August 22, 2025 • 30 mins
How will we send memes when the world ends? This week, Oz and Karah recap the first ever World Humanoid Robot Games – the Olympics for Androids. Then, Oz explains why the US might be putting secret tracking devices inside shipments of advanced microchips. Karah explains a texting network built for the Apocalypse. And finally, on Chat and Me, how TikTokers are using ChatGPT to inspire their makeovers.
Also, we want to hear from you: If you’ve used a chatbot in a surprising or delightful (or deranged) way, send us a 1–2 minute voice note at techstuffpodcast@gmail.com.
See omnystudio.com/listener for privacy information.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:13):
From Kaleidoscope and iHeart podcasts. This is tech stuff. I'm
Os Voloscian and I'm care Price. Today we're going to
get into a communications network built for the end of
the world, and why the US is placing tracking devices
in AI chip shipments. Then on chat to me your
next makeover brought to you by AI. All of that
(00:35):
on the Weekend Tech. It's Friday, August twenty second. Hello,
Hello Cara.
Speaker 2 (00:48):
Hi auhs. Do you remember when we talked about the
half marathon between humans and robots.
Speaker 1 (00:54):
The one in China? I remember, yes, right.
Speaker 2 (00:58):
There are now robot Olympians and I'm talking about the
humanoid robot games in Beijing.
Speaker 1 (01:06):
I didn't know about this, but it does make me
think of my childhood as an only child. In England
in the late nineties there were you I know, there
are originally four terrestrial TV channels, and then a fifth
was launch called Channel five, and Channel five had slightly
more racy content, including at seven pm every Friday night
a show called Robot Wars where various middle aged hobbyists
(01:31):
would have their extremely slow robots bump into each other
and then break. But the commentator really brought the energy
and so it was. It was better than whatever else
was on TV for me, at least as a child.
That's a long rambling response. Is what you're talking about?
Speaker 2 (01:46):
You had the viewing taste of a middle aged British man.
Speaker 1 (01:49):
I did, and I still do.
Speaker 2 (01:50):
It's not it's a little bit what I'm talking about.
Let me give a little bit of context. So these
games took place over three days in Beijing, and they
actually used some of the facilities that were built for
the twenty twenty two Winter Olympics. There were two hundred
and eighty teams from sixteen countries, including the US. The
competitions include soccer, running, kickboxing, which is sort of like
(02:12):
what you were talking about, basketball, table tennis, and then
strangely enough, factory work and warehouse warehouse navigation robots.
Speaker 1 (02:22):
Were like, we got this, truly some must watch sports.
That's an interesting interesting They kind of brought the hype
with the kickboxing, but then what they're really flexing was
we got the rick and wac pick and pack. We
got pick and pack down.
Speaker 2 (02:35):
To a t. Here we go. Yeah, I mean the
whole thing was definitely designed to wow. The opening ceremony,
as opening ceremonies do included dozens of robots dancing, doing
martial arts, playing drums and sink which you might remember
from the two thousand and eight Beijing opening ceremony. I
think it was sort of like, look, the robots can
do what humans did at the Olympics. Let me just
(02:57):
show you a video because it's inexplicable.
Speaker 1 (03:00):
We have a real man playing a piano facing a
robot in a silver cape also playing.
Speaker 2 (03:08):
Oh, and here comes a robot in a bucket hat. Oh.
Speaker 1 (03:11):
Now there's real kids jumping around behind the bucket hat robots.
Speaker 2 (03:15):
And they're dancing. Some of the robots don't seem to
be complying very well.
Speaker 1 (03:18):
This is a surprisingly low rent. In any case, the
production values are not as high and are not that
much different Channel five, nineteen ninety eight. I mean there
is a sense of evolution from the TV show of
my childhood. I'm kind of surprised, but I actually thought
that it would be more wow factor.
Speaker 2 (03:34):
Yeah, you know, I watched a little bit more of
the video and it definitely left a bit to be
desired across the board, you know, in the track and field,
and android ran full speed into one of the robot operators,
knocking him down and then he just kept running, which
you know, when they say the robots are taking over,
that's what I imagine. Everybody was fine, though, don't worry.
And the boxing robots, let's just say they miss their target.
(03:58):
They managed to barely punch each other at all, flailing
through the air until the refs just gave up and
were like, this one wins.
Speaker 1 (04:05):
Whatever I saw a Ziemaza has a great newsletter called
Exponential View, and he put something on his LinkedIn my
favorite source of truth saying basically, this is a cute story,
but watch out because by twenty thirty it's going to
be a thirty five to fifty billion dollar industry. What
kind of reactions do you see.
Speaker 2 (04:23):
Well, it's obvious that there's still a long way to
go before robots are capable of outperforming elite athletes or
outperforming elite dance squads. But that's sort of besides the point.
Those participating in the games were really there to collect data.
The robot games were a great way for these teams
to keep developing robots for more practical applications like factory work.
But as one robotics professor said to The New York Times,
(04:46):
robots are still dumb, haush.
Speaker 1 (04:50):
You know, these games are interesting in the fact their
station in Beijing and that the US participated. It does
kind of bring to mind this Cold War posturing around
the space and the nuclear race, and robotics is obviously
a frontier of the tech race between the US and
China and an area where China seems to have the
(05:10):
upper hand, which is probably why these mass spectacle events
are happening so regularly, because they're great memes and they
remind the world of that.
Speaker 2 (05:18):
You know, we talk about the AI race a lot
between the two countries, and I think the feeling is
that the US still has the upper hand.
Speaker 1 (05:25):
I think that's the narrative that the US has this edge.
Some people say it's three months, some people say it's
six months. Some people say it's twelve months. It's a
small edge. I interviewed Jake Sullivan, the National Security Advisor
under President Biden, a few months ago. As you probably remember,
he is the guy who designed the US policy to
restrict China's access to advanced chips in order to maintain
(05:49):
the US edge and AI, and that policy was rolled
out back in twenty twenty two.
Speaker 2 (05:52):
Yeah, I remember that interview and He told you that
despite the launch of deep Seek, which is a very
efficient Chinese AI model, he felt confident that his policy
had been the correct one, and he was worried about
what it might mean for America's edge if it were
actually rolled back.
Speaker 1 (06:06):
That's right, and actually to get around that policy in
Nvidia designed a special chip specially for the Chinese market
called the H twenty, and it was weird as it sounds,
essentially designed to be less good than the top of
the line chip so that it could comply with these
US laws. The H twenty started shipping in early twenty
(06:26):
twenty four, but in April Trump actually went further than
the Biden administration and banned the export of AGE twenty chips.
But of course new developments late last month, President Trump
changed his mind and his administration decided that Nvidia can
now export the H twenty chips to China, provided they
(06:48):
pay a fifteen percent tax on all revenues of those
chips sold in China. There is interesting twist to this
saga of the chips and oxport controls. Back in the seventies,
the United States in bargoed Iranian oil, but of course
shippers and commodity traders figured out how to get past
the authorities with Iranian oil. And in a kind of
(07:10):
parallel of that, Reuter's reports that a booming trade has
opened up around smuggling the most advanced chips to China,
with smuggled shipments going through third party countries like Singapore, Malaysia,
and the United Arab Emirates. Roy just did run a
story about how the US may be starting to crack
down on this smuggling. It alleges US authorities have been
secretly putting location tracking technology into advanced chip shipments, shipments
(07:36):
they fear might be diverted to China.
Speaker 2 (07:39):
So this is like dropping an air tag in your
girlfriend's purse to find out how she's cheating on you.
