you know, why can't we have our own open source
that competes with these Chinese open source models.
Speaker 4 (06:49):
Well, now there is a lot more investment going into it,
but yeah, it wasn't like fine as financially lucrative. And
you know, even the you hear a lot of rhetoric
from the VCS, right like we we can't stop AI,
like we have to support open source. You weren't seeing
commensurate investment in open source. There are like a lot
(07:11):
of Alan Institute for AI hugging face.
Speaker 2 (07:16):
There was this project called the Adam Project.
Speaker 4 (07:19):
People were really interested in pushing this and having you know,
the US be the center of open source. But yeah,
there wasn't as much investment in it, and the vcs
are all backing the closed sourced companies, and so yeah,
this is the situation we now find ourselves in where
we're playing ketchup.
Speaker 3 (07:37):
Yeah, I feel like we're playing ketchup. And also it's
a losing battle. Like I mean, I know Alex Karp
was talking about this recently, and obviously he's self interested.
I think volunteer uses, maybe the open source models, but
but like it just seems like this idea of this
like closed source situation and like tokens and I don't
(07:58):
I don't know, like everything for profit like that Anthropic
and OPA are built on is just again it's just
a losing game because ultimately open source is going to win.
Speaker 1 (08:08):
I guess if Reid was here, if our play read,
because he's not here this week he's in Boston. But
I mean, I think what he would say is yes,
except that if you are leading edge company or startup
or lab, you need to have access to the best
models otherwise you will fall behind competitively. So like the
future general you say, I may be more and more
open source, but there will always be demand for like
(08:31):
mythos or whatever it is that you have to pay for.
Is that I mean? I guess what Red would say, Well.
Speaker 4 (08:37):
You can, I mean, you could build a state of
the art model on top of open source. You don't
need to start with Anthropic. You know, Anthropic could be
building on top of deep seek. I think, like, you know,
what's really interesting is that the US position, at least
under David Sachs, was that you want to build on
top of the American stack, like you want to make
(09:00):
sure just in the same way that you know, China
has diffused its power by working with you know, companies
in the Middle East and Africa. They were saying, you know,
we want the US, the US stack to be like
the layer on open source. But then the Trump administration
policies are really going against that. And actually Anthropic had said,
(09:20):
you know, they tried to argue that if the US
enforces you know, a stronger stance against distillation, that maybe
the US labs could maintain their lead for twelve to
twenty four months longer. And yeah, it's just it's uh
has led to exactly like what you're you, Taylor. You
mentioned Alex Carp's rant recently, and it's completely related because
(09:45):
companies are extremely frustrated. They feel like, you know, do
they even own the intellectual property that comes from using
your model? Why are they paying for tokens? Like it's
really had people even kind of rethinking the potential business
model and what the value is for the companies if
like they don't own the technology there. Yeah, it's just
(10:07):
I think they feel like it's highly inefficient and not
a good way to to like start the new AI
era building on this technology that could be turned off
by the US government or anthropic could change something, or
they're spying on you.
Speaker 3 (10:21):
Yeah, it seems like ridiculous. It seems like obviously we
should be moving towards an open source model. Obviously China
has the superior sort of structure to the AI environment. Also,
it's better for consumers, it's more affordable, it's better for
business consumers, better for I would say average consumers as well.
I think, I mean, I don't use deep Seek for
like tons of stuff, but I find it completely comparable
(10:42):
to the US you know, companies that are out there,
and I just I also wonder if if it would
help if we if we were investing more in open
source as well, if it would help sort of stem
some of the like curb some of the Antai backlash,
because I feel like so much of the Antai backlash
is not just from companies, like people like Alex Karp
(11:02):
going on rants, but consumers. I think consumers feel squeezed.
They feel like these companies are just doing everything to profit,
the token maxing and minning, the sort of wars. I
just feel like if we yeah had this different sort
of open source ecosystem, it would be a radically different
tech landscape and a better one in my opinion.
Speaker 1 (11:21):
If you're the US government, I mean you're looking at
you're looking at the spaces and anthropic and open AI
as your kind of national champions essentially, right, And so
if you take away the ability by letting too much
distillation happen of these companies to protect their value.
Speaker 3 (11:36):
Like, but let's be clear, let's be clear, open source
like models are not only developed by distilling the anthropic
and open ai like this is what they're alleging, but like,
these open source models are also progressing at a significant rate.
This is why they're so obsessed with competing with China,
Like these open source models are getting better and better,
(11:56):
and it's not just because they're distilling anthropic and open AI, but.
