April 25, 2025 • 37 mins
Should you delete yourself from the internet? This week in the News Roundup, Oz and Karah dig into humanoid robots running a half-marathon, the AI-generated personas helping law enforcement interact with potential suspects and Google’s updated ‘Results About You’ tool. On TechSupport, Jeff Rosenthal, the co-founder of the venture capital firm CIV, discusses the role of private investment in building out energy infrastructure to meet the AI boom.
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):
Welcome to Tech Stuff, a production of iHeart Podcasts and Kaleidoscope.
I'm mas Valoshan and today Karra Price and I will
bring you the headlines this week, including the race of
the humanoid robots. Then on tech Support, we'll talk to
Jeff Rosenthal about the role of private investment in building
out energy infrastructure in the US to meet the AI boom.
Speaker 2 (00:34):
I would argue that this is one of the spaces
that you just see the future manifest regularly, like the
impossible become passible.
Speaker 1 (00:42):
All of that. On the weekend Tech It's Friday, April
twenty fifth. Hello, Hello Cara.
Speaker 3 (00:51):
Hi as well.
Speaker 1 (00:53):
Since we last saw each other, A big important holiday happened,
which was Easter. And did you see how it celebrated
at the White.
Speaker 4 (01:00):
I did, And I can't say that I know many
people who still roll eggs around with wooden spoons, but
the White House likes its traditions.
Speaker 1 (01:09):
Yeah. There was also the human sized bunny who ended
up next to President Trump on the podium, who.
Speaker 4 (01:14):
Looked a lot like Peter Rabbit, but suspiciously with a
different colored jacket.
Speaker 1 (01:19):
This was, in fact Peter the Rabbit, who's been subtly
renamed and restyled face exactly. He has not to infringe
on Beatrix Potter's copyright. This wasn't the only slightly weird
sight to behold at the celebration. There were also sponsorships
and branding from the big tech companies. Meta had its
(01:40):
own tent on the White House lawn, complete with an
AI powered photo op. YouTube had a bunny hop stage,
and a reading nook was provided by Amazon, complete with
a couch and some colorful flowers with the Amazon logo
in pride of place. This was all on the lawn
of the White House at an event traditionally sponsored by
American egg Board.
Speaker 4 (02:01):
I should have done this for my butt, mitzpah. Actually,
just have different tech companies cover different tables.
Speaker 1 (02:07):
Well, future bamans for readers of the Arch of the
Deal will no doubt be able to defray the costs
in years to come.
Speaker 4 (02:12):
Absolutely, you know, nothing says easter to mean quite like
a pink and blue sign with the words expand your
World with Meta AI on it, which is just so
funny to me because Meta execs past and present are
literally testifying right down the street as this is going on.
Speaker 1 (02:26):
I know, I mean, it's surely incredible. What's what's going
on Washington right now? On the one hand, you have
Easter at the White House with all of these corporate
sponsors from the tech industry, and on the other hand,
you have the government's lawyers arguing to break up Google
Meta on trial. It's this, you know, rather delicious seeming
irony that you can't buy friendship, at least not from
(02:46):
President Trump.
Speaker 4 (02:47):
But you can buy a couch at an Easter party exactly,
You know what I mean? You know, it does make
me wonder where this trend will go. I can sort
of imagine a huge tent around Thanksgiving saying the annual
pardoning of the Thanksgiving Turkey sponsored by TikTok.
Speaker 1 (03:04):
Speaking of strange and slightly jarring tech encroachment, I've got
a headline to start us off. The Wall Street Journal
reports that a recent half marathon race in Beijing had
some unusual participants humanoid robots. This thirteen mile race featured
thousands of humans and twenty one robots running. There's a
(03:24):
chance for China to show off its progress in humanoid robotics.
You can check out the video online. It's pretty deep
in the Young Canny Valley to see robots jogging alongside humans.
Speaker 4 (03:34):
This is my favorite kind of story, I think because
of the uncanny valleyness of it, and also because there
was It wasn't perfect, let's just put it that way.
Speaker 1 (03:42):
We'll get to some of the winners and losers.
Speaker 4 (03:44):
We'll talk about that. But you know, China is really
leaning into humanoid robotics. They've said they want to be
the leader by twenty twenty seven, which I don't know
about you, but that's close to when I turned forty,
not far away.
Speaker 1 (03:58):
As you mentioned, the race did show this still quite
a lot room for improvement when it comes to these humanoids.
They had to have a staggered start to the race,
literally so they wouldn't run into each other. There were
pauses between the robots starting. They were also in their
own lane, the HOV lane of the half marathon in Beijing.
Speaker 4 (04:14):
There's so much about this that I love. But my
favorite part is that some of these robots were wearing accessories.
You know, some had like running shoes, tank tops, you know,
they have those little like bicycle hat and to me,
they all kind of look like guys in Fort Green
who are trying to like bag women on hinge.
Speaker 1 (04:29):
Those weren't the only upgrades to the robots. They went
beyond low ups In some cases, the developers had to
modify the robots for the race so that plastic components
wouldn't break off while they were running, and actually recast
and replace those parts with metal. They also given longer legs,
something I would love to.
Speaker 4 (04:45):
Have, which you know, to me, this is a sign
that humans are still superior. These robots trained for months
so they could navigate both the flat and hilly parts
of the course and also take six left turns and
eight right turns, something fortunately I still don't have to
practice for.
Speaker 1 (05:01):
And I think the majority of these robots were actually
remote controlled. But a big issue was stamina. That doesn't
mean they need to take water, break or catch their
breath like we do, but these robots are breat on
batteries which need to be recharged every two hours, and
the race was what over three hours long, so at
different points in the race, the robots were stopped because
they ran out of power and had to swap batteries
in order to keep going, which.
Speaker 4 (05:21):
Is just like a pit stop at a NASCAR race.
Let me just clarify when you say a group of
humanoid robots. Listeners might think this is a robot army,
you know, it's like very Star Wars, where they all
look the same and run at an identical pace. But
that was not what happened here. It was a melee.
Speaker 1 (05:37):
Have you seen Have you seen when dog walkers are
in Central Park and you have every type of dog
from big to small, from crazy to calm.