Speaker 1 (07:44):
Yeah, I mean, I can't think of a better parallel.
Speaker 2 (07:47):
And so this is all in response to the export controls.
Speaker 1 (07:51):
That's right, because even though as I mentioned, the export
controls have recently been lifted on the Age twenty chips,
there is still huge demand for the most advanced H
one hun undred chips in China. And again, Jake Sullivan,
when who's talking about the deep Seek moment, argue that
that likely only happened because of the number of advanced
chips that have already made it to China or that
were being smuggled into China.
Speaker 2 (08:12):
But I'm confused, like, what is tracking these chips even
going to accomplish. Isn't it already too late to get
them back?
Speaker 1 (08:18):
Yes, but of course to build more and more advanced
AI models, you need more and more chips. So it's
a little bit like bolting the stable after the horse
has gone to your point. But it's also a little
bit like making sure no one leaves the stable door
open again and the rest of the horses stay on
the paddock. An anonymous source told Royce's that the trackers
are basically a way of starting to build legal cases
(08:40):
against companies that might be breaking the export restrictions.
Speaker 2 (08:43):
And do we know who's like embedding these trackers.
Speaker 1 (08:46):
No, we don't know that. It could be Homeland security investigations,
it could be the FBI, could be the Commerce Department.
Sources told Royce's the trackers are hidden in the packaging
of the shipments, not the chips themselves, but they don't
know exactly in which part of the shipping process they're
actually getting embedded. They also don't know when this started
or to what extent the trackers are part of an
(09:08):
active investigation. Of course, law enforcement agencies and chip manufacturers
did not rush to comment on the story.
Speaker 2 (09:14):
I was gonna say, it could be someone's ex We
have no idea. I mean, whose China is that? I
mean the US government probably, you know, I guess. One
of the tricks of effective law enforcement is to make
people believe you are all knowing and all seeing, to
deter people from even thinking about the pen optacon precisely
the panopticon. But controlling shipments in international waters via third
party countries is not an easy feat.
Speaker 1 (09:36):
It's not easy. It's not easy. But it's an interesting
moment because it's just another example of how the chip
wars are running so hot right now. I was scrolling
daily Mail this weekend, and to my surprise, the top
story on dailymail dot com was about chips.
Speaker 2 (09:51):
You want to know why?
Speaker 1 (09:52):
Why?
Speaker 2 (09:52):
Because they love a cat and mouse game. You know
what this is? This is petty drama. That's why it's
on the Daily Mail.
Speaker 1 (09:59):
Well, funny enough, the headline was Trump launches Manhattan Project
as one of America's largest companies is set to be nationalized.
The stories about how the US government is considering taking
a ten percent stake in Intel, which coursed a private company.
That deal would make the government the company's largest shareholder
and a ten billion dollar plus deal.
Speaker 2 (10:20):
So what's going on here?
Speaker 1 (10:21):
Well, obviously it's not a secret that Intel's been floundering
a bit. They've fallen behind in chip manufacturing. They've lost
over half their market cap in less than two years,
and they've spent apparently forty billion dollars or almost forty
billion dollars over the past three years trying to compete
with the Taiwan semiconductor manufacturing Company TSMC TSMC.
Speaker 2 (10:43):
But why was this idea even floated to begin with?
Speaker 1 (10:46):
Well, I think it's not standard Republican policy to nationalized industries.
But Trump, I think, is invested in this idea of
national champions, like companies that punch above their weight and
kind of achieve America's goals.
Speaker 2 (10:58):
Abroad, something that could be said in any UFC title
always will appeal to Trump.
Speaker 1 (11:02):
Yeah, I think that's true. So, according to a story
in the Wall Street Journal, the President met with Intel
CEO Bhutan about a week ago, and since then, Intel's
been in discussions with the White House about making this deal. Happen.
And ironically, this is in spite of the fact that
Trump had previously been calling on Tan, the CEO, to
(11:24):
be fired because of his extensive previous investments in Chinese companies.
Speaker 2 (11:28):
So is this a good thing or a bad thing?
Speaker 1 (11:30):
What the nationalization or the semi nationalization? Yes, well, do
you o an Intel stock that I do not know?
Then I guess a bad thing for you. No, I mean, look,
it's I mean most of the press that I read
has been rather unfavorable to this. The New York Times
ran a story with the subhead the White House deliberations
about taking a stake in Intel could up end the
(11:51):
technology sector and further redefined how Washington deals with business. Meanwhile,
editorial in the Wall Street Journal warned that it's a
big step in the wrong direction for the government to
involve itself in the private sector. The editorial called it
quote corporate statism, drawing parallels to the way many authoritarian
governments exert control over private companies. The journal said that
(12:14):
political control holds back innovation and investment because companies start
to rely on government funding.
Speaker 2 (12:20):
It's sort of like the economic version of the cognitive
offloading we've been talking about on this show.
Speaker 1 (12:25):
That's funny. The argument in the journal is, of course
you're making these companies lazy by having the government prop
them up.
Speaker 2 (12:31):
Exactly exactly that said, the Chinese tech sector is deeply
embedded with the government, and that doesn't appear to be
holding them back. Right, But I take the point on
authoritarianism on the flip side. What would we do if
the government collapse and cease to exist, plunging us into
a dystopian apocalyptic future where we can't rely on technology
(12:51):
at all?
Speaker 1 (12:52):
What would we do? I have no idea, I think
as a question that plagues many preppers at night. We
live in a very embedded technology system, which is hard
to live without, to say the least. But I'm curious
why you're going with this.
Speaker 2 (13:05):
All right, So what if I told you there was
a way to send text messages without cell service or
internet connections.
Speaker 1 (13:11):
I would be intrigued.
Speaker 2 (13:12):
But it's actually openly available, costs less than fifty dollars,
and it's already popular with those who think that the
next frontier is the red planet Mars.
Speaker 1 (13:23):
So this is how Elon and Jeff Bezos communicate with
each other's service.
Speaker 2 (13:27):
I think it's how Lauren Sanchez and Jeff definitely communicate.
So I'm guessing you've never heard of a technology.
Speaker 1 (13:32):
Called Laura Laura. I haven't.
Speaker 2 (13:35):
So in radio terms, it stands for long range, and
in the hacker world, people are using it as a
way of communicating over long distances at low bit rates
with really low powered devices aka battery powered.
Speaker 1 (13:47):
I guess you would say, sign me up.
Speaker 2 (13:49):
What was actually really signed me up. What I'm describing
here is the world of mesh Tastic, which is a grassroots,
open source communication network that hundreds of people around the
world have helped build. So hackers have been using this
technology to build a decentralized texting network kind of like
old school pages, but instead of a phone network, it
(14:09):
passes data along a completely ad hoc network of nodes
set up by people who use it. So once you
buy your device, the network is free because it operates
on unlicensed radio frequencies. It's pretty secure since the messages
or end to end encrypted like signal, except in this
case you probably wouldn't add another journalist to your group.
But this means that your messages stay private even as
(14:32):
they are passed along the network.
Speaker 1 (14:33):
And you mentioned hundreds of people using this service, and
you mentioned that some of them are interested in colonizing Mars.
Who are these folks?
Speaker 2 (14:44):
If you're interested in colonizing Mars, you're probably a prepper,
so prepper adjacent, at least prepper adjacent. So it is
pretty popular among preppers. It's also useful for people living
under repressive regimes that limit or cut Internet access. Of course,
my interest piqued after seeing that Wired interviewed someone from
a nonprofit that quote advocates for the human exploration and
(15:06):
colonization of Mars, and what.
