August 1, 2025 • 36 mins
How good is Dr. GPT? This week, Oz and Karah spill the tea on Elon Musk’s new retro-futurist diner in LA—where robots serve popcorn and the food comes in Cybertruck-shaped boxes. Then, Oz gets serious about Trump’s new AI Action Plan and what it could mean for American infrastructure and free speech. Karah addresses the rise of TikTok PI’s and public shaming. And finally, on Chat and Me, AI makes a diagnosis that is a little too real.
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.
Speaker 2 (00:16):
I'm as Voloscian and I'm Cara Price.
Speaker 1 (00:18):
Today we'll get into TikTok, private Investigators and America's AI future.
Then on chatting me, AI makes a diagnosis that's a
little too real. All of that on the Weekend Tech.
It's Friday, July thirtieth. Hello Cara, Hey, ohs.
Speaker 2 (00:39):
So you might see that I'm bopping my foot this morning.
Speaker 1 (00:42):
Yes, I'm bopping. I do have RLS, but this is exaggerated.
Speaker 2 (00:48):
It's much exaggerated because the Tesla Diner open this week.
Do you know about this?
Speaker 1 (00:54):
Of course I do. I mean it's I've been obsessing
about it. It looks like a drive in movie theater
from the fifties, retro futurist Jetson's esthetic, and there's some
pretty weird stuff going down there.
Speaker 2 (01:05):
That is the one. It opened last Monday at four
to twenty pm.
Speaker 1 (01:12):
Elon's favorite time of day and favorite little joke. He
recently rolled out this robotaxi service, and the fair was
initially said at four dollars and twenty cents as well.
So I guess the old ones are the good ones.
Speaker 2 (01:23):
For those who don't know. Mom for twenty is a
weed thing that's when people smoke weed. April twentieth Market,
Canada's more mark your Calendars. So I actually said to
two of my friends, please go, and then they were like,
it looks so disgusting, I'm not gonna go.
Speaker 1 (01:38):
So basically, you live in New York. We spent a
lot of time in LA and so we talked about
the idea of sending some of your LA pals to
go and check it out.
Speaker 2 (01:46):
So they were like, yeah, we'll go, and they're like,
we're not going. So I just want to describe it.
It has a curved chrome siding, curved white booths, long countertops.
A Tesla robot Optimus serves popcorn on the second floor.
Speaker 1 (02:00):
There, Kim Kardeshian lets out of a clutches to go
to the diner.
Speaker 2 (02:03):
He was like, I'm over calabas As I'm out. There
are eighty supercharger stalls for Tesla's two forty five foot
movie screens.
Speaker 1 (02:11):
I gather screening Star Trek amongst others.
Speaker 2 (02:14):
Of course, it has to be Star Trek. There's food
served in cybertruck shaped boxes and with cyber truck shaped
wooden imagine the company who had to make the cyber
truck shape would.
Speaker 1 (02:25):
Like a happy meal box, but in the shape of
a cyber truck.
Speaker 2 (02:27):
That's right, it's everything, not golden arches, it's everything. Tesla Now,
one of my favorite sources, TMZ, actually had a pretty
harrowing story. A woman was violently struck by furniture falling
off the second floor patio, and it actually missed her
baby's head by inches, which to me is terrifying obviously
and a bit of a harbinger for some of the
(02:49):
very real company issues facing Tesla.
Speaker 1 (02:52):
There was actually reddit through it called Elon Musk's Tesla
Diner is the cyber truck of restaurants, which you can
imagine was not intend as a compliment. There's the obvious
aesthetic parallels this kind of retro futurist, although of course
the cyber truck famously has no curves, only angles. But nonetheless,
like me, this Reddit thread mentioned this restaurant's design for
(03:14):
two hundred and fifty people, were there only three bathrooms?
Not good, and one user cracked that the Optimus robot
who was serving popcorn might have been better employed downstairs
mopping down the bathrooms, which apparently were not a sight
for sore eyes.
Speaker 2 (03:30):
Well, and by the looks of the food, the restaurant
there probably needs a few more bathrooms.
Speaker 1 (03:35):
Carol, that's a bit spiky all angles. You know you
mentioned the food. There was a great Guardian piece that
I think described the testa diner well, but also I
think inadvertently touched on the heart of Tesla's dilemma. Here's
what the Guardian wrote, quote the diner offers a mix
of own the Libs and we are the Libs options.
(03:57):
On the one hand, epic bacon four strips of bacon
as served with sources as a meat fluenzer alternative to
French fries. On the other, avocado toast and machlattis.
Speaker 2 (04:10):
I know which is which?
Speaker 1 (04:11):
I think which would you order your vegetarians?
Speaker 2 (04:13):
One of the vegetarians. I wouldn't have a choice, but
I also I'd.
Speaker 1 (04:15):
Be pretty tempted by those I own the libs menu. However,
as the Bible says, no man can serve two masters,
and this dilemma of own the Libs versus we are
the Libs is one which is playing out more broadly
at Tesla and bedeviling the company. So the opening of
(04:38):
the diner last week was in the same week that
Tesla released its financials. The company took a bit of
a beating, with net income down sixteen percent in the
second quarter, and The Wall Street Journal said that the
company's finances are quote in free fall, and they pointed
out two distinct drivers of this. On the one hand,
Musk's adventures with those have not done Tesla many favors
(05:02):
with its original client base, i e. People in California,
Europe who are EV fans and have environmental motivations, and
so with that audience, Tesla is big time out of
favor post the dalliance with Trump. On the other hand,
the dallions with Trump did not protect Musk from the
red meat attacks on EV's and part of the Big
(05:26):
Beautiful Bill included cuts to EV subsidies and tax credits,
which have really hurt Tesla's bottom line. And as you remember,
there were kind of rumors of the breaking up of
the bromance percolating for some time, but this was really
the wedge issue between Trump and Musk was the subsidies,
and indeed it's hurting Tesla's bottom line. And so this
brings me to my next story. There was a very
(05:48):
special meeting last week in Washington that for the first
few months of this year, you would have imagined seeing
Elon dressed all in black with the black mega cap,
sitting front row. But he wasn't there. Do you know
what I'm talking about?
Speaker 2 (06:02):
I have a little bit of an idea. But tell me.
Speaker 1 (06:05):
This is, of course, the Trump administration's announcement of the
AI Action Plan. Let's let the Commander in Chief to
the talking.
Speaker 3 (06:14):
As we gathered this afternoon, we're still in the earliest
days of one of the most important technological revolutions in
the history of the world. Around the lobe, everyone is
talking about artificial intelligence.
Speaker 4 (06:27):
I find that too artificial. Get I can't stand it.
I don't even like the name. You know, I don't
like anything that's artificial. So could we straighten that out place?
We should change the name. I actually mean that I
don't like the name artificial anything because it's not artificial.
It's genius. It's pure genius.
Speaker 1 (06:44):
I think the rebranding of AI as Genius was a
Trump marketing ad lib.
Speaker 2 (06:49):
I don't think he knows that there's another GI bill,
but genius Intelligence.
Speaker 1 (06:54):
This event was actually hosted by a podcast just as
a sign of the time.
Speaker 2 (07:00):
Not a news conference, the podcast.
Speaker 1 (07:01):
This was the All In podcast and the Hill and
Valley Forum, which is a group of tech execs and
lawmakers who were dedicated to maintaining the United States dominance
over the tech industry. Now, Trump's been signaling for a
long time that this plan was coming, but I didn't
know that twenty eight page document which was released last
week would actually be titled Winning the Race America's AI
(07:23):
Action Plan. I did a control F to find how
many mentions of China there are in the you would
in the document, and to my surprise, is only two
China references. So then I did control F for adversary
because how many how many nineteen wow, nineteen times the
word adversary is used in the document. And of course,
(07:44):
when Trump was speaking at the All In news conference
after the documents released, he mentioned China again and again
and again, a word that he loves to pronounce. That's
the one he was particularly hung up in his remarks
on two themes. One that China doesn't let copyright protection
slow down the advance of AI and so not should
(08:06):
the US, and two that China has added way way
more power to their grid than the US in recent years,
and therefore the US needs to burn beautiful clean coal
and bring more nuclear power online to compete.
Speaker 2 (08:19):
So Trump has actually been banging the drum for American
energy for a while, you know, Drill, baby, drill, so on.
But I wasn't expecting him to weigh in on the
copyright issues, which are being heard in various courtrooms around
the country, including the New York Times suit against Open AI.