Speaker 1 (11:59):
In your store or Natasha. Chinese researchers did some research
on this and it turns out that Quen, which is
Ali Barbara's AI model, this is, according to Chinese researchers,
misidentified itself as clawed in nearly one third of all cases.
I mean, that's pretty stunning.
Speaker 2 (12:15):
Yeah, that's true.
Speaker 4 (12:16):
But I will just say that, you know, I have
spoken to researchers who say when they test Quad and
other US models in Chinese and they do the same
thing and they say, who are you? It's as deep seek.
So I mean, like, this is what I'm trying to
say about. You know, it's really challenging when you use
this terminology that means just like a process, a methodology,
(12:40):
a technique, and imply that it's necessarily, you know, a
violation in some way. It might be breaking in terms
of service, which users do all the time and which
you know gets you kicked off of the platform, but
in some cases, like there is something different going on.
It's more of like jail breaking in a way. But
when you kind of use this term so broadly, it's
(13:02):
really hard to tell what exactly is happening and whether
some like whether it's really an egregious violation or just
something that the companies don't really want.
Speaker 1 (13:13):
Tell me about this phrase secret war in the in
the headline of your story, like how is the secret
war playing out? Like who are the combatants? And what
are they doing to each other?
Speaker 4 (13:23):
Well, I would say anthropic is one of the primary ones.
I mean when Taylor is asking about like why isn't
there a lot of open source. Anthropic was a big
voice in DC starting in the Biden administration, talking about
how dangerous open source was because you wouldn't be able
to check the downstream uses as closely, like if somebody
(13:43):
is using it to create a bioweapon or what have you,
because you can just download open source and then kind of,
you know, fork it, use it for yourself and the
company the developers of the open source technology won't necessarily
know what's happened. But I think we've also seen that
we can don't really rely on at least, you know,
when it came to how young people were using this technology,
(14:08):
you know, in terms of like therapy or some of
the things that we've seen with with the mental health issues,
we haven't seen like a robust reports from the companies
about downstream usage or even when it comes to you know,
how is how is it affecting jobs? It's not like
it's transparent to the public or to regulators or to
advocates how these models are being used downstream. I mean,
(14:31):
they put out these reports, but it's like quite filtered.
I would say, it's really hard when you dig into
methodology to really understand how people are using it. So
only the companies have access to that information. But anthropics
argument was very persuasive because you know, they're really worried
about the use of these models for uh, you know,
(14:52):
cyber warfare in some way, or or bioweapons or you know,
developing novel chemicals in some way, and those like once use.
Speaker 1 (15:03):
This is what secure Miss Hassabis talked about this week
right the CEO of DeepMind at Google, when he called
for a US led AI standards group to regulate frontier models,
which got quite a lot of buzz. He also talked
about quote fostering international collaborational key safety issues. I assume
he was talking about China.
Speaker 4 (15:20):
Yes, that's that's the way I read it as well. Yeah,
I mean it's kind of becoming this like a uniform
voice now from because we saw Jack Clark from Aanthropics
say something similar, Dario open Ai. Think what they're pushing
for is like potentially a market basedally, like you'll have
a third party person testing. It won't be through the government,
(15:44):
it would be a startup or a company that has
some independence.
Speaker 2 (15:49):
And yeah, I could.
Speaker 4 (15:50):
See more movement on this because the situation with Mythos
and fable has been so tumultuous and not good for business.
Speaker 1 (15:58):
And do you think that's a good thing, I mean
for the average person in the world if this like
safety initiative happens, or do you think the fact that
it's led by the industry itself automatically creates kind of
perverse incentives.
Speaker 4 (16:11):
I mean, I would say all tech policy in the
US is led by industry. They have you know, they
they have donated to all the nonprofits, they have the
biggest teams able to brief members of Congress. You know,
we just kind of naturally default to the tech executives
to tell us how to develop the technology. And you know,
(16:35):
lots of areas of government that need this kind of
technical expertise have been gutted. I would say there's like
suspicion obviously towards the tech companies, but I would say
all I mean, in the fifteen years that I've been
covering tech, it's all been guided by the companies themselves,
like from the beginning stage to kind of see the ideas,
and then once the law is written, you know, have
(16:57):
their lobbyist tweak things line to line.