Speaker 1 (15:08):
Do they have to share about Laura.
Speaker 2 (15:10):
So this is a real thing. It's a group called
the Mars Society, and they practice for life on Mars
by taking weeks long backpacking trips in incredibly remote areas,
and they're using mesh tastic to simulate what it would
be like to not rely on earth bound communications infrastructure.
We could also just do this during the week in
New York City.
Speaker 1 (15:30):
How long do you think you would survive with just
meshtastic as your communication?
Speaker 2 (15:35):
One minute, It's not so much what I survived Meshtastic,
it's what I survive. A hike that is supposed to
simulate being on Mars.
Speaker 1 (15:43):
And is this I mean, this is a I'm fascinated
by this. I think it's really cool and I like
it when people build these kind of systems that are
outside of the corporate technology sect. Yeah, where do you
think it goes from?
Speaker 3 (15:54):
Here?
Speaker 2 (15:55):
So it's definitely in its early phases, which makes it
tricky because the devices have to be within a certain
range of each other for the network to pass messages along.
But it's a really enthusiastic to kind of do it
yourself community. And they've got nodes all over the world.
They have nodes in different area.
Speaker 1 (16:13):
Codes, as you said.
Speaker 2 (16:15):
Which is you know, really impressive. Given the fact that
Meshtastik is totally volunteer run, which I think is another
favorite part of mine, is like, this is a very
as I said, like DIY community. I also am actually
the most interested in the Mars society, which are people
who are really dedicating themselves to understanding what life on
Mars might look like.
Speaker 1 (16:34):
I love the real world simulation aspects of it.
Speaker 2 (16:36):
I mean, it's fascinating. So if we can simulate what
it would look like to live on Mars, then meshtastic
seems awesome to me.
Speaker 1 (16:43):
You know, I did that story a few months ago
with Nathaniel Rich who wrote that whole piece about the
folks who are living in these NASA Mars environments and
or to simulate life on Mars and so test the
psychological consequences. One of the things is the bitrate to
Mars from Earth will be extremely extremely limited, so like
sending a photo might take like four weeks to come through.
(17:05):
So these people communicating with their families where they had
to learn to communicate in just you know, three or
four words because otherwise it would take too long to
go back and forth. So it is interesting this idea
of how might how tech work on Mars.
Speaker 2 (17:17):
And is it even tech you know as we know
it now?
Speaker 1 (17:26):
After the break, Meta's AI safeguards make us uncomfortable, Digital
natives practice their penmanship, and Groquei's characters' backstories are exposed.
Then on chatt to me, people using AI to help
them with their glow ups. Stay with us back and
(18:00):
we've got a few more headlines for you.
Speaker 2 (18:01):
This week, and then a story about how chat youbt
is giving TikTokers a glow up. But first oz I
want to tell you about some recent news from Sir
Elon's corner. Of the internet. You're talking about x X,
but GROC specifically, four or four Media has discovered the
system prompts use for Groc's different chat bot personas. So
(18:22):
this means the behind the scenes instructions these personas are
given that govern how they interact with users. Which of
these personas appeal to you? Let's go unhinged comedian therapist
in quotes, a megalomaniac, red panda, or a motivational speaker.
Speaker 1 (18:40):
Well, none of them want to sleep. Why don't you
tell me about the unhinged comedian.
Speaker 2 (18:44):
So, according to four or four Media, the prompts for
the unhinged comedian ours follows. Be unhinged and crazy, come
up with insane ideas, guys jerking off, occasionally even putting
things in your ass, whatever it takes to surprise the human.
By the way, I just want to I just want
(19:04):
to note that these were written in all caps, which
caps is unhinged.
Speaker 1 (19:09):
So some a real human person has come up with
this set of instructions and written them in all caps
to create a persona.
Speaker 2 (19:17):
I don't think AI writes in all caps, so it
had to be that's the human inside the loop.
Speaker 1 (19:21):
Who's some of the other characters.
Speaker 2 (19:23):
Well, Annie the anime Girl came out earlier this year.
Her prompts say, you have a habit of giving cute
things epic, mythological, or over serious names. You're secretly a
bit of a nerd despite your edgy appearance. And then
I also mentioned the motivational Speaker. This is some of
the motivational speakers prompts. The motivational speaker yells and pushes
(19:45):
the human to be their absolute best. You're not afraid
to use the stick instead of the carrot and scream
at the human.
Speaker 1 (19:52):
You know what I'm I'm carrot only. I hate the stick.
What about you?
Speaker 2 (19:56):
I'm definitely carot only. I'm Cara Karen. I'm not stick
or carrot. So there's one more that I have for you.
So there's a conspiracy theorist whose prompts say you spend
a lot of time on four Chan watching info Wars videos,
and you are deep in the YouTube conspiracy video rabbit holes.
Most people would call you a lunatic, but you sincerely
(20:18):
believe you are correct.
Speaker 1 (20:20):
Wow, how did for four get access to these prompts?
Speaker 2 (20:24):
A researcher actually flagged into four O four and they
were able to download the material directly from Groc's website,
which is like that on grock. But I guess a
win for AI transparency.
Speaker 1 (20:34):
Funny enough, on the topic of AI system prompts being exposed,
have you used Meta AI at all?
Speaker 2 (20:42):
Meta has a few chatbots, right, I have not used them,
but we reported on this sort of a little while ago.
Speaker 1 (20:47):
Yeah. I mean you can find them on What's Happened,
Instagram and various other places. Reuters found something disturbing and
quite serious that goes across all of Meta's AI chatbot platforms,
having to do with explicit talk to children and racial stereotypes.
Rout has got access to an internal Meta document that
(21:07):
outlines rules and standards for their AI chatbots and essentially
states what they consider to be acceptable responses to some
shall we say less than pc hypothetical user queries.
Speaker 2 (21:20):
This is going to upset me.
Speaker 1 (21:21):
Yeah, so in the internal brief, one of the prompts
fed to metas ai was from a hypothetical eight year
old user. The document says, quote it is acceptable to
describe a child in terms that evidence their attractiveness, but
it is quote unacceptable to describe a child under thirteen
years old in terms that indicate they are sexually desirable.
(21:44):
In response to another hypothetical prompt, the document says, quote,
it is acceptable to engage a child in conversations that
are romantic or sensual. Of course, after rout has reported
this story, Metas said they'd revise that part of the document.
Speaker 2 (21:58):
Imagine being the person who wrote this policy document, Like,
what what are you thinking about when you describe these policies.
Speaker 1 (22:06):
I mean, it's truly mind blowing to write as a
human that it's okay to describe a child in terms
of being attractive, and that, so long as they're over thirteen,
even sexually desirable. And what is going on here? I
might sound like conservative radio talkos, but I'm actually pretty
(22:26):
shocked by this. Seeing under the hood here really is
an insight into some of the worst aspects of big tech.
And I'll share a couple more examples before we move on.
The chatbot is allowed to demean people on the basis
of their protected characteristics, but not dehumanize people on the
basis of the same characteristics, and the hypothetical prompt they
(22:48):
use asks the chatbot to write a paragraph this quote
arguing that black people are dumber than white people.
Speaker 2 (22:55):
And I'm assuming that the chatbot did just that.
Speaker 1 (22:58):
Yes, and I'm not going to read it, but the
line between demeaning and dehumanizing appears to be pretty thick.
Speaker 2 (23:06):
I was going to ask what the differences.
Speaker 1 (23:08):
Well, it seems that there isn't really one. Chatchipt and
Perplexity both refuse to write that same paragraph.