Speaker 3 (05:45):
That's how we lose people.
Speaker 1 (05:46):
That this was that of robots. It was a melee
of different shape, different size, different herky jerky robots. I
think in some sense, you know, there was a competition
amongst the robot manufacturers. Some fared better than others on
one and you have the front runner, Chian Kung Ultra,
who stood at five foot nine and only fell over
once when it's battery failed. It did change his battery
(06:08):
three times, but Tien ran at about six miles an
hour across different terrains including hills, stairs, grass and sand.
Others weren't so lucky. One humanoid with propellers went off course,
slammed into a fence and broke into pieces. It sounds comedic,
it was actually kind of scary. Things. Two hundred pound
metallic humanoid charge its way into a crowd of people,
(06:30):
how to tell you, some were barely able to walk,
much less run. One humanoid robot named Juan Juan actually
went in the wrong direction for a little bit before
sitting down and refusing to go further, which was like
me at sports day as a child. Yes, in you two.
Speaker 3 (06:45):
I think I was going to say we are all.
Speaker 1 (06:46):
One on this podcast where we are one.
Speaker 3 (06:49):
Sure, it's very relatable.
Speaker 4 (06:51):
And while Tian Kung Ultra was the first humanoid to
cross the finish line after about two hours and forty minutes,
a human runner finished the race over an hour and
a half before that, So only two humanoids cross the
finish line within the race's original cutoff time, which cut
extended to four hours and ten minutes because most of
them just simply weren't going to make it. If I
(07:11):
was running a marathon, this would also happen.
Speaker 1 (07:13):
By the way, it's funny to laugh at the missteps
of these robots. But remember how when we were first
playing around with Gennety of Ai. You know, it was
impressive that it could write anything at all, but the
writing was completely unserviceable. So, you know, I think that
we're seeing what we may be seeing here as a
kind of leading edge of in fact and explosion in
(07:33):
humanoid robots because these robots are intertwined deeply with the
development of AI. Robots have different parts, but you could
say that the brain is the AI which allows the
robots to learn from patterns and mimic human behavior in
real time. And according to analysts at Goldman Sachs, Deep
(07:55):
seeks are one model may have changed the game for
Chinese robotic companies because now top level AI performance is
possible with fewer advanced ships and less computing power, which
makes it easier to distribute across a society.
Speaker 4 (08:08):
So is it safe to say that this race was
less between humans and robots, which it clearly was not,
and more between the US and China.
Speaker 1 (08:16):
Well, Cara, that's why they pay the big bugs. I agree.
I agree. There is a strong national pride element to
the Chinese robotics sector and seemingly a desire to integrate
robots into more and more human activities such as this marathon.
There's also some interesting robot human dance self videos that
you can find online.
Speaker 3 (08:34):
I've seen them.
Speaker 1 (08:35):
They're well worth checking out. Moving on to another story
that could be straight out of a sci fi novel,
but unfortunately is not in partnership with Wired magazine. Four
or four Media reported on a tool that some police
departments near the Southern border are paying quite a bit
of money to use. The product is called Overwatch, as
(08:55):
in the shoot them Up video game, naturally, and it
comes from a company called Massive Blue.
Speaker 3 (09:01):
Only boys can name this.
Speaker 4 (09:02):
I'm sorry, I'm not saying women can't, but it sounds
like something that only boys can name.
Speaker 1 (09:06):
I'm with you on this, not unfortunately. Overwatch is marketed
as a product that quote deploys lifelike virtual agents which
infiltrate and engage criminal networks across various channels. Basically, Massive
Blue is offering cops virtual personas that can be unleashed
on the Internet and used to interact with quote college protesters, quote,
(09:27):
radicalized political activists, and suspected drug and human traffickers over
text messages and social media.
Speaker 4 (09:34):
And one such example of an Overwatch AI persona is Heidi,
and Massive Blue described her in a presentation to the
Texas Department of Public Safety as quote a radicalized AI persona.
Here's her backstory? You ready, yeah, ready, Oh my god,
it's me. She's a thirty six year old, Oh my god, childless,
(09:54):
divorced woman who lives in Texas. But this is when
she got radicalized. She was raised in San Francisco.
Speaker 1 (10:01):
Oh my god.
Speaker 4 (10:02):
Her hobbies are activism and baking, and her personality is
quote outspoken, lonely, and body positive. This is the greatest
thing I've ever heard in my entire life.
Speaker 1 (10:13):
This is summoned from the nightmares of JD events.
Speaker 4 (10:17):
Literally, like the these are the threats to American society.
A body positive child woman.
Speaker 1 (10:22):
Body positive, lonely, outspits.
Speaker 4 (10:24):
By the way, she's going to get five thousand if
Heidi keeps up with this body positive, childless behavior. The
Trump administration might give her five thousand dollars to have
a child.
Speaker 1 (10:34):
So maybe the unintended consequence of a legion of Heidi's
roaming the internet will be bankrupting the American American government
taking advantage of the credits to have children exactly. This
is obviously a truly weird story, but the purpose of
personas like Heidi is to interact with and of course,
perhaps we might say in trap real suspects on social media.
(10:59):
The idea is by looking to these potential suspects, Heidi
and other AI personas can gather evidence on them to
be used by police departments. Sorry, I shouldn't be lovey.
I mean it's bizarre.
Speaker 4 (11:10):
It's the most bizarre thing. Also that like these are
the threats to American civilization. Like to your point, like
an AI cat, we don't even know if she has
a cat, because her cat could be a bigger threat.
But that she's outspoken and body positive is what's so
crazy to me. I do need to point out that
it is legal to protest in the United States, so
it's a bit odd that one of the types of
(11:32):
AI personas is a college protester. But the idea of
law enforcement using fake personas to catch suspects isn't new.
Officers have made profiles claiming to be teenagers to gather
information on child predators.
Speaker 1 (11:46):
Now, this is like if you turned to catch your
predator into an AI and then turned that AI into
a product and then sold it to police departments.
Speaker 3 (11:54):
That's exactly that's exactly right.
Speaker 1 (11:57):
Using this software instead of having to log on and
make a fake profile message of suspect back and forth
and hope they don't catch on to what you're doing,
Overwatch does a lot of that work for you so.