Speaker 3 (16:59):
I think so terrifying because like Anthropic kind of reminds
me of Meta in this way where they will they're
obsessed with safety. They want to frame everything as doing
for safety. It's for safety for safety, and often what
those regulations ultimately do is cement their power and consolidate
power in the most powerful tech companies. I mean, I
(17:23):
don't know. The way that Entropic and sort of the
co founders have spoken about freedom of speech and access
to information is like deeply concerning. These are people that
you know, I just I don't trust them at all
to regulate these tech policies in ways that preserve our rights.
And I think that because they go out and they
(17:44):
do these pr tours and they say things that people
want to hear like oh, safety safety, child safety, whatever, whatever.
People somehow think that they're responsible, and I think that
they're not responsible. I think there's I don't know. To me,
it very much feels like this METAP playbook of like
going out and just saying a bunch of stuff and
meanwhile you're just using the rules to tighten control over
(18:05):
our information ecosystem, titan control over the market. It's just
like regulatory capture.
Speaker 4 (18:11):
I will say, it is a step above like what
was you know, pitch during the Biden EO and what's
happening now, which is like voluntary disclosure. So it's the
companies themselves doing these tests and like if you look
very closely at the you know, every time a new
model comes out, they use a different benchmark, they use
different standards. So even just having you know, I don't
(18:35):
want to like stand behind it, because what if we
find out, you know, the companies are like it just
the financial incentives are so skewed towards the companies that
it would be really hard for a third party. However,
I mean even just having a standard across the board
would make it a lot easier to compare models, you know,
to like for outsiders for the public to push for
(18:57):
different kinds of standards, like right now, we have no
say in how they're testing, what they're testing for, you know,
and and like the kinds of language.
Speaker 2 (19:06):
That they use, the methods they use.
Speaker 4 (19:08):
So at least if it were a third party, you know,
we would have that approach. But that, you know, doesn't
mean that it's better if it's a market based approach
versus Some of the testing that we heard was done
through Casey, the NIST organization, which is very very very
small budget but was apparently did like some great testing
(19:29):
of mythos maybe and the UK the UK Ai Safety Institute,
which was one of the first to like jail break
a lot of US models. I think, like a third party, yes,
and then I guess the rest kind of depends on
the on the fine print.
Speaker 1 (19:46):
Just on the topic of AI and China, Taylor, I
saw another story this week they made me think of you,
which is that China has moved to ban Chinese AI
companies developing chatbots that people establish comp onionship with or
any kind of emotional connection whatsoever. I'm curious if you
saw the story and what it made you think.
Speaker 3 (20:06):
Yes, they're specifically sort of trying to forbid chatbots that
are designed for companionship. So, as you mentioned, it's like
the chatbots that are like kind of like the character
AI type ones here where they're specifically designed to provide
a companion like experience, and so they're sort of built
(20:26):
to you know, give more maybe emotionally resident answers. They're
you know, designed to kind of help you navigate your
life problems, somebody to talk to. You know. This is
this is obviously like a very sort of controversial topic.
I think what's so concerning is is that you know,
China has extremely authoritarian uh sort of control over the
(20:50):
content that comes out of the tech ecosystem there, Like
they have a lot of government oversight. I don't think
that we have any evidence like this is not a
very evidence based approach, Like they're saying that they want
to ban it so that you know, basically so that
people have more babies. And I just I don't think
that the reason people are not having more babies in
(21:12):
China is related to the AI chatlots at all. I
think it's just again this is as you normally hear
with them sort of these types of tech laws. It's
not really based in science, it's based more in vibes.
Speaker 1 (21:24):
It is interesting those and I mean it's it's like,
you know, we have, you know, in the US all
this like handering about technology out of control and their
regulation and et cetera, et cetera, and then you know,
you have this policy in China, and it's so hard
to you know, between the between the between those two
paths figure out like how how should societies you know, Well, I.
Speaker 3 (21:45):
Think what's really scary and concerning is that we're seeing
America begin to regulate our information ecosystem increasingly like China.
So we continue to pass laws and actually you'll see
people that are ostensibly liberals praising China and extremely other
authoritarian regimes. And we have like Jonathan Heights, sort of
pseudoscience guy who's amassed a big audience here as a pundit,
(22:07):
praising countries like Malaysia and Indonesia. You know, these are
countries that don't have for free press, that arrest people
for their online speech, and often they're working with them.
I mean, we had even lawmakers praising the UAE and
their restrictions on sort of speech online and access to information.