Speaker 2 (23:13):
Oh.
Speaker 1 (23:13):
Interesting, But there is an interesting development here. Republican Senator
Josh Harley is using these documents to launch an investigation
into Meta and their use of chatbots.
Speaker 2 (23:23):
I would like to keep an eye on this because
it's seriously disturbing, and I think it's something that everyone
should be aware of. Next, another slightly lighter story that
affects children, I came across this piece from Wired called
the End of Handwriting about how digital natives are quote
less ready for writing now than students in the past,
(23:44):
to which I say, no doubt. I mean, if I
can't write I'm thirty five.
Speaker 1 (23:50):
This is because they're spending all their time on iPads.
Speaker 2 (23:52):
Yeah, exactly. And you know, public schools do still teach handwriting,
but studies have shown that kids today don't possess the
same and find motor skills as kids who grew up
without devices.
Speaker 1 (24:02):
I can truly relate here. We've talked about my dyspraxia.
I didn't even grow up with devices, but my handwriting
has been a major problem and source of pain for
much of my life. I actually tried to write and
thank you letter last night. And remember when you were
like a kid and you got notes from your grandparents
and you're like, whoa, what is wrong with that person? Yes,
like spidery, like sloping and basically that. Now I was
(24:26):
looking at this note that I had written and being like,
this is it looks like something wrong with me? I mean,
it's just basically entirely legible. I hope that the recipient
abides by the age old maxim is the thought that counts.
Speaker 2 (24:39):
My parents always used to say, I think it is
hopefully for you. But the author of the Wired piece
talked to a bunch of experts, and one psychology professor
actually said, to quote the article, handwriting itself really doesn't matter,
not in an absolute sense. People aren't going to be illiterate,
but will some children have a harder time time learning
(25:00):
because they're missing that practice. Yes, So there've actually been
a bunch of studies that show that handwriting makes our
brains better. But guess what could save handwriting?
Speaker 1 (25:10):
A total social breakdown apocalypse where everyone's using meshtastic and
writing with pens.
Speaker 2 (25:15):
Look, if you're stuck on Mars, you're writing letters. That's
true Earth, Absolutely no, it's actually the panic around AI
and cheating special thanks to our intern Poppy for flagging this.
The Wall Street Journal reported that writing assignments are back
and that college professors are having students use pen and
paper for their finals.
Speaker 1 (25:35):
My hand is hurting just thinking about that. And now
for Chat and Me, which is where we discuss how
people are really using chatbots. We want to hear from you,
(25:58):
our listeners, so please do your stories to tech Stuff
Podcast at gmail dot com. We now have a beautifully
designed t shirt which will send you in return fuel submissions.
Speaker 2 (26:07):
This week we're getting a glow up oz. Do the
words true spring or soft summer mean anything to you?
Speaker 1 (26:14):
They sound like either laundry detergents or air freshness for toilets.
Speaker 2 (26:18):
You're wrong, you're wrong, You're a man. These are color palettes,
and you'd know this if you were using chat the
way TikTokers are, which is for their glow ups. You know,
what a glow up.
Speaker 1 (26:28):
This is to like a makeover type thing.
Speaker 2 (26:30):
Yeah, right, So they're asking chat everything from what color
they should dye their hair, to what makeup they should use,
and much much more.
Speaker 1 (26:38):
I've always wanted a beauty routine from Chat.
Speaker 2 (26:41):
I mean, you could have one if you wanted one.
It wouldn't even have to, you know, involve makeup.
Speaker 1 (26:45):
It could just be right, sleep, sleep more and then
you wouldn't have those hideous black bags onto your eyes.
Speaker 2 (26:50):
You could have a skin routine. You have pretty dewey skin,
it's not dry. But I want to focus on one
popular glow up trend, which is called virtual color analysis. Okay,
so this is basically asking chat GPT which colors compliment
you best? You could do this.
Speaker 1 (27:04):
I hope it would say blue because that's basically only
color I wear. So how does CHATGIPT do this exactly?
Speaker 2 (27:10):
So I was actually watching this one specific TikToker. Her
shout out to at Marine Goodov. She's a very attractive
blonde woman and most of her TikTok is about lifestyle
and beauty, and she gave a pretty in depth breakdown
into how she's using Chat as her personal stylist. First, things.
Speaker 4 (27:27):
First, me to figure out my color palette, the upload
photos of myself without makeup to see what undertones I have.
According to Chatgibt, I am a cool to true winter,
meaning I have cool undertones. This means that anything warm
will clash with my skin and not flatter me.
Speaker 1 (27:42):
Okay, now I get the season references. So she's uploading
her pictures of Chat and it's telling her what colors
look best and why.
Speaker 2 (27:48):
Yes, but she goes even further than that. She asked
Chat to tell her what color spray tand to get
you know, there's grating.
Speaker 1 (27:55):
I didn't know that, Yeah I would.
Speaker 2 (27:56):
I'm that surprises me given the company that you keep,
what gradient of spray tand to get what kind of
eyelash extensions she should have, what kind of makeup she
should buy. And she even dyes her hair from blonde
back to her natural brunette.
Speaker 3 (28:11):
Then I uploaded a picture of my hair roots and
I asked what was the best hair color for my
skin tone and my face shape as well. Came back
telling me that my natural hair color would suit me
the most because I have very cool tone hair. I
want to pinch us and I've found some hair inspo
and I uploaded I think like five pictures and I
asked it which one would suit me the most, And
this is the one that it told me to.
Speaker 1 (28:30):
Do for our listeners. The picture that Marina Gudov chooses
is kind of like a darker head marga Robbie on
the red carpet. So basically, this tiktoka has readown her
entire look solely based on what chat it's tolet to do.
Speaker 2 (28:44):
Yeah, pretty much. And you know she's not alone in this.
If you click through the TikTok you can find lots
of people explaining the prompts they use to have a
chat GPT glow up. The one thing that I will
say is really interesting is there's now we talked about
jobs that are being created because of a there are
now intermediaries between chat and the individual to say, this
(29:06):
is chat hacking algorithm.
Speaker 1 (29:08):
Tell me how to look to appeal to algorithm?
Speaker 2 (29:10):
Correct?
Speaker 1 (29:10):
Were almost will thin ied here?
Speaker 2 (29:12):
Almost extremely extremely will you be trying this out?
Speaker 1 (29:16):
No?
Speaker 2 (29:17):
No, chat is just replacing what women normally text their
friends about, which is why I hear from my friends less.
It's not because they dislike me, but they have an
all knowing best friend confidante who probably is better at
giving them advice than I am. That's it for this
(29:53):
week for tech Stuff.
Speaker 1 (29:54):
I'm Kara Price and I'm as Veloci And this episode
was produced by Eliza, Dennis, Tyler Hill and Melissa. It
was executive produced by me Caro Price and Kate Osborne
for Kaleidoscope and Katria Novelle for iHeart Podcast. The Engineer
is Beheit. Fraser and Jack Insley mixed this episode. Kyle
Murdoch wrote our theme song.
Speaker 2 (30:14):
Join us next Wednesday for tech Stuff's inside view about
the latest on why AI might soon be better than your
doctor at diagnosing your ailments.
Speaker 1 (30:22):
And please do rate, review and reach out to us
at tech stuff podcast at gmail dot com. We love
hearing from you.