Another of watched profile the police departments can use is
child trafficking AI persona, also known as Jason. Jason is
fourteen years old, likes video games, and has a hard
(12:17):
time interacting with girls.
Speaker 3 (12:19):
He's an insult or just a normal.
Speaker 1 (12:20):
Fourteen year old boy. To be fair to him, I
wasn't such a big video gamer, but I certainly had
those other two characteristics. In the presentation obtained by four
or for Media, there is a screenshot of a sample
text exchange between Jason and someone who is presumably a
predatory adult, asking these alone, Jason replies to the message, quote,
just chillin by myself. Man. My mom's at at Symbol
(12:43):
work and my dad's out of town, so it's just
me and my VID games video game controller emoji.
Speaker 4 (12:50):
I would have already been under a dictionary. I wouldn't
even be able to talk to this. You know, we're
not clear if this is a real exchange. We do
know that Massive Blues signed a three hundred and six
d thousand dollars contract with Panal County in Arizona, and
the county use an anti human trafficking grant to pay
for this exchange. The Panal County Sheriff's office did tell
(13:10):
four A four media that Overwatch has so far not
been used for any arrests.
Speaker 3 (13:14):
It is used for a lot of laughs.
Speaker 1 (13:17):
Yeah, I'm curious if the fact it hasn't been used
for any arrests it's considered to be a good thing
or a bad thing.
Speaker 3 (13:22):
That's what I was going to say. Yeah, exactly.
Speaker 1 (13:24):
So, Karen, while we're on the topic of what kind
of information can we gathered about people online, You've got
a headline for us about removing yourself from the Internet.
Speaker 3 (13:33):
Yeah.
Speaker 4 (13:33):
So the Wall Street Journal ran an article with the headline,
go delete yourself from the Internet, and it talks about
how Google recently updated its results about You feature and
you can plug in details like your name, various email addresses,
phone numbers, or street addresses, and Google will literally show
you what kind of things about you or out there
on the internet.
Speaker 1 (13:52):
Yeah. I looked at this and I was like, him
is basically asking me for all of my personal data
to run a search on what personal data exists on
the Internet. So, of course I I did it anyway,
And indeed, I think this is this is less about
a kind of cursory Google search and more of them
doing like a deep web search to reveal what exists
about you online? Is that true?
Speaker 4 (14:10):
Yes, it's things like where you live, if you've ever
gotten a speeding ticket, magazine subscriptions, which is very risky
for me. And these little bits of data can add
up to a pretty substantial amount of information about you
that's just floating around. You know, it can expose you
to annoyances like getting more junk mail or scarier situations
like identity theft or doxing or receiving unwonted flowers.
Speaker 1 (14:32):
Has that ever happened to you? I can't talk about
it and not to mention. Data brokers scrape this stuff,
package it and then sell it on as a dossier.
We're talking about very personal stuff here, like license plate
numbers of your vehicles, list of your family members, and
the dossiers are getting more and more detailed as time
goes on and more data becomes available. Yes.
Speaker 4 (14:51):
So just to be clear, like Google's results about you
only provides information, there's a separate request process for removing
things from search results, and there are are also other
services you can pay for to proactively remove personal data
from the Internet.
Speaker 1 (15:05):
One thing I can't stop thinking about is that deleting
yourself from the Internet is not something you can do
once and then it's done forever. I mean, it's like
a never ending task.
Speaker 4 (15:13):
It is a never ending story, and it's ongoing for
a couple of reasons. First, you know, if you don't
live in a state with data privacy laws like California
that require, for example, people search sites to take down
your data upon request, that information might just stay up there.
And second, data that's been taken down can reappear. And
I think this is an unfortunate part of being in
(15:35):
the modern world, is that like your data can just
be out there.
Speaker 1 (15:39):
I mean, I feel bad for people who live in
states where mugshot databases automatically uploaded to the Internet. I mean,
I find that so deeply unfair that you're kind of
haunted by your mugshot forever. Even if you know Sheriff's
department takes it down subsequently, it will never never leave
the Internet. Also, it's one of those things where we know,
you know, I know we have to be much more
(15:59):
careful with day to privacy, but it's just so hard
to actually do it. I mean, it's one of those
things where it seems I until something bad happens to you,
it's almost impossible to motivate yourself to take more care.
Speaker 4 (16:11):
And also, you and I always talk about this, like
what is to be gained by giving away my data
quite a bit? What is lost by giving away my
data quite a bit? But I don't care as much.
Speaker 1 (16:21):
Yeah, I think there are things you can do at
the margins, like make a fake email address or a
burner email address for every time you sign up for
a site online, have a different password every time you know.
Those are steps you can take that will be substantially protective.
It's time for a quick break now. When we come back,
we'll run through the short headlines and speak with Jeff Rosenthal,
(16:43):
co founder of the VC firm SIEV about investing in
energy infrastructure. Carrow it back, and it's time to get
into the shorties.
Speaker 3 (17:06):
I like that.
Speaker 1 (17:07):
I'll start with one of my favorite publications, The Aviationist.
Speaker 3 (17:13):
Do you really read that?
Speaker 1 (17:14):
You read it online? I read it online. They have
a story with the headline British Army radio wave weapon
cooks multiple drones swarms simultaneously. If you can write headlines
like that, you deserve to be anybody's favorite publication.
Speaker 3 (17:29):
Peabody, Babe Peabody.
Speaker 1 (17:31):
So the British Army is testing a new air defense
weapon that can simultaneously destroy more than one hundred drones
in one deployment. The so called rapid Destroyer we Brits
to God with names, damages or disrupts critical electronic components
inside drones using just radio waves at high frequencies. This
(17:51):
causes the drones to crash or malfunction during flight, and
it could actually be a transformative defensive weapon for conflicts
like the War in Ukraine. To the UK's Defense Intelligence,
Ukraine had to defend against attacks from more than eighteen
thousand drones in twenty twenty four.
Speaker 4 (18:07):
So bad news for members of jen Alpha trying to
pose his adults on Instagram. The Verge reports that Meta
is taking its AI driven age detection to the next level.