So I think this is really worrying. There's also a
(22:27):
bunch of research that has come out around sort of
information and how sort of different llms criticize government. Basically,
in countries with lower levels of media freedom, when you
query models in the language of that country about politics
or the government or whatever, you are much more likely
to get a pro regime authoritarian answer than if you
(22:49):
ask in a different language. So these models, and this
came out in like a paper last month. Met As
Oversight Board is doing research into this. But basically what
we realize is that like training, data is shaped by
the sort of political ecosystem of that country. This has
consequences for LLM output. We need to make sure that
we are preserving our free and open information ecosystem here
(23:13):
in the US. And I don't think we're doing any
of that. And I think when we look at laws
like what China is doing and we sort of praise it,
it makes me very nervous.
Speaker 1 (23:28):
When we come back, Tayla tells us why all the
girls having a hot surveillance summer stay with us? Welcome back, Taylor.
Where does the phrase hot surveillance summer come from? What
(23:50):
does it mean?
Speaker 3 (23:52):
So it's actually from this podcaster Natalie, who hosts this
podcast Boys Club It's like a tech podcast. She wrote
this post about having a surveillance summer that was basically
pro Meta smart glasses. She came out right along when
sort of Meta began to launch this campaign with Kylie
Jenner and all the influencers are having, you know, their
(24:15):
sort of surveillance summer moment. Basically what she said is
like masurveillances everywhere, so why not participate in it?
Speaker 2 (24:22):
Which is I.
Speaker 3 (24:23):
Trust, it's horrifying. I hate it. But a lot of influencers,
if you look at the replies to her posts, were like, yes,
let's go. And you know, I was in New York
just last week and I saw the first pair of
Meta smart glasses in the wild on this girl. I
don't know if she was an influencer, but she was
really pretty, and I imagine that she was an influencer
because she had the early access to these glasses. But they're, yeah,
(24:46):
they're getting popular and people just kind of are treating
it as this like fun thing and they love the
Kylie glasses and yeah, they're talking about sort of proactively
having a hot surveillanceummer in like a positive way.
Speaker 1 (24:58):
Tally can you kind of lay out the history smart glasses.
I remember when I was a TV producer back in
like the twenty eleven period, Sebastian Thrun came on the
show I worked, only had this like Bionic Man prototype
of the Google glass that didn't quite catch on. Then
they were the first generation of Snapchat spectacles. Then metas
(25:18):
ray Band partnership kind of brought these somewhat into the mainstream.
Obviously Kylie is now has her own Meta ray Ban collab,
and then later this year or early next year, snap
had to going to release these new spectacles are apparently
going to cost two thousand dollars. I mean, what's the
kind of what's the technological history of these glasses? And
are they are there things I've actually I'm not sure
i've ever seen anyone actually wearing them in the wild
(25:40):
as well as I know.
Speaker 3 (25:41):
I've seen them unfortunately never Google glass. But but but yeah,
I mean these, I mean Meta RayBan glasses are popular.
I mean people have them. I've seen them in you
in LA. I think what's interesting to me, I mean
you just sort of like laid out the general progression
of a smart glas is I think the idea of
having something on your face that is constantly recording. That
(26:04):
can you know, act as this sort of second brain
for you, remind you of someone you can't remember, someone
you know, Like that's always been the sort of like
sci fi fantasy for years. We've had recording capabilities. So
I actually had the Snapchat spectacles back in twenty sixteen,
and you know, you could record very easily, you could
post a snapchat, so and that is sort of what
(26:24):
pioneered I think the popularity, the niche popularity, I guess,
of the original sort of meta rap ban glasses. What
makes this moment interesting to me and why I think
we're reaching a turning point in the normalization of these
products is that you now have sort of plausible deniability.
Previously they were only for recording, right, And so it's
kind of weird, like your people know that that's why
(26:46):
you're wearing them. It's like, what do you creep? Why
do you have to record? Why you know? Why can't
you just take out your phone like a normal person.
But two things have happened. One, we've actually normalized taking
out our phone and recording at all times. That's basically
been completely destigmatized ten years ago. That was actually still stigmatized.
Now there was like houn its like influencers in the
wild that would like make fun of people recording in public.
Now everyone records in public. Second of all is that
(27:09):
these glasses have AI integrated into them, and so they're
being sold as this like smart AI assistant. And so
the people that I spoke to in the replies of
this tweet that I was like, why would you do this,
They're like, well, I want AI. I want somebody to
see that everything that I see and help me. And
that is I think when I sort of see the
discussion around these consumer products, they're like, well, yeah, people
(27:31):
that record all day around them? Are you know, if
you're sort of like recording women with your glasses, you're
a greep. But I actually use them for help for AI.