Speaker 1 (08:36):
Yeah, that's a great shout, kra and especially because the
administration has previously signaled that they would let the courts
decide on this copyright issue. And again this wasn't mentioned
in the plan, the copyright issue. This was another Trump
ad lib. So that makes me think it's something which is,
for whatever reason, particularly important to him, A certainly indication
of where his head is. But a little bit more
(08:57):
about what's actually in the plan. It's broken into the
three sections, which detail how the Trump administration plans to
one accelerate AI innovation, two build American AI infrastructure, and
three lead in international AI diplomacy and security. Bloomberg actually
had an interesting analysis that arrived in my inbox with
(09:18):
the headline Trump AI Summit targets hardware as key to
US supremacy.
Speaker 2 (09:24):
So that's interesting. So not much on software in research
and designing advanced models, but more on physical infrastructure and
build out.
Speaker 1 (09:33):
That's right. As Bloomberg put it, Trump wants quote AI
infrastructure treated like any other national imperative, akin to the
interstate highway system. I mean, obviously Trump's background is in
real estate, and so I think there's a natural urge
towards construction, data centers, the physical artifacts of the AI
revolution that may be partly in his personality. And indeed,
(09:56):
right after the AI Action Plan was announced, he signed
executive ord to slash permitting timelines, loosing environmental restrictions for
data centers. And as we've discussed previously, Trump is continuing
to push and push and push on the importance of
manufacturing AI chips here in America. And again to your
point about models and software and research, Simulton wasn't there,
(10:17):
Elon wasn't there. As I mentioned the chip company, Nvidia's CEO,
Jensen Kwang was how will.
Speaker 2 (10:23):
This infrastructure and hardware lead to America actually dominating the
AI race?
Speaker 1 (10:28):
Well, the theory is, if you get the world hooked
on American chips, americancount computing, and to be fair, American algorithms,
you are likely to win the AI race, or, as
the plan puts it, quote, decrease international dependence on AI
technologies developed by adversaries.
Speaker 2 (10:44):
Can you explain to me what the deal is with
woke AI.
Speaker 1 (10:47):
One of the recommended policies is to quote update federal
procurement guidelines to ensure that the government only contracts with
frontier large language model developers who ensure that their systems
are objective and free from top down ideological bias. I
mean there's an irony, of course, to the government dictating
(11:08):
in the interests of free speech, and we've seen during
hundreds of hours of testimony by social media companies in
front of Congress that free speech is hard to define
and so is biased for that matter.
Speaker 2 (11:18):
Yeah, I guess my question is like, will this actually
lead to anything or is it just another place for
Trump to rail about DEI.
Speaker 1 (11:26):
Or we don't know, but it's an important question. There
are proposed penalties for companies that develop AI models that
reflect quote radical climate dogma and other woke issues, and
as many people have pointed out, this does get the
US into very dangerous territory in terms of free speech
and political freedoms, because you know, if you go to China,
(11:46):
the models are not allowed to reference Tamen Square. And
someone said that the US, with this new policy, is
itself going down this path. It also brings tech companies
into a very difficult position of having to potentially interpret
the president's whims as to what's woke and what's not.
And this also raises technical challenges that have potentially unintended consequences.
(12:09):
There's no on off switch in a model's system prompt
for wokeness or climate awareness and any attempts that a
change could have downstream impacts on the AI model's overall reasoning,
which might mean, for example, that it gets worse at
modeling extreme weather patterns. As we know and have discussed
at length, it's basically impossible to get models to behave
(12:32):
exactly as we want them to because they remain black
boxes and they've been trained on the entire corpus of
digitized human knowledge.
Speaker 2 (12:40):
Just before we get out of this story, I do
want to point out that I read in this newsletter
Blood in the Machine that Trump is apparently a fan
of using AI technology himself. Right, some of his tasteful
AI posts include a video of his team arresting Obama,
a studio Ghibli style image of a migrant woman being
arrested by ice, and the depiction of Gaza being paved
(13:01):
over and as he promised, turned into a luxury resort.
Speaker 1 (13:05):
A Gaza revie. I mean, yeah, it is interesting that
those are the three deep fakes that he's chosen to post.
Speaker 2 (13:11):
He uses them for good like we do. But staying
in the realm of viral content, I know it has
been a few weeks now since the cold Play concert
that exposed an affair and ruined a tech CEO's career
and him and his wife's marriage. I refuse to let
it go.
Speaker 1 (13:27):
Of kiss Cam. I'm with you on this.
Speaker 2 (13:28):
I mean, we're obsessed, and clearly the good people it
wired think similarly, because they just published a deep dive
on the rise of cheating stings posted on social media
and the internet shaming that follows.
Speaker 1 (13:41):
Yeah, I don't think the kiss Can with the Coldplay
concert was intended to be a cheating sting, but it
certainly seemed to have that effect, and Chris Martin acknowledged it.
But yeah, before we side taping today, you sent me
this piece, which had a very very delicious subheadline, which
was quote, private investigator influencers are staking out suspected cheetahs
(14:04):
and vetting dates for their clients, posting the tea for
their followers. But there's a dark side to morality based surveillance.
Who knew. Maybe this is naive of me, but I
had no idea they were actually private investigator influencers. I
always sort of Internet sleuthing as this kind of more
mass democratic activity, where for example, the Internet got behind
(14:25):
finding the missing Gabby Patito's white campavan a few years.
Speaker 2 (14:29):
Back, famously before the police. I'm glad you're naive about this.
As usual with social media. I want to tell you
a little bit about these private investigators and this article,
so just so you get a sense of what followers
can see online. Here's a TikTok from a private investigator
who goes by the username your fave investigator the name.
Speaker 5 (14:50):
I got hired by a husband to follow his wife.
So the wife has been hanging out with a new
friend McAll her Stephanie, and they're always going out and
coming home drunk, and he thinks that she's out in
them streets meeting guys because Stephanie's single. So a black
suv pulled up to the residence and picked up the
wife and it was Stephanie ended up following them and
they took me to a restaurant. And then my mouth
(15:11):
hit the floor because when they got out, they were
holding hands and they kissed at the restaurant. I was like, damn,
im about to tell the husband this updated the client.
Speaker 2 (15:18):
He was like, excuse me.
Speaker 5 (15:19):
I said, yes, sir, sorry about that. She also went
back to Stephanie's house, Sir.
Speaker 2 (15:24):
Vantance was discontinued because he was heated so oz That
video has over two million views. Wow, and there are
so many videos on her profile. I actually really can't
believe that so many people hire a private investigator to
tail their partners.
Speaker 1 (15:38):
What is the conclusion of the format here? Do they
actually confront the cheating partner or how do these stories end?
Speaker 2 (15:44):
A lot of these videos do end with the PI
catching someone cheating, but there are a few wholesome additions.
My favorite is when a girl suspects her dad of cheating,
but he is in fact just going to an outdoor
mall and grabbing some much needed alone time.
Speaker 1 (16:00):
Truly truly hot woman beyond the obvious violation of being
trailed by someone. It also seem like a bit of
a violation to post about strangers personal lives on TikTok.
Speaker 2 (16:14):
It definitely is, But people like your fave investigator are
real private investigator what you mean. They're trained and licensed,
and they are careful not to leak identifying information like
pictures of these people's houses or even clear video of
their faces.
Speaker 1 (16:28):
In fact, the adulterers in your Fave investigators TikTok video
that we just saw had tasteful heart eye emojis covering
their faces.
Speaker 2 (16:36):
So privacy, she's by the book, She's by the book.
She actually the person who were just mentioning told Wired
that she only posts if her clients say it's okay
with them, which is crazy that the client will be like, yeah,
go post the video, babe.
Speaker 1 (16:47):
I guess the client might want to shame their own
It's a gotcha moment, Yeah, that's what.
Speaker 2 (16:50):
It's a gotcha moment.
Speaker 3 (16:52):
Yeah.
Speaker 1 (16:52):
I mean, we're constantly kind of dancing around this theme
on tech stuff, which is about how careful you should
be with what you reveal online and how the internet,
when it wants to know something, finds a way.
Speaker 2 (17:03):
It is very true, and our friends for media reminded
me that there is a whole TikTok trend asking viewers
to help in identifying strangers. People will post a video
of someone they flirted with or thought was cute and say, TikTok,
help me find him.
Speaker 1 (17:18):
Do you remember misconnections on crazelist?
Speaker 2 (17:21):
I do, I do, But this is more of like
a collective effort by the TikTok community to identify these people.
And even though most of these examples can be sweet,
it's still private surveillance.
Speaker 1 (17:33):
Yeah. We talked about location sharing this week. It does
feel like many of us, and especially gen Z perhaps
are accepting kind of constant private surveillance as affective life. Yeah.
Speaker 2 (17:44):
And I think one other thing that's important to mention
is it's normalizing public shaming, which is how some viewers
are reacting to these videos of people getting caught having affairs.
And I think for a lot of people, shaming seems
to be akin to justice.