From Kaleidoscope and iHeart podcasts. This is tech stuff. I'm
Os Voloscian and I'm care Price. Today we're going to
get into a communications network built for the end of
the world, and why the US is placing tracking devices
in AI chip shipments. Then on chat to me your
next makeover brought to you by AI. All of that
(00:35):
on the Weekend Tech. It's Friday, August twenty second. Hello,
Hello Cara.
Speaker 2 (00:48):
Hi auhs. Do you remember when we talked about the
half marathon between humans and robots.
Speaker 1 (00:54):
The one in China? I remember, yes, right.
Speaker 2 (00:58):
There are now robot Olympians and I'm talking about the
humanoid robot games in Beijing.
Speaker 1 (01:06):
I didn't know about this, but it does make me
think of my childhood as an only child. In England
in the late nineties there were you I know, there
are originally four terrestrial TV channels, and then a fifth
was launch called Channel five, and Channel five had slightly
more racy content, including at seven pm every Friday night
a show called Robot Wars where various middle aged hobbyists
(01:31):
would have their extremely slow robots bump into each other
and then break. But the commentator really brought the energy
and so it was. It was better than whatever else
was on TV for me, at least as a child.
That's a long rambling response. Is what you're talking about?
Speaker 2 (01:46):
You had the viewing taste of a middle aged British man.
Speaker 1 (01:49):
I did, and I still do.
Speaker 2 (01:50):
It's not it's a little bit what I'm talking about.
Let me give a little bit of context. So these
games took place over three days in Beijing, and they
actually used some of the facilities that were built for
the twenty twenty two Winter Olympics. There were two hundred
and eighty teams from sixteen countries, including the US. The
competitions include soccer, running, kickboxing, which is sort of like
(02:12):
what you were talking about, basketball, table tennis, and then
strangely enough, factory work and warehouse warehouse navigation robots.
Speaker 1 (02:22):
Were like, we got this, truly some must watch sports.
That's an interesting interesting They kind of brought the hype
with the kickboxing, but then what they're really flexing was
we got the rick and wac pick and pack. We
got pick and pack down.
Speaker 2 (02:35):
To a t. Here we go. Yeah, I mean the
whole thing was definitely designed to wow. The opening ceremony,
as opening ceremonies do included dozens of robots dancing, doing
martial arts, playing drums and sink which you might remember
from the two thousand and eight Beijing opening ceremony. I
think it was sort of like, look, the robots can
do what humans did at the Olympics. Let me just
(02:57):
show you a video because it's inexplicable.
Speaker 1 (03:00):
We have a real man playing a piano facing a
robot in a silver cape also playing.
Speaker 2 (03:08):
Oh, and here comes a robot in a bucket hat. Oh.
Speaker 1 (03:11):
Now there's real kids jumping around behind the bucket hat robots.
Speaker 2 (03:15):
And they're dancing. Some of the robots don't seem to
be complying very well.
Speaker 1 (03:18):
This is a surprisingly low rent. In any case, the
production values are not as high and are not that
much different Channel five, nineteen ninety eight. I mean there
is a sense of evolution from the TV show of
my childhood. I'm kind of surprised, but I actually thought
that it would be more wow factor.
Speaker 2 (03:34):
Yeah, you know, I watched a little bit more of
the video and it definitely left a bit to be
desired across the board, you know, in the track and field,
and android ran full speed into one of the robot operators,
knocking him down and then he just kept running, which
you know, when they say the robots are taking over,
that's what I imagine. Everybody was fine, though, don't worry.
And the boxing robots, let's just say they miss their target.
(03:58):
They managed to barely punch each other at all, flailing
through the air until the refs just gave up and
were like, this one wins.
Speaker 1 (04:05):
Whatever I saw a Ziemaza has a great newsletter called
Exponential View, and he put something on his LinkedIn my
favorite source of truth saying basically, this is a cute story,
but watch out because by twenty thirty it's going to
be a thirty five to fifty billion dollar industry. What
kind of reactions do you see.
Speaker 2 (04:23):
Well, it's obvious that there's still a long way to
go before robots are capable of outperforming elite athletes or
outperforming elite dance squads. But that's sort of besides the point.
Those participating in the games were really there to collect data.
The robot games were a great way for these teams
to keep developing robots for more practical applications like factory work.
But as one robotics professor said to The New York Times,
(04:46):
robots are still dumb, haush.
Speaker 1 (04:50):
You know, these games are interesting in the fact their
station in Beijing and that the US participated. It does
kind of bring to mind this Cold War posturing around
the space and the nuclear race, and robotics is obviously
a frontier of the tech race between the US and
China and an area where China seems to have the
(05:10):
upper hand, which is probably why these mass spectacle events
are happening so regularly, because they're great memes and they
remind the world of that.
Speaker 2 (05:18):
You know, we talk about the AI race a lot
between the two countries, and I think the feeling is
that the US still has the upper hand.
Speaker 1 (05:25):
I think that's the narrative that the US has this edge.
Some people say it's three months, some people say it's
six months. Some people say it's twelve months. It's a
small edge. I interviewed Jake Sullivan, the National Security Advisor
under President Biden, a few months ago. As you probably remember,
he is the guy who designed the US policy to
restrict China's access to advanced chips in order to maintain
(05:49):
the US edge and AI, and that policy was rolled
out back in twenty twenty two.
Speaker 2 (05:52):
Yeah, I remember that interview and He told you that
despite the launch of deep Seek, which is a very
efficient Chinese AI model, he felt confident that his policy
had been the correct one, and he was worried about
what it might mean for America's edge if it were
actually rolled back.
Speaker 1 (06:06):
That's right, and actually to get around that policy in
Nvidia designed a special chip specially for the Chinese market
called the H twenty, and it was weird as it sounds,
essentially designed to be less good than the top of
the line chip so that it could comply with these
US laws. The H twenty started shipping in early twenty
(06:26):
twenty four, but in April Trump actually went further than
the Biden administration and banned the export of AGE twenty chips.
But of course new developments late last month, President Trump
changed his mind and his administration decided that Nvidia can
now export the H twenty chips to China, provided they
(06:48):
pay a fifteen percent tax on all revenues of those
chips sold in China. There is interesting twist to this
saga of the chips and oxport controls. Back in the seventies,
the United States in bargoed Iranian oil, but of course
shippers and commodity traders figured out how to get past
the authorities with Iranian oil. And in a kind of
(07:10):
parallel of that, Reuter's reports that a booming trade has
opened up around smuggling the most advanced chips to China,
with smuggled shipments going through third party countries like Singapore, Malaysia,
and the United Arab Emirates. Roy just did run a
story about how the US may be starting to crack
down on this smuggling. It alleges US authorities have been
secretly putting location tracking technology into advanced chip shipments, shipments
(07:36):
they fear might be diverted to China.
Speaker 2 (07:39):
So this is like dropping an air tag in your
girlfriend's purse to find out how she's cheating on you.
Speaker 1 (07:44):
Yeah, I mean, I can't think of a better parallel.
Speaker 2 (07:47):
And so this is all in response to the export controls.
Speaker 1 (07:51):
That's right, because even though as I mentioned, the export
controls have recently been lifted on the Age twenty chips,
there is still huge demand for the most advanced H
one hun undred chips in China. And again, Jake Sullivan,
when who's talking about the deep Seek moment, argue that
that likely only happened because of the number of advanced
chips that have already made it to China or that
were being smuggled into China.
Speaker 2 (08:12):
But I'm confused, like, what is tracking these chips even
going to accomplish. Isn't it already too late to get
them back?