If a user claims to be an adult, meaning they
lie about their birth year, but then an AI tool
finds evidence that they are in fact a teenager, evidence
like messages from friends saying happy sixteenth birthday, smoking that's
(18:29):
a big smoking gun, Meta will automatically switch that user
to a teen account, So watch out who your friends are.
Speaker 3 (18:35):
These accounts are more restrictive.
Speaker 4 (18:37):
For example, they're private by default and Instagram limits the
kind of content teen c.
Speaker 1 (18:42):
Here's one for all the people still ignoring cryptocurrency. The
Wall Street Journal has a story under the headline Crypto
knocks on the door of a banking world, shut it out,
and it reports that some crypto firms plan to apply
for bank charters or licenses, allowing them to operate more
like traditional lenders that can take deposits and make loans.
This comes just a few years after the FTX collapse,
(19:04):
when many major banks cut ties with crypto firms. But
where a bank license comes striccher regulation. At least for now,
all eyes are on the crypto friendly Trump administration to
see where the crypto companies will have to play by
the same rules in order to get these licenses.
Speaker 4 (19:20):
And lastly, I wanted to highlight a new startup with
an unfathomable mission to achieve quote the full automation of
all work, so to basically achieve humans having to do
no work. This story is from tech Crunch, and the
company called Mechanize intends to research the limitations of AI
because they believe quote the explosive economic growth likely to
(19:43):
result from completely automating labor could generate vast abundance, much
higher standards of living, and new goods and services that
we can't even imagine today. Unsurprisingly, the announcement has received
considerable disdain and backlash on social media, but the company,
in its affiliate nonprofit AI research organization, actually has significant
backing from industry insiders, and they want you to know
(20:07):
that they are hiring humans.
Speaker 1 (20:09):
Not fully mechanized yet not yet. It is interesting, though,
how quickly things which seem like pipe dreams of the
tech industry become the bread and butter of our daily lives.
I mean, think no further than generative AI chat GPT.
Speaker 3 (20:25):
It's all I can hear about.
Speaker 4 (20:26):
It's all people focus on, and even people who don't
know I host a tech podcast are talking to me
about it. I'm hearing stories at book club, at my passoversator,
in all of my group chats. Everyone wants to tell
me ways in which they are using generative AI.
Speaker 1 (20:41):
Have there been any particularly striking usages that kind of
stuck with you?
Speaker 4 (20:46):
I actually was able to do a palm reading on
my mother. I'm sorry, mom for giving chat GPT your palm.
We're gonna have to look into that on Google.
Speaker 1 (20:54):
Chatchept will say yes to a request to do a
palm reading.
Speaker 4 (20:57):
Oh absolutely, and then it gives you all of your
different life lines and your bumps on your hand and
all of that. And then there's the lowest hanging fruit,
which is what do I text back to the guy
from Hinge, which seems to be the most popular use.
Speaker 1 (21:09):
Yeah. I love the idea of AI palmistry. I mean,
all of these industries that would have thought when the
writers were striking in Hollywood two years ago, this will
never happen to me.
Speaker 3 (21:19):
Of all of it, I.
Speaker 1 (21:19):
Should have laughed. But literally, really, a palm with AI
is insane.
Speaker 4 (21:23):
We're going to see all those storefronts in the West
Village and Greenwich Village just shutting down because chatgvt has
taken their jobs.
Speaker 1 (21:30):
Just computer screens and palm readers. Yeah. Something I think
back too often is my conversation with azeemas are Back
in March, he writes the Exponential View newsletter, and he
talks about the adoption of AI. Back in March, he'd
just given a presentation at south By Southwest about the
intersection between AI and energy demand, specifically how much energy
(21:50):
is needed to power all this new usage, and he
says something that really stuck with me, which is that
USAI dominance could be challenged by how quickly this country
can bring new power onto the grid.
Speaker 4 (22:02):
Yeah, and the tech companies are addressing this head on.
Microsoft made a deal with the owners of the Three
Mile Island nuclear plant, which suffered a meltdown in the
nineteen seventies, to reopen it, and in return, Microsoft committed
to paying almost double market energy rates over a twenty
year contract.
Speaker 1 (22:19):
So obviously energy is absolutely critical to the future of
the tech industry, and as we know US private money
increasingly plays a huge role in tech buildout. And just
look at Stargate, which are both fascinated by.
Speaker 4 (22:31):
I don't know if i'd say fave, but I'm definitely
interested in Stargate.
Speaker 1 (22:35):
This week, per Bloomberg quote, a new venture capital firm
called SIV has raised an inaugural fund of two hundred
million dollars to invest in startups tackling projects like nuclear
energy and manufacturing, joining a broader movement in Silicon Valley
to back physical world companies with national implications. Hits tell
us more is Jeff Rosenthal, the co founder and managing
(22:58):
partner of SIV. Jeff, thanks so much for being here today.
Speaker 2 (23:01):
Thanks so much for having me Oz. It's great to
be here.
Speaker 1 (23:03):
So tell us a little bit about SIV.
Speaker 4 (23:04):
So.
Speaker 2 (23:04):
SIV is a firm built with a very singular purpose
in mind. It's to back and build businesses that are
focusing on critical infrastructure and industry domestically. Here in the US.
There are these three multi decade megatrends that are really
meeting us quickly in this moment, and those are AI
and compute and the power needs behind them in the
digital infrastructure required. As you mentioned, the rev limitter to
(23:28):
AI really isn't GPUs, it is electrons. It's the reshoring
and reindustrialization of global industry after decades of offshoring and
globalization really following COVID and Ukraine and now accelerated with
the new administration. And then the third is the electrification
of everything and really just the need for energy abundance domestically,
not only because of economic prosperity, but also as a
(23:51):
pertinent national security requirement. So we see that as a
potential seventy trillion dollar market opportunity over the next fifteen years.
To put that in perspective, that would be like building
the rail system and the highway system every six weeks
for fifteen years here in the United States.
Speaker 1 (24:07):
And what's the name SIEV means SIV is short for civilization,
So big, big goals.
Speaker 2 (24:12):
I mean, you know, shoot for the stars, land on
the moon.