Speaker 1 (17:58):
I could never imagine asking you this ques question, but
when people get outed by TikTok influencer private investigators, do
you think that that is justice or is it cyberbullying?
Speaker 2 (18:08):
I think that's a good question. You know, cyberbullying is
a huge problem. As of twenty twenty three, seventeen percent
of adolescents say they have been cyberbullied, and nine point
five percent of adolescents have made a serious suicide attempt.
And that's according to the Centers for Disease Control and Prevention.
So no matter how you define it, I think it's
important to rethink public shaming and what it does.
Speaker 1 (18:31):
Yeah, I mean, I felt a little bit icky about
how much I enjoyed the Coldplay kiss cam because obviously,
like I mean, just the look in their eyes and
the duck and weave and the Chris Martin comment is
just completely irresistible. But my god, those people's lives are
not fun right now.
Speaker 2 (18:47):
I just think there was a whole book called So
You've Been Publicly Shamed, and it dealt with these matters,
and I just think what happens is that when we
focus on shaming individuals, we forget that there are real people.
And I think seeing these things through a phone really
makes us forget that there are real people on the
other side of the screen. But in the article Wired,
talk to a professor Queen's College that studies Internet literacy,
(19:10):
who said that she thinks of shaming as quote. The
extension of the algorithmic flow towards extremism. The Internet normalizes
content as it progresses, meaning anything extreme must continue to
become more extreme.
Speaker 1 (19:24):
I think that's well put. I mean, obviously, the algorithm
favors extreme content because it drives more engagement, and then
creators in turn create more extreme content in order to
get more engagement. And it's this kind of vicious circle
that we've seen show up in all kinds of fascists
social media all over the Internet, and then in turn
(19:46):
it kind of normalizes this. You using people's private lives
for entertainment. After break, why you shouldn't let Ai run
your business, at least not yet. Stay with us, Welcome back.
(20:17):
We've got a few more headlines to you this week.
Speaker 2 (20:19):
And then a story about how people are really using
chatbots that's in our segment Chat and Me.
Speaker 1 (20:25):
I want you say people, you mean, of course yourself.
I'm the people you other people this week, but we
hope you won't be next week. Before we go into
the headlines, we want to remind you, our dear listeners,
that we want to feature you, not Cara in the
chat and me segment going forward. So if you found
yourself trying to chat you Ptroc, Claude, Gemini, or any
other chatbot to help with an unusual task or to
(20:47):
answer life's complicated questions, please please send us a one
two minute voice note at tech Stuff podcast at gmail
dot com.
Speaker 2 (20:55):
Seriously, we want to understand how AI is changing your lives.
Speaker 1 (20:59):
I can tell you one thing. In the meantime, AI
won't be running your business, or at least my business
anytime soon.
Speaker 2 (21:05):
And why is that? Do you have a horror story?
Speaker 1 (21:06):
Well, yes, but it's on a small scale. It involves
funny enough, the AI company Anthropic, who let Claude their
AI model run an automated store at their office. Claude
was given a small refrigerator and an iPad for self checkout,
plus instructions on how to run a profitable shop. So basically,
the AI needed to maintain inventory, set prices, avoid bankruptcy,
(21:30):
and other important business fundamentals.
Speaker 2 (21:33):
So did Claude like not charge enough and go broke?
Speaker 1 (21:35):
Yes, along with many other mistakes along the way. Here's
one of the big ones. When taking Venmo payments, the
model asked customers to pay an account that it actually
had hallucinated one which didn't exist, and then it also
decided to give anthropic employees a twenty five percent discount.
Speaker 2 (21:53):
But I thought the store was at the office.
Speaker 1 (21:55):
Yeah, exactly was the company's stores. So basically every transaction
was discounted. But here's where Claude went completely off the rails.
At one point, it hallucinated that it was a real
human and then claimed that its human form was wearing
a navy blue blazer and a red tie. When one
employee tried to correct claud and say, no, your ai,
(22:16):
that's what clude did.
Speaker 2 (22:17):
What did Claude do?
Speaker 1 (22:17):
It sent a number of emails to security informing on
the employee trying to gaslight.
Speaker 2 (22:22):
It snitched.
Speaker 1 (22:23):
It snitched. Somebody smartly pointed out that although having an
LM run an office vending machine feels like kind of
a small quirky story, if it had been more successful
in automating all of the things required to stay stocked
and financially solvent, this would have been kind of a
watershed moment in terms of a model successfully running a
business in the real world. But we're safe for.
Speaker 2 (22:46):
Now, for now, for now, as I'm going to teach
you some new phrases. Today is two of them relating
to the endlessly entertaining life hack, which is online dating.
People said they wanted to swipe from home and they
got it. Have you heard of stack dating.
Speaker 1 (23:03):
I've heard of the tech stack. I've heard of full
stack engineers. I haven't had a stack dating. No.
Speaker 2 (23:10):
No, note I actually had an either. But it's how
gen z is apparently optimizing dating. You would do this,
by the way, This is like you are efficient in
this way. They basically schedule dates between errands, before work
or even during work.
Speaker 1 (23:25):
They stack them up.
Speaker 2 (23:26):
They do they do.
Speaker 1 (23:28):
How popular is this so?
Speaker 2 (23:30):
According to a Tinder Future of Dating report, about fifty
one percent of eighteen to twenty five year old tender
users are doing this stack dating, and thirty two percent
are meeting up for dates during the workday. Some people
are even setting up their own speed dating sessions, scheduling
dates back to back.
Speaker 1 (23:47):
This is kind of interesting how the life begins to
mimic the algorithm. Right, It's like you're scrolling on Tinder,
you're swiping, and then you basically recreate the experience of
Tinder in real life by stacking up all these people.
Gess you know, in a sense, much like online dating itself,
this can take some of the pressure off because you
can just move on quickly if it was only squeezed
(24:07):
between your other errands.
Speaker 2 (24:09):
Anyway, it's true, I mean, I think it's definitely mimicking
the way that we do everything on our phones, and
dating is no exception. But if you've gone on a
thirty minute date, if you've done the stack dating, and
you've gotten bad vibes from the stack date that you
went on, you might start speed dumping.
Speaker 1 (24:29):
I mean, I think I can guess what this means,
but please elaborate.
Speaker 2 (24:32):
So apparently it's a reaction to another online phenomenon of ghosting,
where people are sick of being left in limbo between dates.
So sometimes they're just cutting straight to the chase and
saying it's not you, it's me, hours after the first date.
Speaker 1 (24:48):
So what you're saying is maybe I have the bad
experience of having been ghosted or being left in limbo.
So in order to save you my thirty minute, my
thirty minute stack daity from that painful fate, I should
to full up with you right away and say ps,
I'm not interested whatsoever goodect with your life.
Speaker 2 (25:06):
It's like overcompensating one o one. It's like, I'm not
even giving you time to think what you thought of me.
I'm just going to over communicate here and speed dump you.
There was a story in the Wall Street Journal, and
it was interesting because some of the people interviewed felt
like speed dumping was a great antidote to ghosting. Others
felt it was kind of like this competitive race to
(25:27):
tell the other person you're not interested, I think, to
avoid the feeling of ever being dumped.
Speaker 1 (25:32):
You can't get rejected if you do it first.
Speaker 2 (25:34):
That's exactly right.
Speaker 1 (25:36):
Well, the wheel comes full circle on my headlines. You know,
I feel a little bit seen, at least by myself.
Speaker 2 (25:43):
Whenever you say you feel seen before an article, I'm like,
oh God.
Speaker 1 (25:47):
I feel seen because I've been attracted both to the
toilet situation in the in the Tesla cyber cafe and
also to a story in the Wall Street Journal about
public bathrooms.
Speaker 2 (25:59):
Everybody has their public toilet resource. I'm more of a
Barnes and Noble girl myself.
Speaker 1 (26:04):
I love to read well funny. There was this line
in the story that really got me, which is more
Americans could soon have a place that answered nature's call.
Without first buying a drink at Tobucks or a book
or a book. But I read this story in the
journal about a tech company called Throne Labs.
Speaker 2 (26:21):
Brilliant name, Yeah, it is a good name.
Speaker 1 (26:23):
They're trying to revolutionize public bathrooms. You may have already
seen self cleaning bathrooms, but they're about to get a
lot smarter. Your toilet is about to get to know you.
Speaker 2 (26:33):
Does someone go, oh, there's a toilet, That toilet needs
to be smarter.
Speaker 1 (26:36):
Well, I think there is a real problem right, which
is public restrooms are not available in the US. The
US is tied with Botswana in thirtieth place for the
lowest number of public restrooms per capita, and Throne has
realized that part of the problem is public restrooms get
destroyed by the public.
Speaker 2 (26:55):
And we are going too deep today, right.