And this is like part of the sort of girl
bossification of AI, right, Like I use my metal smart
glasses all day long and they help me figure out
what to buy at arrow on or you know, I
don't know, remember to shop stop by Alo or something.
(27:53):
I don't know. But it's yeah, which is I think
notable because when you look at who they've been giving
these glasses to, it is a lot of the like
West Village girlies, like lifestyle influencers, these aspirational women.
Speaker 1 (28:03):
Natasha, you've been flowing this story, yes.
Speaker 4 (28:05):
I mean like I of course went through the whole
Kylie photodump of the of the Meta glasses. And I
think when I moved to San Francisco was just when,
like maybe slightly after, people were being accosted in bars
for wearing the Google glasses, which are you know, like
extremely it looked you can tell when somebody was wearing
(28:27):
Google glasses, So I did see them here. I have
only seen the Metal ones in La one time, and
people were giving the guy a hard time. He was
on it was like on Abbot Kinney and you know,
it was just like what are you doing?
Speaker 2 (28:39):
U creep? But this was maybe in January, so a
while ago.
Speaker 4 (28:45):
Yeah, I mean, I understand the appeal of potentially having
something like as an always on assistant. I for some reason,
I always think about that scene in Devilwaar's Prada where
they're like whispering the name of the person to you.
(29:06):
I understand why we would want to feel like we're
the boss and we have you know, minions helping us.
Speaker 2 (29:13):
But I don't know. I mean at least when I
put on.
Speaker 4 (29:15):
The Google glasses, I remember too that it was just
it was not the seamless minority report experience that that
Sci Fi promised us.
Speaker 2 (29:23):
So I don't know.
Speaker 4 (29:26):
I mean, what do you think Taylor about, like how
how savvy they've been with targeting the right influencers. Do
you feel like going through women has been like, has
led to more adoption than they would have if they
tried to go like the Google glass route.
Speaker 3 (29:43):
Undeniably. I think, like number one, they've destigmatized the idea
that these are creep glasses used to surveil women, which
is what people sort of thought of metaglasses before. And
I think they've really made it aspirational. I think the
HIGHI partnership was brilliant. I think the sort of rollout
has been great, and I think you're seeing a lot
(30:03):
of women get reflexively defensive and be like, well I
want I mean I talked to my friend the other
day who wants to buy them because she has her kids,
and she's like, oh, it would be so great, Like
my hands are always full, I could just talk to
the glasses and also then I could record moments with
my children. I don't have to have my phone out
and like, I totally get it. I totally get it.
Speaker 4 (30:22):
Like I.
Speaker 3 (30:24):
Hate the mass surveillance aspect. I think like, if we
are going to normalize this level of masurveillance, let's have
like serious conversations around policy and privacy and privacy protections.
I think it's interesting as well. There was that Wired
report about that system called face Card that Meta was
going to ask you about this.
Speaker 1 (30:38):
This is interesting idea that the lasses were cross reference
with the database of all faces, so you could kind
of live identify people. But then they discontinued that, or
they kind of sum set it for the time being
off the wide article or what happened.
Speaker 3 (30:51):
Well sort of, And you know, Boz was on this
podcast with Nicholas Thompson, the editor of The Atlantic, a
Meta executive, and he's like, you know, I can't believe
there was so much backlash. Like, first of all, he
confirmed the story after Andy Stone, the comms person for Meta,
was like on Twitter all day denying it. This is nonsense,
this is nonsense. Then you have bas an executive going
(31:12):
on a podcast and being like, actually, we're really proud
of this work and it's actually really amazing. My favorite
part of that interview is that he's like, oh, people
think that there's just going to be some master database
and like it's just going to be pulling everyone for
the face databases. No, it's going to be pulling every
it's going to be pulling from a database that shows
you your connections, if you met that person once or whatever.
It's just like, Okay, that is such a meaningless distinction.
(31:33):
The point is that you are creating a database of
biometric data and facial recognition data and tying that to people,
and that is what we are mad about, you know.
Speaker 1 (31:43):
So it's amazing. I remember when Facebook first came out,
and I was at high school. Maybe not when it
first came out, but when it first started diffuse and
I was supposed sixteen or seventeen, I remember looking at
him thinking, oh my god, like there was soon not
in the not too in futury, possible to overlay everyone voluntary,
voluntarily kind of uploading themselves to this platform and just
having no expectation of privacy and people being able to
(32:04):
know who you are all the time. And it's like, wh.