Speaker 1 (24:15):
I'm especially interested in energy, and we've spoken about Microsoft's
investments in nuclear power. Do you at SIV have a
kind of ten year view on how America's energy needs
are going to be addressed?
Speaker 2 (24:27):
Absolutely? Yeah. So we're saying we want to triple energy consumption,
you know, in a decade. Right. It's it's just the
United States. It's like a lot of these things that
people say. It's like, hey, if we're going to meet
the needs of the time, if we're going to meet
the demand curve of AI and reshoring and electrification, we
need to triple energy. It just doesn't really work that
way because we're talking about infrastructure. This is steel, it's rebar,
(24:48):
it's concrete, it's tens of thousands of people on work
sites every single day. And so it's not AI, it's
not software. You can't just say, like, we have a
million more users, let's just scale it. You know, when
we looked at this and we saw inflect demand with
relatively fixed supply. We thought that there was really nothing
that could completely cross that chasm except maybe the thing
that we invented seventy years ago, which is a scaled vision.
(25:10):
So you see plenty of money and venture going into
new reactors. And so when we look at the United States,
we have ninety four operational reactors today, we're the leading
nuclear superpower in the world. However, there's about fifty different
designs active, which means that they're all special snowflakes. And
when you build anything in of one as a singular project,
that's the most expensive way you could ever do things.
(25:31):
And so what we innovated on at the Nuclear Company,
which we co founded on our platform actually before raising
outside capital, is really a fleet scale approach to developing
and deploying approved nuclear.
Speaker 1 (25:41):
Reactors, a fleet of reactors.
Speaker 2 (25:43):
Yes, so instead of building one approved reactor gigawatt scale reactor,
we're actually building in of six. And the idea is
that you have huge cost declines and huge time declines
from reactor one to two to three to five. You
can build a real supply chain, you can build real
trades that roll from reactor to reactor, and this is
all through observed data.
Speaker 1 (26:04):
Power generation is one challenge, power storage is another. Onboarding
to the grid is another, Dealing with regulations is another.
What's the kind of total picture of the ways in
which you are addressing the opportunity of energy demand in
the US.
Speaker 2 (26:23):
Yeah, I think it's a very broad set of needs.
To your point, it's energy generation, it's storage, it's dealing
with more and more volatility on the grid which increases cost.
It's dealing with, as you mentioned in the opening, the
veritable arms race that the hyperscalers are going through for power.
This is a existential crisis for them to secure their
(26:44):
future growth. There is industrial efficiency technology. There's everything in
the vertical of digital infrastructure, cooling, high capacity lines, you know,
next generation racks and servers, innovation and acceleration in GPU,
and this is all Chapter one. That's the most exciting
part is like this is the worst that this technology
will ever be. And this is the very beginning of
(27:06):
the innovation cycle for all of these things. When you
use chat GPT's the worst it will ever be. The
way we really model what we personally back or build
as we say, hey, is this scale tech versus deep tech.
We think ninety five percent of the solutions are here today.
The chasm to cross is more one of execution than
one of pure intervention.
Speaker 1 (27:25):
Just just to clarify, scale tech means something that already
exists that you can accelerate by investing in, whereas deep
tech means essentially replicating the role of university or funding
for research that may only bear commercial fruits a decade
from now.
Speaker 2 (27:40):
That's right. I think it's too long duration to capital
intensive for a vehicle like ours, and we're also not PhDs,
so we really don't have much to offer you if
you're still in a lab, so we won't really take
binary science or technology risk. We then look for proven
customer demand, hence our LP base of true operators in
the spaces in which we invest. We want to understand
the appetite for adoption, the imminence of that adoption, when
(28:04):
would you adopt it, and then we want to back
into competitive unit economics. Businesses with those features really scale exponentially.
They can really redefine markets.
Speaker 4 (28:15):
After the break, OZ has more from Jeff Rosenthal, a
VC firmative that's on tech Stuff's tech support.
Speaker 3 (28:21):
Stay with us.
Speaker 1 (28:35):
On the nuclear side. This is kind of one of
the great debates I think going on in the tech
world will be how much of text amount of energy
is met by nuclear. Of course, one of the facts
in this is that people are scared of nuclear power.
I mean they're scared of how you suppose that the
wast they're scared of meltdowns. Did you have any concerns
about the first project in this space? You know, you're
(28:56):
putting your name to being nuclear.
Speaker 2 (28:58):
I think it gives us solid that zero people have
died from nuclear in the United States. You know, like
hundreds of fallen off of windmills. If you think about
coal plants or all the other ways in which we
consume power, there's trade offs to everything. I personally feel
that splitting atoms to create chain reactions to power giant
turbines that give us the energy and prosperity for our
(29:19):
cities in our society is almost the most amazing thing
we've ever done as a species. And we're not handling
nuclear waste ourselves. It's not like I'm taking it out
in a bag in the back of my house. We
work with nuclear operating utilities, and so this is constellation
or floor to power in light. They handle the nuclear
waste today without issue. We do believe that this is
an essential part of the power mix of the future.
(29:40):
It won't be all of it, you know, Like today,
the most efficient way to really run a data center
on power is two gas turbines because that achieves a
capacity factor, meaning they can stay on close to one
hundred percent of the time. The problem with intermittent energy
powering data centers is you really can't do training or.
Speaker 1 (29:57):
Inference intimatum being basically solar.
Speaker 2 (30:01):
Wind, et cetera. Yeah, so the value of twenty four
x seven clean base power is exponential for the entire system.
It's really really important. So it's not all just apples
to apples.
Speaker 1 (30:12):
There is I mean, there are cool storage solutions around,
you know, capturing wind and solar energy to make it
on nine twenty four seven or investments in.
Speaker 2 (30:20):
That space explicitly true. There are emergent technologies for large
long duration battery storage for wind and solar. There's things
like form energy and in Tora energy, there's interesting startups
that are looking to store thing and it's store power
and heat. So these things are amazing, they're so inspiring,
but they are not really here today.
Speaker 1 (30:41):
We talked a lot on this show about how some
of these things that seemed like pipe dreams or seemed
like science fiction, uh, you know, becoming science fact in
our lifetimes.