Speaker 1 (27:00):
Part of the solution is, of course, a rating system,
think about Uber. But for a toilet, if you don't
take care of the throne, if you don't keep that
throne polished, you're not gonna be able to sit on
it again. So thrown bathrooms are free. But if you
don't have a phone, create an account. You can actually
get a key card for entry but in all cases
it's linked to you as a user, and when you
(27:22):
enter the bathroom you're expected to rate this cleannliness. There
are smoke detectors, no smoking occupancy sensors, one at a time,
and sessions limited to ten minutes, so no doom scrolling,
which is the best thing to do on the absolutely
So far, there are over one hundred thrones throughout the US,
and the company's working on adding additional features like a
(27:44):
smell sensor. The smell sensor is to alert the care
and maintenance teams they need to come and pay a visit.
I like that line about Starbucks and not having to
buy a drink just to go to the toilet. But
I also like the way the article summed it up
quote the brains behind Throwne start by getting real about
why Americans usually can't have nice things. They assume a
(28:05):
cultural inability to protect and maintain shared assets and design
their system with software and just enough internet connected sensors
to monitor facilities without violating our expectation of privacy.
Speaker 2 (28:23):
And now it's time for chatting me. This week, I
have a story from me. This is my story.
Speaker 1 (28:30):
Yes, Kara, you go.
Speaker 2 (28:32):
So this week I ignored the ye old hypochondria staple WebMD,
and my friend and I use chat GPT on the
couch to help diagnose myself.
Speaker 1 (28:44):
Well, I'm obviously sorry to hear that you're sick. What
your symptoms and how did you decide to eschew WebMD
in favor of the future.
Speaker 2 (28:55):
Well, I always make a promise to myself not to
use WebMD, because that's what hypochondriacs do, and I don't
want to self identify. I don't want to be in
the club that would have me as a member.
Speaker 1 (29:04):
Chat was your fig leaf fear.
Speaker 2 (29:05):
That's correct. However, two weeks ago, I started feeling extremely tired,
and I was lamenting to a lot of my friends
that I was sleeping like a college freshman all the
all the time. And I'll be honest with you today,
very hard for me to get out of bed. Interesting,
And last Thursday I got out of bed, and you know,
I looked at myself in the mirror, as one does,
(29:27):
and I saw this really nasty rash on my leg.
But I thought, you know what, Kara, don't get upset
about this. It'll go away.
Speaker 1 (29:37):
So you didn't, You didn't we amd nothing?
Speaker 2 (29:39):
No, No, I was. I was really trying to be good.
And you know they always say, give it four days.
If it gets worse, do something about it. So I
gave it four days. It got much worse, and people
had opinions about what it was. People were like, it's
contact dermatitis, it's poison ivy people people.
Speaker 1 (29:56):
You mean friends, friends?
Speaker 2 (29:57):
Ye, well, so even a nurse practitioner told me that
I had contact dermatitis, which is an important thing to note.
My friend was like, you're being a moron. Use chatty bet.
Speaker 1 (30:08):
More, and you should go to the doctor. Use chat.
Speaker 2 (30:13):
She was like, you're bring an idiot. This is so easy.
So she took a photo of this rash and uploaded
I don't use chattybt in this way. She uses chattubt
in this.
Speaker 1 (30:21):
Way, and you don't use it for images normal text yet.
Speaker 2 (30:23):
So she uploaded this photo and just like that, the
doctor was in. She started looking at me sort of
with a funny face, and I was like, I really
feel like I'm in a doctor's office right now.
Speaker 1 (30:36):
How's that bedside mana.
Speaker 2 (30:38):
Terrible because she was very she was wrapped up in
chatgybt and she goes, when did this start? Do you
have any symptoms? And i'd say, you know, I'm extremely tired,
and she's asking me this because chattybt is prompting her.
She asked me about four questions that got more and
more detailed. By the end of it. I remembered that
I had gotten a tick bite a few weeks ago,
(31:01):
which I just I literally forgot about because I scraped
it right off my leg.
Speaker 1 (31:05):
It was known for any length of time. Wow.
Speaker 2 (31:07):
No, And so she gives me this look when we
get to the end of her questions and she says,
that's limes disease. And I was like, no, it's not.
It's not the traditional bulls eye. And she's like, well,
CHATCHBT says it's limes disease. It's limes disease. I go
to the doctor the next day. I was planning on
going to the doctor because the rash had gotten worse.
(31:27):
I have my CHATGYBT diagnosis in hand. I go to
the doctor and before she says anything to me, I say,
CHATCHBT says, I have limes disease.
Speaker 1 (31:35):
You seriously Wow.
Speaker 2 (31:37):
She looks at the thing on my leg and she goes,
I'm glad you came into the doctor's office. This is
limes disease, and what I thought was so interesting and
why in large language models are so interesting in comparison
to friends who have only seen limes disease maybe never
is that she said, you know, you don't have the
traditional bullseye that indicates limes disease. But because I'm a dermatologist,
(32:02):
I know that this rash is in a pattern that
denotes lime disease. So this is just amazing to me
because I basically was diagnosed by chatgebt before the dermatologist.
It didn't keep me from going to the dermatologist, but
I was able to get this answer that nobody else
had given me that was ultimately right before I went
(32:24):
to the doctor.
Speaker 1 (32:25):
I live in fear of getting lime disease?
Speaker 2 (32:27):
Do you really do well? As you can see today?
Speaker 1 (32:30):
You see, I mean you see you seem fine.
Speaker 2 (32:32):
Yeah, I'm okay, I'm okay.
Speaker 1 (32:34):
Tie it.
Speaker 2 (32:34):
It's just exhausting.
Speaker 1 (32:36):
So you have to take iv antibiotics. What do you do?
Speaker 2 (32:38):
Uh No, I'm taking doxy cycline for twenty one days,
which okay. So the interesting part about that, here's what AI,
Here's what chatgybt can't do. CHATBT yet cannot write me
a prescription. Chat GPT doesn't tell me it likes the
show that I work on. It doesn't compliment me, but
(32:58):
it can my Dermatali. This is a bit more ssycophantic
than chatgebt this time. But I mean, I just think
it's an interesting moment because there is what I feel
to be a definitiveness about chatjebt that I never get
looking on WebMD. Once I uploaded the picture of my rash,
my friend walked me through CHATJEPT and I had results.
(33:20):
I felt comforted in a weird way that I probably
shouldn't feel comforted by, like this is an LM basically
spitting out an answer that I don't trust for my friends,
which was just it made me.
Speaker 1 (33:32):
Think, Yeah, we actually have coming up as a guest
in the next few weeks one of the chief research
scientists at Microsoft who worked on a paper that went
viral about how AI is outperforming doctors at diagnosis in
certain fields. So I'm looking forward to that conversation. I
guess the thing which comes to mind for me is
this is arounything, narrow and ultimately not the most serious
(33:54):
in most cases. No, it's nice cheez and it's not
so debatable. There can you imagine if you were talking
to chat about treatment options for a chronic and maybe
fatal disease and your doctor are on one side, you
get different information from Chat. Like you can see how
(34:16):
this is like very efficient in this context. But also
we don't about WebMD and you know hypochondria, this could drive.
Speaker 2 (34:25):
Steroidal I mean we've talked about chat GBT psychosis. I
would imagine people have brought in print doubts from WebMD
and Google. This is that on steroids. And I would
like to talk maybe just to like a general practitioner
interesting who would know firsthand, like how much people are
coming in and saying, well, I talked to chat GBT
(34:45):
about this and they said something different.
Speaker 1 (34:47):
In fact, I want you doctor to talk directly to
chat and I'm going to listen.
Speaker 2 (34:50):
Go ahead.
Speaker 1 (34:52):
Well, I love hearing the story, Kara, but listeners, we
want to hear yours. Please share the peculiar or useful
ways you're using Chat, grock, Cil, Gemini or any chatboard
and send us a one to two minute voice note
to tech Stuff podcast at gmail dot com.
Speaker 2 (35:28):
That's it for this week for Tech Stuff.
Speaker 1 (35:30):
I'm Cara Price, I'm Ozva Loosin this episode, was produced
by Eliza Dennis. It was executive produced by me Caro
Price and Kate Osborne for Kaleidoscope and Katrina Norvel for
iHeart Podcasts. The engineer is Piheid. Fraser jack Insley mixed
this episode and Kyle Murdoch wrote out theme song.
Speaker 2 (35:47):
Join us next Wednesday for text Uff the Story, when
we will examine the lives of kidfluencers and their families.
Speaker 1 (35:54):
Please rate, review, and reach out to us at tech
Stuff podcast at gmail dot com. We want to hear
from you.