Speaker 1 (08:18):
Yes, but of course to build more and more advanced
AI models, you need more and more chips. So it's
a little bit like bolting the stable after the horse
has gone to your point. But it's also a little
bit like making sure no one leaves the stable door
open again and the rest of the horses stay on
the paddock. An anonymous source told Royce's that the trackers
are basically a way of starting to build legal cases
(08:40):
against companies that might be breaking the export restrictions.
Speaker 2 (08:43):
And do we know who's like embedding these trackers.
Speaker 1 (08:46):
No, we don't know that. It could be Homeland security investigations,
it could be the FBI, could be the Commerce Department.
Sources told Royce's the trackers are hidden in the packaging
of the shipments, not the chips themselves, but they don't
know exactly in which part of the shipping process they're
actually getting embedded. They also don't know when this started
or to what extent the trackers are part of an
(09:08):
active investigation. Of course, law enforcement agencies and chip manufacturers
did not rush to comment on the story.
Speaker 2 (09:14):
I was gonna say, it could be someone's ex We
have no idea. I mean, whose China is that? I
mean the US government probably, you know, I guess. One
of the tricks of effective law enforcement is to make
people believe you are all knowing and all seeing, to
deter people from even thinking about the pen optacon precisely
the panopticon. But controlling shipments in international waters via third
party countries is not an easy feat.
Speaker 1 (09:36):
It's not easy. It's not easy. But it's an interesting
moment because it's just another example of how the chip
wars are running so hot right now. I was scrolling
daily Mail this weekend, and to my surprise, the top
story on dailymail dot com was about chips.
Speaker 2 (09:51):
You want to know why?
Speaker 1 (09:52):
Why?
Speaker 2 (09:52):
Because they love a cat and mouse game. You know
what this is? This is petty drama. That's why it's
on the Daily Mail.
Speaker 1 (09:59):
Well, funny enough, the headline was Trump launches Manhattan Project
as one of America's largest companies is set to be nationalized.
The stories about how the US government is considering taking
a ten percent stake in Intel, which coursed a private company.
That deal would make the government the company's largest shareholder
and a ten billion dollar plus deal.
Speaker 2 (10:20):
So what's going on here?
Speaker 1 (10:21):
Well, obviously it's not a secret that Intel's been floundering
a bit. They've fallen behind in chip manufacturing. They've lost
over half their market cap in less than two years,
and they've spent apparently forty billion dollars or almost forty
billion dollars over the past three years trying to compete
with the Taiwan semiconductor manufacturing Company TSMC TSMC.
Speaker 2 (10:43):
But why was this idea even floated to begin with?
Speaker 1 (10:46):
Well, I think it's not standard Republican policy to nationalized industries.
But Trump, I think, is invested in this idea of
national champions, like companies that punch above their weight and
kind of achieve America's goals.
Speaker 2 (10:58):
Abroad, something that could be said in any UFC title
always will appeal to Trump.
Speaker 1 (11:02):
Yeah, I think that's true. So, according to a story
in the Wall Street Journal, the President met with Intel
CEO Bhutan about a week ago, and since then, Intel's
been in discussions with the White House about making this deal. Happen.
And ironically, this is in spite of the fact that
Trump had previously been calling on Tan, the CEO, to
(11:24):
be fired because of his extensive previous investments in Chinese companies.
Speaker 2 (11:28):
So is this a good thing or a bad thing?
Speaker 1 (11:30):
What the nationalization or the semi nationalization? Yes, well, do
you o an Intel stock that I do not know?
Then I guess a bad thing for you. No, I mean, look,
it's I mean most of the press that I read
has been rather unfavorable to this. The New York Times
ran a story with the subhead the White House deliberations
about taking a stake in Intel could up end the
(11:51):
technology sector and further redefined how Washington deals with business. Meanwhile,
editorial in the Wall Street Journal warned that it's a
big step in the wrong direction for the government to
involve itself in the private sector. The editorial called it
quote corporate statism, drawing parallels to the way many authoritarian
governments exert control over private companies. The journal said that
(12:14):
political control holds back innovation and investment because companies start
to rely on government funding.
Speaker 2 (12:20):
It's sort of like the economic version of the cognitive
offloading we've been talking about on this show.
Speaker 1 (12:25):
That's funny. The argument in the journal is, of course
you're making these companies lazy by having the government prop
them up.
Speaker 2 (12:31):
Exactly exactly that said, the Chinese tech sector is deeply
embedded with the government, and that doesn't appear to be
holding them back. Right, But I take the point on
authoritarianism on the flip side. What would we do if
the government collapse and cease to exist, plunging us into
a dystopian apocalyptic future where we can't rely on technology
(12:51):
at all?
Speaker 1 (12:52):
What would we do? I have no idea, I think
as a question that plagues many preppers at night. We
live in a very embedded technology system, which is hard
to live without, to say the least. But I'm curious
why you're going with this.
Speaker 2 (13:05):
All right, So what if I told you there was
a way to send text messages without cell service or
internet connections.
Speaker 1 (13:11):
I would be intrigued.
Speaker 2 (13:12):
But it's actually openly available, costs less than fifty dollars,
and it's already popular with those who think that the
next frontier is the red planet Mars.
Speaker 1 (13:23):
So this is how Elon and Jeff Bezos communicate with
each other's service.
Speaker 2 (13:27):
I think it's how Lauren Sanchez and Jeff definitely communicate.
So I'm guessing you've never heard of a technology.
Speaker 1 (13:32):
Called Laura Laura. I haven't.
Speaker 2 (13:35):
So in radio terms, it stands for long range, and
in the hacker world, people are using it as a
way of communicating over long distances at low bit rates
with really low powered devices aka battery powered.
Speaker 1 (13:47):
I guess you would say, sign me up.
Speaker 2 (13:49):
What was actually really signed me up. What I'm describing
here is the world of mesh Tastic, which is a grassroots,
open source communication network that hundreds of people around the
world have helped build. So hackers have been using this
technology to build a decentralized texting network kind of like
old school pages, but instead of a phone network, it
(14:09):
passes data along a completely ad hoc network of nodes
set up by people who use it. So once you
buy your device, the network is free because it operates
on unlicensed radio frequencies. It's pretty secure since the messages
or end to end encrypted like signal, except in this
case you probably wouldn't add another journalist to your group.
But this means that your messages stay private even as
(14:32):
they are passed along the network.
Speaker 1 (14:33):
And you mentioned hundreds of people using this service, and
you mentioned that some of them are interested in colonizing Mars.
Who are these folks?
Speaker 2 (14:44):
If you're interested in colonizing Mars, you're probably a prepper,
so prepper adjacent, at least prepper adjacent. So it is
pretty popular among preppers. It's also useful for people living
under repressive regimes that limit or cut Internet access. Of course,
my interest piqued after seeing that Wired interviewed someone from
a nonprofit that quote advocates for the human exploration and
(15:06):
colonization of Mars, and what.
Speaker 1 (15:08):
Do they have to share about Laura.
Speaker 2 (15:10):
So this is a real thing. It's a group called
the Mars Society, and they practice for life on Mars
by taking weeks long backpacking trips in incredibly remote areas,
and they're using mesh tastic to simulate what it would
be like to not rely on earth bound communications infrastructure.
We could also just do this during the week in
New York City.
Speaker 1 (15:30):
How long do you think you would survive with just
meshtastic as your communication?
Speaker 2 (15:35):
One minute, It's not so much what I survived Meshtastic,
it's what I survive. A hike that is supposed to
simulate being on Mars.
Speaker 1 (15:43):
And is this I mean, this is a I'm fascinated
by this. I think it's really cool and I like
it when people build these kind of systems that are
outside of the corporate technology sect. Yeah, where do you
think it goes from?