Speaker 2 (30:49):
These stories have been all around us for a long time.
It's more an awareness issue than it is like an
invention issue. So people are like, where's my iPhone? You know,
I got my first iPhone? Change my life, life Gmail
and change my life. Where are these things? And so
I would argue that this is one of the spaces
that you just see the future manifest regularly, like the
impossible become possible. It's such a fun innovation cycle.
Speaker 1 (31:13):
You know.
Speaker 2 (31:13):
There's things like Cigar Lake. Cigar Lake is in eastern Saskatchewan, Canada.
It's owned by Camico. They discovered it in the early eighties.
It's underneath a snow melt lake fourteen hundred feet deep.
It's a billion three year old deposit of uranium. Most uranium,
say in Africa point one two point two percent uranium
per ounce of war, right, and so these massive sulphuric
(31:35):
acid leaching fields. This is like twenty four percent uranium
war It's the second most concentrated uranium deposit in the world.
So they spent decades figuring out how to freeze the lake.
They free on this entire lake. Okay, then they rebarred it,
and then they drilled it fourteen hundred feet down and
they've been mining it with robots for three decades. And
it's where something like a quarter of our uranium comes
(31:56):
from in the United States. And so all I'm saying
the reason I offer you this, this is like a
wonder of the world. There are one hundred of these.
We just don't really know about them because we haven't
been excited about this space for a long time.
Speaker 1 (32:07):
Bloomberg used the phrase physical world companies with national implications.
I thought was a good phrase because the world you're
investing in and building, of course, intersects very heavily with politics.
But on clean energy, it's kind of interesting because President
Trump has talked about beautiful coal, beautiful, clean coal, beautiful beautiful,
(32:28):
and has certainly stepped away from the Biden administration's commitment
to like keleen energy and green energy. At the same time,
I think Q one of twenty twenty five was the
biggest bonanza quarter for investment in climate tech the US
has ever seen. Five million dollars invested in one quarter
up nearly sixty five percent year on year. What's going
on there is private capitals seeing an opportunity in public
(32:51):
drawback or how do you know?
Speaker 2 (32:52):
I mean, this is the Golden Age. So first and foremost,
there's this thing called Geben's paradox, and with Jebens paradox,
exis ammons is that for you know, one hundred and
fifty years, humans have essentially gotten one percent more efficient
every single year. We invent LEDs where we figure out
a more efficient electric motor, and then we use that
gain an efficiency which drops pricing to increase consumption by
(33:15):
one point three percent. So every time we increase efficiency gains,
we actually use more of what we have.
Speaker 1 (33:21):
And this was why when deep Sea came out, the
Nvidia stock price went down, a lot of small people said, well,
actually this might be a good thing ultimately because although
they've done it without the very very advanced chips, this
will ultimately drive more demand for chips one.
Speaker 2 (33:34):
Percent and more demand for AI. I believe we're pretty
energy blind in our society to a degree. In the
case of these clean energy companies and these efficiency technologies.
For the twenty years leading up to this we were
in an abundance market. You just did not need these things.
There was no market driver, there was no demand driver
for innovation in these spaces, and so that you can
see the returns being pretty abysmal.
Speaker 1 (33:56):
Historical energy demand was flat until Jerry, we.
Speaker 2 (33:59):
Were solving problem that didn't exist yet. And so it's
not just AI that's a major contributor. It's AI. It's
the reshoring and reindustrialization of global industry. If you want
to smelt steel in America, it takes a lot of
power building flush rocks in America if you want to
I mean, let alone, refinement of critical minerals and then
the mining of critical minerals like these are all things
that are nonpartisan in my opinion. Like this is a
(34:21):
movement that's been happening for quite some time. And we see,
you know, the tariffs more as a canary in the
coal mine.
Speaker 1 (34:26):
So the energy sector has been particularly punished by the
swings and roundabouts on the tariffs. I think abound sixteen
percent this year. That's not something that concerns you as
somebody investing in this space, I.
Speaker 2 (34:38):
Think it certainly concerns me. I think that you know,
you talk to anybody in retail markets are softening, consumers
are spending less money. But I think that these trends
that we're talking about are multi decade mega trends that
you know, you're going to need these things if you
want to have a prosperous society in the future.
Speaker 1 (34:58):
When people think about the sa they tend to think
about consumer like you know, will by Parker or Goop. Well,
I think about SaaS, I think about you know, Slack
or Shopify or whatever else. But you're not the only
person to be looking addressing this kind of infrastructure opportunity
through VC. Andrewson, I think, announced a six and roe
(35:18):
million dollar fun last year to address American dynamism.
Speaker 2 (35:22):
What's striving this necessity is the mother of innovation first
and foremost. You know, you have these mega trens we've identified,
I think many other people have. I think that there's
also a more durable moat in physical products in a
sense now, so while you know, hardware the first four letters,
and hardware is hard and so Benuura is always like, yeah,
you know, I'll stick to enterprise SaaS. That is way
(35:44):
less defensible than it used to be. You know, now
with windsurf and your own time, no staff, you can
code and develop, you know, your own apps or platforms
or technologies. Suites that used to take millions of dollars
in years and dozens of people on front end back
in nearing development. None of that is necessary anymore.
Speaker 1 (36:03):
So.
Speaker 2 (36:03):
The defensibility of what was really like the center of
the Venn diagram for venture deals a decade ago has
kind of disappeared to a degree. I think people are
conscious of that, and so we'd rather work with companies
that are utilizing the tools of the day to really
reimagine and reinvent critical industries versus just being like yet
another digital service provider.
Speaker 1 (36:25):
Jeff, thank you for joining us on tech Stuff.
Speaker 2 (36:27):
Of course, thank you for having me.
Speaker 3 (36:32):
That's it for this week for tech Stuff.
Speaker 1 (36:34):
I'm Karra Price and I'm as Valoshin. This episode was
produced by Eliza Dennis, Victoria Dominguez and Adriana Tapia. Is
Executive produced by me Cara Price and Kate Osborne for
Kaleidoscope and Katrina Norvel for iHeart Podcasts. The engineers are
Bahiit Fraser and Graham Gibson. Kyle Murdock makes this episode
and he also rid of theme.