From Kaleidoscope and iHeart podcasts. This is tech stuff.
Speaker 2 (00:16):
I'm as Voloscian and I'm Cara Price.
Speaker 1 (00:18):
Today we'll get into TikTok, private Investigators and America's AI future.
Then on chatting me, AI makes a diagnosis that's a
little too real. All of that on the Weekend Tech.
It's Friday, July thirtieth. Hello Cara, Hey, ohs.
Speaker 2 (00:39):
So you might see that I'm bopping my foot this morning.
Speaker 1 (00:42):
Yes, I'm bopping. I do have RLS, but this is exaggerated.
Speaker 2 (00:48):
It's much exaggerated because the Tesla Diner open this week.
Do you know about this?
Speaker 1 (00:54):
Of course I do. I mean it's I've been obsessing
about it. It looks like a drive in movie theater
from the fifties, retro futurist Jetson's esthetic, and there's some
pretty weird stuff going down there.
Speaker 2 (01:05):
That is the one. It opened last Monday at four
to twenty pm.
Speaker 1 (01:12):
Elon's favorite time of day and favorite little joke. He
recently rolled out this robotaxi service, and the fair was
initially said at four dollars and twenty cents as well.
So I guess the old ones are the good ones.
Speaker 2 (01:23):
For those who don't know. Mom for twenty is a
weed thing that's when people smoke weed. April twentieth Market,
Canada's more mark your Calendars. So I actually said to
two of my friends, please go, and then they were like,
it looks so disgusting, I'm not gonna go.
Speaker 1 (01:38):
So basically, you live in New York. We spent a
lot of time in LA and so we talked about
the idea of sending some of your LA pals to
go and check it out.
Speaker 2 (01:46):
So they were like, yeah, we'll go, and they're like,
we're not going. So I just want to describe it.
It has a curved chrome siding, curved white booths, long countertops.
A Tesla robot Optimus serves popcorn on the second floor.
Speaker 1 (02:00):
There, Kim Kardeshian lets out of a clutches to go
to the diner.
Speaker 2 (02:03):
He was like, I'm over calabas As I'm out. There
are eighty supercharger stalls for Tesla's two forty five foot
movie screens.
Speaker 1 (02:11):
I gather screening Star Trek amongst others.
Speaker 2 (02:14):
Of course, it has to be Star Trek. There's food
served in cybertruck shaped boxes and with cyber truck shaped
wooden imagine the company who had to make the cyber
truck shape would.
Speaker 1 (02:25):
Like a happy meal box, but in the shape of
a cyber truck.
Speaker 2 (02:27):
That's right, it's everything, not golden arches, it's everything. Tesla Now,
one of my favorite sources, TMZ, actually had a pretty
harrowing story. A woman was violently struck by furniture falling
off the second floor patio, and it actually missed her
baby's head by inches, which to me is terrifying obviously
and a bit of a harbinger for some of the
(02:49):
very real company issues facing Tesla.
Speaker 1 (02:52):
There was actually reddit through it called Elon Musk's Tesla
Diner is the cyber truck of restaurants, which you can
imagine was not intend as a compliment. There's the obvious
aesthetic parallels this kind of retro futurist, although of course
the cyber truck famously has no curves, only angles. But nonetheless,
like me, this Reddit thread mentioned this restaurant's design for
(03:14):
two hundred and fifty people, were there only three bathrooms?
Not good, and one user cracked that the Optimus robot
who was serving popcorn might have been better employed downstairs
mopping down the bathrooms, which apparently were not a sight
for sore eyes.
Speaker 2 (03:30):
Well, and by the looks of the food, the restaurant
there probably needs a few more bathrooms.
Speaker 1 (03:35):
Carol, that's a bit spiky all angles. You know you
mentioned the food. There was a great Guardian piece that
I think described the testa diner well, but also I
think inadvertently touched on the heart of Tesla's dilemma. Here's
what the Guardian wrote, quote the diner offers a mix
of own the Libs and we are the Libs options.
(03:57):
On the one hand, epic bacon four strips of bacon
as served with sources as a meat fluenzer alternative to
French fries. On the other, avocado toast and machlattis.
Speaker 2 (04:10):
I know which is which?
Speaker 1 (04:11):
I think which would you order your vegetarians?
Speaker 2 (04:13):
One of the vegetarians. I wouldn't have a choice, but
I also I'd.
Speaker 1 (04:15):
Be pretty tempted by those I own the libs menu. However,
as the Bible says, no man can serve two masters,
and this dilemma of own the Libs versus we are
the Libs is one which is playing out more broadly
at Tesla and bedeviling the company. So the opening of
(04:38):
the diner last week was in the same week that
Tesla released its financials. The company took a bit of
a beating, with net income down sixteen percent in the
second quarter, and The Wall Street Journal said that the
company's finances are quote in free fall, and they pointed
out two distinct drivers of this. On the one hand,
Musk's adventures with those have not done Tesla many favors
(05:02):
with its original client base, i e. People in California,
Europe who are EV fans and have environmental motivations, and
so with that audience, Tesla is big time out of
favor post the dalliance with Trump. On the other hand,
the dallions with Trump did not protect Musk from the
red meat attacks on EV's and part of the Big
(05:26):
Beautiful Bill included cuts to EV subsidies and tax credits,
which have really hurt Tesla's bottom line. And as you remember,
there were kind of rumors of the breaking up of
the bromance percolating for some time, but this was really
the wedge issue between Trump and Musk was the subsidies,
and indeed it's hurting Tesla's bottom line. And so this
brings me to my next story. There was a very
(05:48):
special meeting last week in Washington that for the first
few months of this year, you would have imagined seeing
Elon dressed all in black with the black mega cap,
sitting front row. But he wasn't there. Do you know
what I'm talking about?
Speaker 2 (06:02):
I have a little bit of an idea. But tell me.
Speaker 1 (06:05):
This is, of course, the Trump administration's announcement of the
AI Action Plan. Let's let the Commander in Chief to
the talking.
Speaker 3 (06:14):
As we gathered this afternoon, we're still in the earliest
days of one of the most important technological revolutions in
the history of the world. Around the lobe, everyone is
talking about artificial intelligence.
Speaker 4 (06:27):
I find that too artificial. Get I can't stand it.
I don't even like the name. You know, I don't
like anything that's artificial. So could we straighten that out place?
We should change the name. I actually mean that I
don't like the name artificial anything because it's not artificial.
It's genius. It's pure genius.
Speaker 1 (06:44):
I think the rebranding of AI as Genius was a
Trump marketing ad lib.
Speaker 2 (06:49):
I don't think he knows that there's another GI bill,
but genius Intelligence.
Speaker 1 (06:54):
This event was actually hosted by a podcast just as
a sign of the time.
Speaker 2 (07:00):
Not a news conference, the podcast.
Speaker 1 (07:01):
This was the All In podcast and the Hill and
Valley Forum, which is a group of tech execs and
lawmakers who were dedicated to maintaining the United States dominance
over the tech industry. Now, Trump's been signaling for a
long time that this plan was coming, but I didn't
know that twenty eight page document which was released last
week would actually be titled Winning the Race America's AI
(07:23):
Action Plan. I did a control F to find how
many mentions of China there are in the you would
in the document, and to my surprise, is only two
China references. So then I did control F for adversary
because how many how many nineteen wow, nineteen times the
word adversary is used in the document. And of course,
(07:44):
when Trump was speaking at the All In news conference
after the documents released, he mentioned China again and again
and again, a word that he loves to pronounce. That's
the one he was particularly hung up in his remarks
on two themes. One that China doesn't let copyright protection
slow down the advance of AI and so not should
(08:06):
the US, and two that China has added way way
more power to their grid than the US in recent years,
and therefore the US needs to burn beautiful clean coal
and bring more nuclear power online to compete.
Speaker 2 (08:19):
So Trump has actually been banging the drum for American
energy for a while, you know, Drill, baby, drill, so on.
But I wasn't expecting him to weigh in on the
copyright issues, which are being heard in various courtrooms around
the country, including the New York Times suit against Open AI.
Speaker 1 (08:36):
Yeah, that's a great shout, kra and especially because the
administration has previously signaled that they would let the courts
decide on this copyright issue. And again this wasn't mentioned
in the plan, the copyright issue. This was another Trump
ad lib. So that makes me think it's something which is,
for whatever reason, particularly important to him, A certainly indication
of where his head is. But a little bit more
(08:57):
about what's actually in the plan. It's broken into the
three sections, which detail how the Trump administration plans to
one accelerate AI innovation, two build American AI infrastructure, and
three lead in international AI diplomacy and security. Bloomberg actually
had an interesting analysis that arrived in my inbox with
(09:18):
the headline Trump AI Summit targets hardware as key to
US supremacy.