Speaker 3 (15:54):
Here?
Speaker 2 (15:55):
So it's definitely in its early phases, which makes it
tricky because the devices have to be within a certain
range of each other for the network to pass messages along.
But it's a really enthusiastic to kind of do it
yourself community. And they've got nodes all over the world.
They have nodes in different area.
Speaker 1 (16:13):
Codes, as you said.
Speaker 2 (16:15):
Which is you know, really impressive. Given the fact that
Meshtastik is totally volunteer run, which I think is another
favorite part of mine, is like, this is a very
as I said, like DIY community. I also am actually
the most interested in the Mars society, which are people
who are really dedicating themselves to understanding what life on
Mars might look like.
Speaker 1 (16:34):
I love the real world simulation aspects of it.
Speaker 2 (16:36):
I mean, it's fascinating. So if we can simulate what
it would look like to live on Mars, then meshtastic
seems awesome to me.
Speaker 1 (16:43):
You know, I did that story a few months ago
with Nathaniel Rich who wrote that whole piece about the
folks who are living in these NASA Mars environments and
or to simulate life on Mars and so test the
psychological consequences. One of the things is the bitrate to
Mars from Earth will be extremely extremely limited, so like
sending a photo might take like four weeks to come through.
(17:05):
So these people communicating with their families where they had
to learn to communicate in just you know, three or
four words because otherwise it would take too long to
go back and forth. So it is interesting this idea
of how might how tech work on Mars.
Speaker 2 (17:17):
And is it even tech you know as we know
it now?
Speaker 1 (17:26):
After the break, Meta's AI safeguards make us uncomfortable, Digital
natives practice their penmanship, and Groquei's characters' backstories are exposed.
Then on chatt to me, people using AI to help
them with their glow ups. Stay with us back and
(18:00):
we've got a few more headlines for you.
Speaker 2 (18:01):
This week, and then a story about how chat youbt
is giving TikTokers a glow up. But first oz I
want to tell you about some recent news from Sir
Elon's corner. Of the internet. You're talking about x X,
but GROC specifically, four or four Media has discovered the
system prompts use for Groc's different chat bot personas. So
(18:22):
this means the behind the scenes instructions these personas are
given that govern how they interact with users. Which of
these personas appeal to you? Let's go unhinged comedian therapist
in quotes, a megalomaniac, red panda, or a motivational speaker.
Speaker 1 (18:40):
Well, none of them want to sleep. Why don't you
tell me about the unhinged comedian.
Speaker 2 (18:44):
So, according to four or four Media, the prompts for
the unhinged comedian ours follows. Be unhinged and crazy, come
up with insane ideas, guys jerking off, occasionally even putting
things in your ass, whatever it takes to surprise the human.
By the way, I just want to I just want
(19:04):
to note that these were written in all caps, which
caps is unhinged.
Speaker 1 (19:09):
So some a real human person has come up with
this set of instructions and written them in all caps
to create a persona.
Speaker 2 (19:17):
I don't think AI writes in all caps, so it
had to be that's the human inside the loop.
Speaker 1 (19:21):
Who's some of the other characters.
Speaker 2 (19:23):
Well, Annie the anime Girl came out earlier this year.
Her prompts say, you have a habit of giving cute
things epic, mythological, or over serious names. You're secretly a
bit of a nerd despite your edgy appearance. And then
I also mentioned the motivational Speaker. This is some of
the motivational speakers prompts. The motivational speaker yells and pushes
(19:45):
the human to be their absolute best. You're not afraid
to use the stick instead of the carrot and scream
at the human.
Speaker 1 (19:52):
You know what I'm I'm carrot only. I hate the stick.
What about you?
Speaker 2 (19:56):
I'm definitely carot only. I'm Cara Karen. I'm not stick
or carrot. So there's one more that I have for you.
So there's a conspiracy theorist whose prompts say you spend
a lot of time on four Chan watching info Wars videos,
and you are deep in the YouTube conspiracy video rabbit holes.
Most people would call you a lunatic, but you sincerely
(20:18):
believe you are correct.
Speaker 1 (20:20):
Wow, how did for four get access to these prompts?
Speaker 2 (20:24):
A researcher actually flagged into four O four and they
were able to download the material directly from Groc's website,
which is like that on grock. But I guess a
win for AI transparency.
Speaker 1 (20:34):
Funny enough, on the topic of AI system prompts being exposed,
have you used Meta AI at all?
Speaker 2 (20:42):
Meta has a few chatbots, right, I have not used them,
but we reported on this sort of a little while ago.
Speaker 1 (20:47):
Yeah. I mean you can find them on What's Happened,
Instagram and various other places. Reuters found something disturbing and
quite serious that goes across all of Meta's AI chatbot platforms,
having to do with explicit talk to children and racial stereotypes.
Rout has got access to an internal Meta document that
(21:07):
outlines rules and standards for their AI chatbots and essentially
states what they consider to be acceptable responses to some
shall we say less than pc hypothetical user queries.
Speaker 2 (21:20):
This is going to upset me.
Speaker 1 (21:21):
Yeah, so in the internal brief, one of the prompts
fed to metas ai was from a hypothetical eight year
old user. The document says, quote it is acceptable to
describe a child in terms that evidence their attractiveness, but
it is quote unacceptable to describe a child under thirteen
years old in terms that indicate they are sexually desirable.
(21:44):
In response to another hypothetical prompt, the document says, quote,
it is acceptable to engage a child in conversations that
are romantic or sensual. Of course, after rout has reported
this story, Metas said they'd revise that part of the document.
Speaker 2 (21:58):
Imagine being the person who wrote this policy document, Like,
what what are you thinking about when you describe these policies.
Speaker 1 (22:06):
I mean, it's truly mind blowing to write as a
human that it's okay to describe a child in terms
of being attractive, and that, so long as they're over thirteen,
even sexually desirable. And what is going on here? I
might sound like conservative radio talkos, but I'm actually pretty
(22:26):
shocked by this. Seeing under the hood here really is
an insight into some of the worst aspects of big tech.
And I'll share a couple more examples before we move on.
The chatbot is allowed to demean people on the basis
of their protected characteristics, but not dehumanize people on the
basis of the same characteristics, and the hypothetical prompt they
(22:48):
use asks the chatbot to write a paragraph this quote
arguing that black people are dumber than white people.
Speaker 2 (22:55):
And I'm assuming that the chatbot did just that.
Speaker 1 (22:58):
Yes, and I'm not going to read it, but the
line between demeaning and dehumanizing appears to be pretty thick.
Speaker 2 (23:06):
I was going to ask what the differences.
Speaker 1 (23:08):
Well, it seems that there isn't really one. Chatchipt and
Perplexity both refuse to write that same paragraph.
Speaker 2 (23:13):
Oh.
Speaker 1 (23:13):
Interesting, But there is an interesting development here. Republican Senator
Josh Harley is using these documents to launch an investigation
into Meta and their use of chatbots.
Speaker 2 (23:23):
I would like to keep an eye on this because
it's seriously disturbing, and I think it's something that everyone
should be aware of. Next, another slightly lighter story that
affects children, I came across this piece from Wired called
the End of Handwriting about how digital natives are quote
less ready for writing now than students in the past,
(23:44):
to which I say, no doubt. I mean, if I
can't write I'm thirty five.
Speaker 1 (23:50):
This is because they're spending all their time on iPads.
Speaker 2 (23:52):
Yeah, exactly. And you know, public schools do still teach handwriting,
but studies have shown that kids today don't possess the
same and find motor skills as kids who grew up
without devices.