Speaker 4 (36:54):
Song Join us next Wednesday for tex Stuff the Story,
when we will share an in depth conversation with journalist
and podcast host Evan Ratliffe about a radical group of
young tech people concerned by the existential threat of AI,
who are on trial for murder.
Speaker 1 (37:09):
Please rate, review, and reach out to us at tech
Stuff podcast at gmail dot com. We really want to
hear from you.
Welcome to Tech Stuff, a production of iHeart Podcasts and Kaleidoscope.
I'm mas Valoshan and today Karra Price and I will
bring you the headlines this week, including the race of
the humanoid robots. Then on tech Support, we'll talk to
Jeff Rosenthal about the role of private investment in building
out energy infrastructure in the US to meet the AI boom.
Speaker 2 (00:34):
I would argue that this is one of the spaces
that you just see the future manifest regularly, like the
impossible become passible.
Speaker 1 (00:42):
All of that. On the weekend Tech It's Friday, April
twenty fifth. Hello, Hello Cara.
Speaker 3 (00:51):
Hi as well.
Speaker 1 (00:53):
Since we last saw each other, A big important holiday happened,
which was Easter. And did you see how it celebrated
at the White.
Speaker 4 (01:00):
I did, And I can't say that I know many
people who still roll eggs around with wooden spoons, but
the White House likes its traditions.
Speaker 1 (01:09):
Yeah. There was also the human sized bunny who ended
up next to President Trump on the podium, who.
Speaker 4 (01:14):
Looked a lot like Peter Rabbit, but suspiciously with a
different colored jacket.
Speaker 1 (01:19):
This was, in fact Peter the Rabbit, who's been subtly
renamed and restyled face exactly. He has not to infringe
on Beatrix Potter's copyright. This wasn't the only slightly weird
sight to behold at the celebration. There were also sponsorships
and branding from the big tech companies. Meta had its
(01:40):
own tent on the White House lawn, complete with an
AI powered photo op. YouTube had a bunny hop stage,
and a reading nook was provided by Amazon, complete with
a couch and some colorful flowers with the Amazon logo
in pride of place. This was all on the lawn
of the White House at an event traditionally sponsored by
American egg Board.
Speaker 4 (02:01):
I should have done this for my butt, mitzpah. Actually,
just have different tech companies cover different tables.
Speaker 1 (02:07):
Well, future bamans for readers of the Arch of the
Deal will no doubt be able to defray the costs
in years to come.
Speaker 4 (02:12):
Absolutely, you know, nothing says easter to mean quite like
a pink and blue sign with the words expand your
World with Meta AI on it, which is just so
funny to me because Meta execs past and present are
literally testifying right down the street as this is going on.
Speaker 1 (02:26):
I know, I mean, it's surely incredible. What's what's going
on Washington right now? On the one hand, you have
Easter at the White House with all of these corporate
sponsors from the tech industry, and on the other hand,
you have the government's lawyers arguing to break up Google
Meta on trial. It's this, you know, rather delicious seeming
irony that you can't buy friendship, at least not from
(02:46):
President Trump.
Speaker 4 (02:47):
But you can buy a couch at an Easter party exactly,
You know what I mean? You know, it does make
me wonder where this trend will go. I can sort
of imagine a huge tent around Thanksgiving saying the annual
pardoning of the Thanksgiving Turkey sponsored by TikTok.
Speaker 1 (03:04):
Speaking of strange and slightly jarring tech encroachment, I've got
a headline to start us off. The Wall Street Journal
reports that a recent half marathon race in Beijing had
some unusual participants humanoid robots. This thirteen mile race featured
thousands of humans and twenty one robots running. There's a
(03:24):
chance for China to show off its progress in humanoid robotics.
You can check out the video online. It's pretty deep
in the Young Canny Valley to see robots jogging alongside humans.
Speaker 4 (03:34):
This is my favorite kind of story, I think because
of the uncanny valleyness of it, and also because there
was It wasn't perfect, let's just put it that way.
Speaker 1 (03:42):
We'll get to some of the winners and losers.
Speaker 4 (03:44):
We'll talk about that. But you know, China is really
leaning into humanoid robotics. They've said they want to be
the leader by twenty twenty seven, which I don't know
about you, but that's close to when I turned forty,
not far away.
Speaker 1 (03:58):
As you mentioned, the race did show this still quite
a lot room for improvement when it comes to these humanoids.
They had to have a staggered start to the race,
literally so they wouldn't run into each other. There were
pauses between the robots starting. They were also in their
own lane, the HOV lane of the half marathon in Beijing.
Speaker 4 (04:14):
There's so much about this that I love. But my
favorite part is that some of these robots were wearing accessories.
You know, some had like running shoes, tank tops, you know,
they have those little like bicycle hat and to me,
they all kind of look like guys in Fort Green
who are trying to like bag women on hinge.
Speaker 1 (04:29):
Those weren't the only upgrades to the robots. They went
beyond low ups In some cases, the developers had to
modify the robots for the race so that plastic components
wouldn't break off while they were running, and actually recast
and replace those parts with metal. They also given longer legs,
something I would love to.
Speaker 4 (04:45):
Have, which you know, to me, this is a sign
that humans are still superior. These robots trained for months
so they could navigate both the flat and hilly parts
of the course and also take six left turns and
eight right turns, something fortunately I still don't have to
practice for.
Speaker 1 (05:01):
And I think the majority of these robots were actually
remote controlled. But a big issue was stamina. That doesn't
mean they need to take water, break or catch their
breath like we do, but these robots are breat on
batteries which need to be recharged every two hours, and
the race was what over three hours long, so at
different points in the race, the robots were stopped because
they ran out of power and had to swap batteries
in order to keep going, which.
Speaker 4 (05:21):
Is just like a pit stop at a NASCAR race.
Let me just clarify when you say a group of
humanoid robots. Listeners might think this is a robot army,
you know, it's like very Star Wars, where they all
look the same and run at an identical pace. But
that was not what happened here. It was a melee.
Speaker 1 (05:37):
Have you seen Have you seen when dog walkers are
in Central Park and you have every type of dog
from big to small, from crazy to calm.