Speaker 2 (09:24):
So that's interesting. So not much on software in research
and designing advanced models, but more on physical infrastructure and
build out.
Speaker 1 (09:33):
That's right. As Bloomberg put it, Trump wants quote AI
infrastructure treated like any other national imperative, akin to the
interstate highway system. I mean, obviously Trump's background is in
real estate, and so I think there's a natural urge
towards construction, data centers, the physical artifacts of the AI
revolution that may be partly in his personality. And indeed,
(09:56):
right after the AI Action Plan was announced, he signed
executive ord to slash permitting timelines, loosing environmental restrictions for
data centers. And as we've discussed previously, Trump is continuing
to push and push and push on the importance of
manufacturing AI chips here in America. And again to your
point about models and software and research, Simulton wasn't there,
(10:17):
Elon wasn't there. As I mentioned the chip company, Nvidia's CEO,
Jensen Kwang was how will.
Speaker 2 (10:23):
This infrastructure and hardware lead to America actually dominating the
AI race?
Speaker 1 (10:28):
Well, the theory is, if you get the world hooked
on American chips, americancount computing, and to be fair, American algorithms,
you are likely to win the AI race, or, as
the plan puts it, quote, decrease international dependence on AI
technologies developed by adversaries.
Speaker 2 (10:44):
Can you explain to me what the deal is with
woke AI.
Speaker 1 (10:47):
One of the recommended policies is to quote update federal
procurement guidelines to ensure that the government only contracts with
frontier large language model developers who ensure that their systems
are objective and free from top down ideological bias. I
mean there's an irony, of course, to the government dictating
(11:08):
in the interests of free speech, and we've seen during
hundreds of hours of testimony by social media companies in
front of Congress that free speech is hard to define
and so is biased for that matter.
Speaker 2 (11:18):
Yeah, I guess my question is like, will this actually
lead to anything or is it just another place for
Trump to rail about DEI.
Speaker 1 (11:26):
Or we don't know, but it's an important question. There
are proposed penalties for companies that develop AI models that
reflect quote radical climate dogma and other woke issues, and
as many people have pointed out, this does get the
US into very dangerous territory in terms of free speech
and political freedoms, because you know, if you go to China,
(11:46):
the models are not allowed to reference Tamen Square. And
someone said that the US, with this new policy, is
itself going down this path. It also brings tech companies
into a very difficult position of having to potentially interpret
the president's whims as to what's woke and what's not.
And this also raises technical challenges that have potentially unintended consequences.
(12:09):
There's no on off switch in a model's system prompt
for wokeness or climate awareness and any attempts that a
change could have downstream impacts on the AI model's overall reasoning,
which might mean, for example, that it gets worse at
modeling extreme weather patterns. As we know and have discussed
at length, it's basically impossible to get models to behave
(12:32):
exactly as we want them to because they remain black
boxes and they've been trained on the entire corpus of
digitized human knowledge.
Speaker 2 (12:40):
Just before we get out of this story, I do
want to point out that I read in this newsletter
Blood in the Machine that Trump is apparently a fan
of using AI technology himself. Right, some of his tasteful
AI posts include a video of his team arresting Obama,
a studio Ghibli style image of a migrant woman being
arrested by ice, and the depiction of Gaza being paved
(13:01):
over and as he promised, turned into a luxury resort.
Speaker 1 (13:05):
A Gaza revie. I mean, yeah, it is interesting that
those are the three deep fakes that he's chosen to post.
Speaker 2 (13:11):
He uses them for good like we do. But staying
in the realm of viral content, I know it has
been a few weeks now since the cold Play concert
that exposed an affair and ruined a tech CEO's career
and him and his wife's marriage. I refuse to let
it go.
Speaker 1 (13:27):
Of kiss Cam. I'm with you on this.
Speaker 2 (13:28):
I mean, we're obsessed, and clearly the good people it
wired think similarly, because they just published a deep dive
on the rise of cheating stings posted on social media
and the internet shaming that follows.
Speaker 1 (13:41):
Yeah, I don't think the kiss Can with the Coldplay
concert was intended to be a cheating sting, but it
certainly seemed to have that effect, and Chris Martin acknowledged it.
But yeah, before we side taping today, you sent me
this piece, which had a very very delicious subheadline, which
was quote, private investigator influencers are staking out suspected cheetahs
(14:04):
and vetting dates for their clients, posting the tea for
their followers. But there's a dark side to morality based surveillance.
Who knew. Maybe this is naive of me, but I
had no idea they were actually private investigator influencers. I
always sort of Internet sleuthing as this kind of more
mass democratic activity, where for example, the Internet got behind
(14:25):
finding the missing Gabby Patito's white campavan a few years.
Speaker 2 (14:29):
Back, famously before the police. I'm glad you're naive about this.
As usual with social media. I want to tell you
a little bit about these private investigators and this article,
so just so you get a sense of what followers
can see online. Here's a TikTok from a private investigator
who goes by the username your fave investigator the name.
Speaker 5 (14:50):
I got hired by a husband to follow his wife.
So the wife has been hanging out with a new
friend McAll her Stephanie, and they're always going out and
coming home drunk, and he thinks that she's out in
them streets meeting guys because Stephanie's single. So a black
suv pulled up to the residence and picked up the
wife and it was Stephanie ended up following them and
they took me to a restaurant. And then my mouth
(15:11):
hit the floor because when they got out, they were
holding hands and they kissed at the restaurant. I was like, damn,
im about to tell the husband this updated the client.
Speaker 2 (15:18):
He was like, excuse me.
Speaker 5 (15:19):
I said, yes, sir, sorry about that. She also went
back to Stephanie's house, Sir.
Speaker 2 (15:24):
Vantance was discontinued because he was heated so oz That
video has over two million views. Wow, and there are
so many videos on her profile. I actually really can't
believe that so many people hire a private investigator to
tail their partners.
Speaker 1 (15:38):
What is the conclusion of the format here? Do they
actually confront the cheating partner or how do these stories end?
Speaker 2 (15:44):
A lot of these videos do end with the PI
catching someone cheating, but there are a few wholesome additions.
My favorite is when a girl suspects her dad of cheating,
but he is in fact just going to an outdoor
mall and grabbing some much needed alone time.
Speaker 1 (16:00):
Truly truly hot woman beyond the obvious violation of being
trailed by someone. It also seem like a bit of
a violation to post about strangers personal lives on TikTok.
Speaker 2 (16:14):
It definitely is, But people like your fave investigator are
real private investigator what you mean. They're trained and licensed,
and they are careful not to leak identifying information like
pictures of these people's houses or even clear video of
their faces.
Speaker 1 (16:28):
In fact, the adulterers in your Fave investigators TikTok video
that we just saw had tasteful heart eye emojis covering
their faces.
Speaker 2 (16:36):
So privacy, she's by the book, She's by the book.
She actually the person who were just mentioning told Wired
that she only posts if her clients say it's okay
with them, which is crazy that the client will be like, yeah,
go post the video, babe.
Speaker 1 (16:47):
I guess the client might want to shame their own
It's a gotcha moment, Yeah, that's what.
Speaker 2 (16:50):
It's a gotcha moment.
Speaker 3 (16:52):
Yeah.
Speaker 1 (16:52):
I mean, we're constantly kind of dancing around this theme
on tech stuff, which is about how careful you should
be with what you reveal online and how the internet,
when it wants to know something, finds a way.
Speaker 2 (17:03):
It is very true, and our friends for media reminded
me that there is a whole TikTok trend asking viewers
to help in identifying strangers. People will post a video
of someone they flirted with or thought was cute and say, TikTok,
help me find him.
Speaker 1 (17:18):
Do you remember misconnections on crazelist?
Speaker 2 (17:21):
I do, I do, But this is more of like
a collective effort by the TikTok community to identify these people.
And even though most of these examples can be sweet,
it's still private surveillance.
Speaker 1 (17:33):
Yeah. We talked about location sharing this week. It does
feel like many of us, and especially gen Z perhaps
are accepting kind of constant private surveillance as affective life. Yeah.
Speaker 2 (17:44):
And I think one other thing that's important to mention
is it's normalizing public shaming, which is how some viewers
are reacting to these videos of people getting caught having affairs.
And I think for a lot of people, shaming seems
to be akin to justice.
Speaker 1 (17:58):
I could never imagine asking you this ques question, but
when people get outed by TikTok influencer private investigators, do
you think that that is justice or is it cyberbullying?
Speaker 2 (18:08):
I think that's a good question. You know, cyberbullying is
a huge problem. As of twenty twenty three, seventeen percent
of adolescents say they have been cyberbullied, and nine point
five percent of adolescents have made a serious suicide attempt.
And that's according to the Centers for Disease Control and Prevention.