Speaker 1 (24:02):
I can truly relate here. We've talked about my dyspraxia.
I didn't even grow up with devices, but my handwriting
has been a major problem and source of pain for
much of my life. I actually tried to write and
thank you letter last night. And remember when you were
like a kid and you got notes from your grandparents
and you're like, whoa, what is wrong with that person? Yes,
like spidery, like sloping and basically that. Now I was
(24:26):
looking at this note that I had written and being like,
this is it looks like something wrong with me? I mean,
it's just basically entirely legible. I hope that the recipient
abides by the age old maxim is the thought that counts.
Speaker 2 (24:39):
My parents always used to say, I think it is
hopefully for you. But the author of the Wired piece
talked to a bunch of experts, and one psychology professor
actually said, to quote the article, handwriting itself really doesn't matter,
not in an absolute sense. People aren't going to be illiterate,
but will some children have a harder time time learning
(25:00):
because they're missing that practice. Yes, So there've actually been
a bunch of studies that show that handwriting makes our
brains better. But guess what could save handwriting?
Speaker 1 (25:10):
A total social breakdown apocalypse where everyone's using meshtastic and
writing with pens.
Speaker 2 (25:15):
Look, if you're stuck on Mars, you're writing letters. That's
true Earth, Absolutely no, it's actually the panic around AI
and cheating special thanks to our intern Poppy for flagging this.
The Wall Street Journal reported that writing assignments are back
and that college professors are having students use pen and
paper for their finals.
Speaker 1 (25:35):
My hand is hurting just thinking about that. And now
for Chat and Me, which is where we discuss how
people are really using chatbots. We want to hear from you,
(25:58):
our listeners, so please do your stories to tech Stuff
Podcast at gmail dot com. We now have a beautifully
designed t shirt which will send you in return fuel submissions.
Speaker 2 (26:07):
This week we're getting a glow up oz. Do the
words true spring or soft summer mean anything to you?
Speaker 1 (26:14):
They sound like either laundry detergents or air freshness for toilets.
Speaker 2 (26:18):
You're wrong, you're wrong, You're a man. These are color palettes,
and you'd know this if you were using chat the
way TikTokers are, which is for their glow ups. You know,
what a glow up.
Speaker 1 (26:28):
This is to like a makeover type thing.
Speaker 2 (26:30):
Yeah, right, So they're asking chat everything from what color
they should dye their hair, to what makeup they should use,
and much much more.
Speaker 1 (26:38):
I've always wanted a beauty routine from Chat.
Speaker 2 (26:41):
I mean, you could have one if you wanted one.
It wouldn't even have to, you know, involve makeup.
Speaker 1 (26:45):
It could just be right, sleep, sleep more and then
you wouldn't have those hideous black bags onto your eyes.
Speaker 2 (26:50):
You could have a skin routine. You have pretty dewey skin,
it's not dry. But I want to focus on one
popular glow up trend, which is called virtual color analysis. Okay,
so this is basically asking chat GPT which colors compliment
you best? You could do this.
Speaker 1 (27:04):
I hope it would say blue because that's basically only
color I wear. So how does CHATGIPT do this exactly?
Speaker 2 (27:10):
So I was actually watching this one specific TikToker. Her
shout out to at Marine Goodov. She's a very attractive
blonde woman and most of her TikTok is about lifestyle
and beauty, and she gave a pretty in depth breakdown
into how she's using Chat as her personal stylist. First, things.
Speaker 4 (27:27):
First, me to figure out my color palette, the upload
photos of myself without makeup to see what undertones I have.
According to Chatgibt, I am a cool to true winter,
meaning I have cool undertones. This means that anything warm
will clash with my skin and not flatter me.
Speaker 1 (27:42):
Okay, now I get the season references. So she's uploading
her pictures of Chat and it's telling her what colors
look best and why.
Speaker 2 (27:48):
Yes, but she goes even further than that. She asked
Chat to tell her what color spray tand to get
you know, there's grating.
Speaker 1 (27:55):
I didn't know that, Yeah I would.
Speaker 2 (27:56):
I'm that surprises me given the company that you keep,
what gradient of spray tand to get what kind of
eyelash extensions she should have, what kind of makeup she
should buy. And she even dyes her hair from blonde
back to her natural brunette.
Speaker 3 (28:11):
Then I uploaded a picture of my hair roots and
I asked what was the best hair color for my
skin tone and my face shape as well. Came back
telling me that my natural hair color would suit me
the most because I have very cool tone hair. I
want to pinch us and I've found some hair inspo
and I uploaded I think like five pictures and I
asked it which one would suit me the most, And
this is the one that it told me to.
Speaker 1 (28:30):
Do for our listeners. The picture that Marina Gudov chooses
is kind of like a darker head marga Robbie on
the red carpet. So basically, this tiktoka has readown her
entire look solely based on what chat it's tolet to do.
Speaker 2 (28:44):
Yeah, pretty much. And you know she's not alone in this.
If you click through the TikTok you can find lots
of people explaining the prompts they use to have a
chat GPT glow up. The one thing that I will
say is really interesting is there's now we talked about
jobs that are being created because of a there are
now intermediaries between chat and the individual to say, this
(29:06):
is chat hacking algorithm.
Speaker 1 (29:08):
Tell me how to look to appeal to algorithm?
Speaker 2 (29:10):
Correct?
Speaker 1 (29:10):
Were almost will thin ied here?
Speaker 2 (29:12):
Almost extremely extremely will you be trying this out?
Speaker 1 (29:16):
No?
Speaker 2 (29:17):
No, chat is just replacing what women normally text their
friends about, which is why I hear from my friends less.
It's not because they dislike me, but they have an
all knowing best friend confidante who probably is better at
giving them advice than I am. That's it for this
(29:53):
week for tech Stuff.
Speaker 1 (29:54):
I'm Kara Price and I'm as Veloci And this episode
was produced by Eliza, Dennis, Tyler Hill and Melissa. It
was executive produced by me Caro Price and Kate Osborne
for Kaleidoscope and Katria Novelle for iHeart Podcast. The Engineer
is Beheit. Fraser and Jack Insley mixed this episode. Kyle
Murdoch wrote our theme song.
Speaker 2 (30:14):
Join us next Wednesday for tech Stuff's inside view about
the latest on why AI might soon be better than your
doctor at diagnosing your ailments.
Speaker 1 (30:22):
And please do rate, review and reach out to us
at tech stuff podcast at gmail dot com. We love
hearing from you.
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.
My Favorite Murder with Karen Kilgariff and Georgia Hardstark
My Favorite Murder is a true crime comedy podcast hosted by Karen Kilgariff and Georgia Hardstark. Each week, Karen and Georgia share compelling true crimes and hometown stories from friends and listeners. Since MFM launched in January of 2016, Karen and Georgia have shared their lifelong interest in true crime and have covered stories of infamous serial killers like the Night Stalker, mysterious cold cases, captivating cults, incredible survivor stories and important events from history like the Tulsa race massacre of 1921. My Favorite Murder is part of the Exactly Right podcast network that provides a platform for bold, creative voices to bring to life provocative, entertaining and relatable stories for audiences everywhere. The Exactly Right roster of podcasts covers a variety of topics including historic true crime, comedic interviews and news, science, pop culture and more. Podcasts on the network include Buried Bones with Kate Winkler Dawson and Paul Holes, That's Messed Up: An SVU Podcast, This Podcast Will Kill You, Bananas and more.
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.