Speaker 3 (32:07):
Boy, there, which, by the way, we normalize that, I mean,
I like, I mean, I did the same thing with Facebook,
like I would stock down everyone in my class, Like
I think I added everyone in my college dorm, you
know when I got into that. And so it's like
I think that, like people used to have a huge
amount of privacy, we have watched privacy be obliterated over
(32:30):
the past couple decades with the rise of social media.
Now we have AI which is super charging privacy violations.
I think that we are not having any conversations about
privacy like we should be having. We should be passing
privacy laws. Instead, we're passing the opposite. We're actually mandating.
Over twenty five states have mandated identity verification laws to
(32:52):
use Internet and then or to use parts of the Internet,
I should say. And then I think there's forty seven
states that are that are that have total either pass
these laws or are considering past them. And then we
have the Federal Kids Act discussions. So we're about to
mandate that Meta start harvesting biometric data, including on children
while they're building this technology. So if we want, if
(33:12):
we're really against surveillance, then we have to have a
bigger discussion about privacy.
Speaker 1 (33:18):
Just to close, Natasha, I want to ask you about
and There's two interesting stories I saw this week that
are kind of adjacent to this. One is Apple suing
open Ai for allegedly getting interviewing Apple execs and basically
using the interviews to steal Apple trade secrets around device manufacturing.
And the other was a report in Bloomberg about open
(33:39):
ai bringing its first device to market, which is not
a phone and not glasses. It sounds more like an
Alexa or a Sonos. But can you give us some
contexts on those two stories.
Speaker 4 (33:50):
Yeah, the Apple lawsuit was super interesting because it was
chock full of details about, you know, even having access
to people's texts and messages where they're like, lol, I
guess I still have access to this or I didn't
know you could, you know, bring a piece of an
Apple device out of the out of the company because
(34:13):
this long time Apple executive who moved to open ai
suggested that when you come to open ai, you bring
like an actual physical thing for a little show and
tell session. But one thing I thought was also really
interesting is Tony Fidel, the CEO of Nest. He reached
out after this newsletter writer Ben Thompson Etstra Tetri wrote
(34:39):
about it, and he said, this is how Apple scares
employees like this is I think what he was doing
was trying to frame it as not necessarily going after
open Ai, like you know, being sure that they'll they'll
be able to you know, get damages or prevent something
from open Ai, but rather as a message to Apple employees,
of which open ai has hired four hundreds.
Speaker 1 (35:00):
But these were texts onto Apple employees on their work phones,
which therefore Apple could read. Or how did Apple start
to like see this, Corsemond.
Speaker 4 (35:07):
Yeah, yeah, what like the a few of the texts
that they already had were on their Apple devices remaining
on their Apple devices, and they believe that, you know,
if they take it to discovery, they'll be able to
access even more. And they also mentioned this, you know,
they mentioned the Johnny I've partnership, the Johnny Ive company
that was that was acquired by open Ai, and they
(35:29):
were saying that you know, Johnny went to it was
something in particular about the metal finishing, like a proprietary
metal finishing that only Apple knows, and Johnny I've it
sounded like basically he's going to their suppliers and using
like Apple terminology, and so they had the impression this
was for an Apple product, And yeah, I guess we'll
(35:49):
see when this new Alexa comes out. I thought it
was supposed to be that tiny thing that we saw
like a few executives carrying around and get photographed with
like a compact almost with like in a dark metal finish,
with I think a little open AI logo. It looked
like a little clamshaw kind of thing and people assumed
(36:10):
that that was the device.
Speaker 2 (36:11):
But yeah, it sounds much less exciting.
Speaker 1 (36:16):
That's all we have time for this week. Taylor, Natasha,
thank you.
Speaker 2 (36:19):
Good to see you.
Speaker 3 (36:20):
Thanks for having us.
Speaker 1 (36:43):
For tech stuff. I'm os Voloshin. This episode was produced
by Eliza Dennis. It was executive produced by me and
Julian Nutta for Kaleidoscope and Katrina norvelf iHeart Podcasts. Jack
Insley makes this episode and Kyle Murdoch wrote our theme song.
A special thank you to Taylor Lorenz and Natasha Tiku.
Please check out all the work they put out into
(37:03):
the world. We're lucky to call them friends of the Pod.
TechStuff News
Hosts And Creators
Oz Woloshyn
Karah Preiss
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
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.