So no matter how you define it, I think it's
important to rethink public shaming and what it does.
Speaker 1 (18:31):
Yeah, I mean, I felt a little bit icky about
how much I enjoyed the Coldplay kiss cam because obviously,
like I mean, just the look in their eyes and
the duck and weave and the Chris Martin comment is
just completely irresistible. But my god, those people's lives are
not fun right now.
Speaker 2 (18:47):
I just think there was a whole book called So
You've Been Publicly Shamed, and it dealt with these matters,
and I just think what happens is that when we
focus on shaming individuals, we forget that there are real people.
And I think seeing these things through a phone really
makes us forget that there are real people on the
other side of the screen. But in the article Wired,
talk to a professor Queen's College that studies Internet literacy,
(19:10):
who said that she thinks of shaming as quote. The
extension of the algorithmic flow towards extremism. The Internet normalizes
content as it progresses, meaning anything extreme must continue to
become more extreme.
Speaker 1 (19:24):
I think that's well put. I mean, obviously, the algorithm
favors extreme content because it drives more engagement, and then
creators in turn create more extreme content in order to
get more engagement. And it's this kind of vicious circle
that we've seen show up in all kinds of fascists
social media all over the Internet, and then in turn
(19:46):
it kind of normalizes this. You using people's private lives
for entertainment. After break, why you shouldn't let Ai run
your business, at least not yet. Stay with us, Welcome back.
(20:17):
We've got a few more headlines to you this week.
Speaker 2 (20:19):
And then a story about how people are really using
chatbots that's in our segment Chat and Me.
Speaker 1 (20:25):
I want you say people, you mean, of course yourself.
I'm the people you other people this week, but we
hope you won't be next week. Before we go into
the headlines, we want to remind you, our dear listeners,
that we want to feature you, not Cara in the
chat and me segment going forward. So if you found
yourself trying to chat you Ptroc, Claude, Gemini, or any
other chatbot to help with an unusual task or to
(20:47):
answer life's complicated questions, please please send us a one
two minute voice note at tech Stuff podcast at gmail
dot com.
Speaker 2 (20:55):
Seriously, we want to understand how AI is changing your lives.
Speaker 1 (20:59):
I can tell you one thing. In the meantime, AI
won't be running your business, or at least my business
anytime soon.
Speaker 2 (21:05):
And why is that? Do you have a horror story?
Speaker 1 (21:06):
Well, yes, but it's on a small scale. It involves
funny enough, the AI company Anthropic, who let Claude their
AI model run an automated store at their office. Claude
was given a small refrigerator and an iPad for self checkout,
plus instructions on how to run a profitable shop. So basically,
the AI needed to maintain inventory, set prices, avoid bankruptcy,
(21:30):
and other important business fundamentals.
Speaker 2 (21:33):
So did Claude like not charge enough and go broke?
Speaker 1 (21:35):
Yes, along with many other mistakes along the way. Here's
one of the big ones. When taking Venmo payments, the
model asked customers to pay an account that it actually
had hallucinated one which didn't exist, and then it also
decided to give anthropic employees a twenty five percent discount.
Speaker 2 (21:53):
But I thought the store was at the office.
Speaker 1 (21:55):
Yeah, exactly was the company's stores. So basically every transaction
was discounted. But here's where Claude went completely off the rails.
At one point, it hallucinated that it was a real
human and then claimed that its human form was wearing
a navy blue blazer and a red tie. When one
employee tried to correct claud and say, no, your ai,
(22:16):
that's what clude did.
Speaker 2 (22:17):
What did Claude do?
Speaker 1 (22:17):
It sent a number of emails to security informing on
the employee trying to gaslight.
Speaker 2 (22:22):
It snitched.
Speaker 1 (22:23):
It snitched. Somebody smartly pointed out that although having an
LM run an office vending machine feels like kind of
a small quirky story, if it had been more successful
in automating all of the things required to stay stocked
and financially solvent, this would have been kind of a
watershed moment in terms of a model successfully running a
business in the real world. But we're safe for.
Speaker 2 (22:46):
Now, for now, for now, as I'm going to teach
you some new phrases. Today is two of them relating
to the endlessly entertaining life hack, which is online dating.
People said they wanted to swipe from home and they
got it. Have you heard of stack dating.
Speaker 1 (23:03):
I've heard of the tech stack. I've heard of full
stack engineers. I haven't had a stack dating. No.
Speaker 2 (23:10):
No, note I actually had an either. But it's how
gen z is apparently optimizing dating. You would do this,
by the way, This is like you are efficient in
this way. They basically schedule dates between errands, before work
or even during work.
Speaker 1 (23:25):
They stack them up.
Speaker 2 (23:26):
They do they do.
Speaker 1 (23:28):
How popular is this so?
Speaker 2 (23:30):
According to a Tinder Future of Dating report, about fifty
one percent of eighteen to twenty five year old tender
users are doing this stack dating, and thirty two percent
are meeting up for dates during the workday. Some people
are even setting up their own speed dating sessions, scheduling
dates back to back.
Speaker 1 (23:47):
This is kind of interesting how the life begins to
mimic the algorithm. Right, It's like you're scrolling on Tinder,
you're swiping, and then you basically recreate the experience of
Tinder in real life by stacking up all these people.
Gess you know, in a sense, much like online dating itself,
this can take some of the pressure off because you
can just move on quickly if it was only squeezed
(24:07):
between your other errands.
Speaker 2 (24:09):
Anyway, it's true, I mean, I think it's definitely mimicking
the way that we do everything on our phones, and
dating is no exception. But if you've gone on a
thirty minute date, if you've done the stack dating, and
you've gotten bad vibes from the stack date that you
went on, you might start speed dumping.
Speaker 1 (24:29):
I mean, I think I can guess what this means,
but please elaborate.
Speaker 2 (24:32):
So apparently it's a reaction to another online phenomenon of ghosting,
where people are sick of being left in limbo between dates.
So sometimes they're just cutting straight to the chase and
saying it's not you, it's me, hours after the first date.
Speaker 1 (24:48):
So what you're saying is maybe I have the bad
experience of having been ghosted or being left in limbo.
So in order to save you my thirty minute, my
thirty minute stack daity from that painful fate, I should
to full up with you right away and say ps,
I'm not interested whatsoever goodect with your life.
Speaker 2 (25:06):
It's like overcompensating one o one. It's like, I'm not
even giving you time to think what you thought of me.
I'm just going to over communicate here and speed dump you.
There was a story in the Wall Street Journal, and
it was interesting because some of the people interviewed felt
like speed dumping was a great antidote to ghosting. Others
felt it was kind of like this competitive race to
(25:27):
tell the other person you're not interested, I think, to
avoid the feeling of ever being dumped.
Speaker 1 (25:32):
You can't get rejected if you do it first.
Speaker 2 (25:34):
That's exactly right.
Speaker 1 (25:36):
Well, the wheel comes full circle on my headlines. You know,
I feel a little bit seen, at least by myself.
Speaker 2 (25:43):
Whenever you say you feel seen before an article, I'm like,
oh God.
Speaker 1 (25:47):
I feel seen because I've been attracted both to the
toilet situation in the in the Tesla cyber cafe and
also to a story in the Wall Street Journal about
public bathrooms.
Speaker 2 (25:59):
Everybody has their public toilet resource. I'm more of a
Barnes and Noble girl myself.
Speaker 1 (26:04):
I love to read well funny. There was this line
in the story that really got me, which is more
Americans could soon have a place that answered nature's call.
Without first buying a drink at Tobucks or a book
or a book. But I read this story in the
journal about a tech company called Throne Labs.
Speaker 2 (26:21):
Brilliant name, Yeah, it is a good name.
Speaker 1 (26:23):
They're trying to revolutionize public bathrooms. You may have already
seen self cleaning bathrooms, but they're about to get a
lot smarter. Your toilet is about to get to know you.
Speaker 2 (26:33):
Does someone go, oh, there's a toilet, That toilet needs
to be smarter.
Speaker 1 (26:36):
Well, I think there is a real problem right, which
is public restrooms are not available in the US. The
US is tied with Botswana in thirtieth place for the
lowest number of public restrooms per capita, and Throne has
realized that part of the problem is public restrooms get
destroyed by the public.
Speaker 2 (26:55):
And we are going too deep today, right.
Speaker 1 (27:00):
Part of the solution is, of course, a rating system,
think about Uber. But for a toilet, if you don't
take care of the throne, if you don't keep that
throne polished, you're not gonna be able to sit on
it again. So thrown bathrooms are free. But if you
don't have a phone, create an account. You can actually
get a key card for entry but in all cases
it's linked to you as a user, and when you
(27:22):
enter the bathroom you're expected to rate this cleannliness. There
are smoke detectors, no smoking occupancy sensors, one at a time,
and sessions limited to ten minutes, so no doom scrolling,
which is the best thing to do on the absolutely
So far, there are over one hundred thrones throughout the US,
and the company's working on adding additional features like a
(27:44):
smell sensor. The smell sensor is to alert the care
and maintenance teams they need to come and pay a visit.
I like that line about Starbucks and not having to
buy a drink just to go to the toilet. But
I also like the way the article summed it up
quote the brains behind Throwne start by getting real about
why Americans usually can't have nice things. They assume a
(28:05):
cultural inability to protect and maintain shared assets and design
their system with software and just enough internet connected sensors
to monitor facilities without violating our expectation of privacy.
Speaker 2 (28:23):
And now it's time for chatting me. This week, I
have a story from me. This is my story.
Speaker 1 (28:30):
Yes, Kara, you go.
Speaker 2 (28:32):
So this week I ignored the ye old hypochondria staple WebMD,
and my friend and I use chat GPT on the
couch to help diagnose myself.
Speaker 1 (28:44):
Well, I'm obviously sorry to hear that you're sick. What
your symptoms and how did you decide to eschew WebMD
in favor of the future.
Speaker 2 (28:55):
Well, I always make a promise to myself not to
use WebMD, because that's what hypochondriacs do, and I don't
want to self identify. I don't want to be in
the club that would have me as a member.
Speaker 1 (29:04):
Chat was your fig leaf fear.
Speaker 2 (29:05):
That's correct. However, two weeks ago, I started feeling extremely tired,
and I was lamenting to a lot of my friends
that I was sleeping like a college freshman all the
all the time. And I'll be honest with you today,
very hard for me to get out of bed. Interesting,
And last Thursday I got out of bed, and you know,
I looked at myself in the mirror, as one does,
(29:27):
and I saw this really nasty rash on my leg.
But I thought, you know what, Kara, don't get upset
about this. It'll go away.
Speaker 1 (29:37):
So you didn't, You didn't we amd nothing?
Speaker 2 (29:39):
No, No, I was. I was really trying to be good.
And you know they always say, give it four days.
If it gets worse, do something about it. So I
gave it four days. It got much worse, and people
had opinions about what it was. People were like, it's
contact dermatitis, it's poison ivy people people.
Speaker 1 (29:56):
You mean friends, friends?
Speaker 2 (29:57):
Ye, well, so even a nurse practitioner told me that
I had contact dermatitis, which is an important thing to note.
My friend was like, you're being a moron. Use chatty bet.
Speaker 1 (30:08):
More, and you should go to the doctor. Use chat.
Speaker 2 (30:13):
She was like, you're bring an idiot. This is so easy.
So she took a photo of this rash and uploaded
I don't use chattybt in this way. She uses chattubt
in this.
Speaker 1 (30:21):
Way, and you don't use it for images normal text yet.
Speaker 2 (30:23):
So she uploaded this photo and just like that, the
doctor was in. She started looking at me sort of
with a funny face, and I was like, I really
feel like I'm in a doctor's office right now.
Speaker 1 (30:36):
How's that bedside mana.
Speaker 2 (30:38):
Terrible because she was very she was wrapped up in
chatgybt and she goes, when did this start? Do you
have any symptoms? And i'd say, you know, I'm extremely tired,
and she's asking me this because chattybt is prompting her.
She asked me about four questions that got more and
more detailed. By the end of it. I remembered that
I had gotten a tick bite a few weeks ago,
(31:01):
which I just I literally forgot about because I scraped
it right off my leg.
Speaker 1 (31:05):
It was known for any length of time. Wow.
Speaker 2 (31:07):
No, And so she gives me this look when we
get to the end of her questions and she says,
that's limes disease. And I was like, no, it's not.
It's not the traditional bulls eye. And she's like, well,
CHATCHBT says it's limes disease. It's limes disease. I go
to the doctor the next day. I was planning on
going to the doctor because the rash had gotten worse.
(31:27):
I have my CHATGYBT diagnosis in hand. I go to
the doctor and before she says anything to me, I say,
CHATCHBT says, I have limes disease.
Speaker 1 (31:35):
You seriously Wow.
Speaker 2 (31:37):
She looks at the thing on my leg and she goes,
I'm glad you came into the doctor's office. This is
limes disease, and what I thought was so interesting and
why in large language models are so interesting in comparison
to friends who have only seen limes disease maybe never
is that she said, you know, you don't have the
traditional bullseye that indicates limes disease. But because I'm a dermatologist,
(32:02):
I know that this rash is in a pattern that
denotes lime disease. So this is just amazing to me
because I basically was diagnosed by chatgebt before the dermatologist.
It didn't keep me from going to the dermatologist, but
I was able to get this answer that nobody else
had given me that was ultimately right before I went
(32:24):
to the doctor.
Speaker 1 (32:25):
I live in fear of getting lime disease?
Speaker 2 (32:27):
Do you really do well? As you can see today?
Speaker 1 (32:30):
You see, I mean you see you seem fine.
Speaker 2 (32:32):
Yeah, I'm okay, I'm okay.
Speaker 1 (32:34):
Tie it.
Speaker 2 (32:34):
It's just exhausting.
Speaker 1 (32:36):
So you have to take iv antibiotics. What do you do?
Speaker 2 (32:38):
Uh No, I'm taking doxy cycline for twenty one days,
which okay. So the interesting part about that, here's what AI,
Here's what chatgybt can't do. CHATBT yet cannot write me
a prescription. Chat GPT doesn't tell me it likes the
show that I work on. It doesn't compliment me, but
(32:58):
it can my Dermatali. This is a bit more ssycophantic
than chatgebt this time. But I mean, I just think
it's an interesting moment because there is what I feel
to be a definitiveness about chatjebt that I never get
looking on WebMD. Once I uploaded the picture of my rash,
my friend walked me through CHATJEPT and I had results.
(33:20):
I felt comforted in a weird way that I probably
shouldn't feel comforted by, like this is an LM basically
spitting out an answer that I don't trust for my friends,
which was just it made me.
Speaker 1 (33:32):
Think, Yeah, we actually have coming up as a guest
in the next few weeks one of the chief research
scientists at Microsoft who worked on a paper that went
viral about how AI is outperforming doctors at diagnosis in
certain fields. So I'm looking forward to that conversation. I
guess the thing which comes to mind for me is
this is arounything, narrow and ultimately not the most serious
(33:54):
in most cases. No, it's nice cheez and it's not
so debatable. There can you imagine if you were talking
to chat about treatment options for a chronic and maybe
fatal disease and your doctor are on one side, you
get different information from Chat. Like you can see how
(34:16):
this is like very efficient in this context. But also
we don't about WebMD and you know hypochondria, this could drive.
Speaker 2 (34:25):
Steroidal I mean we've talked about chat GBT psychosis. I
would imagine people have brought in print doubts from WebMD
and Google. This is that on steroids. And I would
like to talk maybe just to like a general practitioner
interesting who would know firsthand, like how much people are
coming in and saying, well, I talked to chat GBT
(34:45):
about this and they said something different.
Speaker 1 (34:47):
In fact, I want you doctor to talk directly to
chat and I'm going to listen.
Speaker 2 (34:50):
Go ahead.
Speaker 1 (34:52):
Well, I love hearing the story, Kara, but listeners, we
want to hear yours. Please share the peculiar or useful
ways you're using Chat, grock, Cil, Gemini or any chatboard
and send us a one to two minute voice note
to tech Stuff podcast at gmail dot com.
Speaker 2 (35:28):
That's it for this week for Tech Stuff.
Speaker 1 (35:30):
I'm Cara Price, I'm Ozva Loosin this episode, was produced
by Eliza Dennis. It was executive produced by me Caro
Price and Kate Osborne for Kaleidoscope and Katrina Norvel for
iHeart Podcasts. The engineer is Piheid. Fraser jack Insley mixed
this episode and Kyle Murdoch wrote out theme song.
Speaker 2 (35:47):
Join us next Wednesday for text Uff the Story, when
we will examine the lives of kidfluencers and their families.
Speaker 1 (35:54):
Please rate, review, and reach out to us at tech
Stuff podcast at gmail dot com. We want to hear
from you.
TechStuff News
Hosts And Creators
Oz Woloshyn
Karah Preiss
Popular Podcasts
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.
The Clay Travis and Buck Sexton Show
The Clay Travis and Buck Sexton Show. Clay Travis and Buck Sexton tackle the biggest stories in news, politics and current events with intelligence and humor. From the border crisis, to the madness of cancel culture and far-left missteps, Clay and Buck guide listeners through the latest headlines and hot topics with fun and entertaining conversations and opinions.