Speaker 3 (05:45):
That's how we lose people.
Speaker 1 (05:46):
That this was that of robots. It was a melee
of different shape, different size, different herky jerky robots. I
think in some sense, you know, there was a competition
amongst the robot manufacturers. Some fared better than others on
one and you have the front runner, Chian Kung Ultra,
who stood at five foot nine and only fell over
once when it's battery failed. It did change his battery
(06:08):
three times, but Tien ran at about six miles an
hour across different terrains including hills, stairs, grass and sand.
Others weren't so lucky. One humanoid with propellers went off course,
slammed into a fence and broke into pieces. It sounds comedic,
it was actually kind of scary. Things. Two hundred pound
metallic humanoid charge its way into a crowd of people,
(06:30):
how to tell you, some were barely able to walk,
much less run. One humanoid robot named Juan Juan actually
went in the wrong direction for a little bit before
sitting down and refusing to go further, which was like
me at sports day as a child. Yes, in you two.
Speaker 3 (06:45):
I think I was going to say we are all.
Speaker 1 (06:46):
One on this podcast where we are one.
Speaker 3 (06:49):
Sure, it's very relatable.
Speaker 4 (06:51):
And while Tian Kung Ultra was the first humanoid to
cross the finish line after about two hours and forty minutes,
a human runner finished the race over an hour and
a half before that, So only two humanoids cross the
finish line within the race's original cutoff time, which cut
extended to four hours and ten minutes because most of
them just simply weren't going to make it. If I
(07:11):
was running a marathon, this would also happen.
Speaker 1 (07:13):
By the way, it's funny to laugh at the missteps
of these robots. But remember how when we were first
playing around with Gennety of Ai. You know, it was
impressive that it could write anything at all, but the
writing was completely unserviceable. So, you know, I think that
we're seeing what we may be seeing here as a
kind of leading edge of in fact and explosion in
(07:33):
humanoid robots because these robots are intertwined deeply with the
development of AI. Robots have different parts, but you could
say that the brain is the AI which allows the
robots to learn from patterns and mimic human behavior in
real time. And according to analysts at Goldman Sachs, Deep
(07:55):
seeks are one model may have changed the game for
Chinese robotic companies because now top level AI performance is
possible with fewer advanced ships and less computing power, which
makes it easier to distribute across a society.
Speaker 4 (08:08):
So is it safe to say that this race was
less between humans and robots, which it clearly was not,
and more between the US and China.
Speaker 1 (08:16):
Well, Cara, that's why they pay the big bugs. I agree.
I agree. There is a strong national pride element to
the Chinese robotics sector and seemingly a desire to integrate
robots into more and more human activities such as this marathon.
There's also some interesting robot human dance self videos that
you can find online.
Speaker 3 (08:34):
I've seen them.
Speaker 1 (08:35):
They're well worth checking out. Moving on to another story
that could be straight out of a sci fi novel,
but unfortunately is not in partnership with Wired magazine. Four
or four Media reported on a tool that some police
departments near the Southern border are paying quite a bit
of money to use. The product is called Overwatch, as
(08:55):
in the shoot them Up video game, naturally, and it
comes from a company called Massive Blue.
Speaker 3 (09:01):
Only boys can name this.
Speaker 4 (09:02):
I'm sorry, I'm not saying women can't, but it sounds
like something that only boys can name.
Speaker 1 (09:06):
I'm with you on this, not unfortunately. Overwatch is marketed
as a product that quote deploys lifelike virtual agents which
infiltrate and engage criminal networks across various channels. Basically, Massive
Blue is offering cops virtual personas that can be unleashed
on the Internet and used to interact with quote college protesters, quote,
(09:27):
radicalized political activists, and suspected drug and human traffickers over
text messages and social media.
Speaker 4 (09:34):
And one such example of an Overwatch AI persona is Heidi,
and Massive Blue described her in a presentation to the
Texas Department of Public Safety as quote a radicalized AI persona.
Here's her backstory? You ready, yeah, ready, Oh my god,
it's me. She's a thirty six year old, Oh my god, childless,
(09:54):
divorced woman who lives in Texas. But this is when
she got radicalized. She was raised in San Francisco.
Speaker 1 (10:01):
Oh my god.
Speaker 4 (10:02):
Her hobbies are activism and baking, and her personality is
quote outspoken, lonely, and body positive. This is the greatest
thing I've ever heard in my entire life.
Speaker 1 (10:13):
This is summoned from the nightmares of JD events.
Speaker 4 (10:17):
Literally, like the these are the threats to American society.
A body positive child woman.
Speaker 1 (10:22):
Body positive, lonely, outspits.
Speaker 4 (10:24):
By the way, she's going to get five thousand if
Heidi keeps up with this body positive, childless behavior. The
Trump administration might give her five thousand dollars to have
a child.
Speaker 1 (10:34):
So maybe the unintended consequence of a legion of Heidi's
roaming the internet will be bankrupting the American American government
taking advantage of the credits to have children exactly. This
is obviously a truly weird story, but the purpose of
personas like Heidi is to interact with and of course,
perhaps we might say in trap real suspects on social media.
(10:59):
The idea is by looking to these potential suspects, Heidi
and other AI personas can gather evidence on them to
be used by police departments. Sorry, I shouldn't be lovey.
I mean it's bizarre.
Speaker 4 (11:10):
It's the most bizarre thing. Also that like these are
the threats to American civilization. Like to your point, like
an AI cat, we don't even know if she has
a cat, because her cat could be a bigger threat.
But that she's outspoken and body positive is what's so
crazy to me. I do need to point out that
it is legal to protest in the United States, so
it's a bit odd that one of the types of
(11:32):
AI personas is a college protester. But the idea of
law enforcement using fake personas to catch suspects isn't new.
Officers have made profiles claiming to be teenagers to gather
information on child predators.
Speaker 1 (11:46):
Now, this is like if you turned to catch your
predator into an AI and then turned that AI into
a product and then sold it to police departments.
Speaker 3 (11:54):
That's exactly that's exactly right.
Speaker 1 (11:57):
Using this software instead of having to log on and
make a fake profile message of suspect back and forth
and hope they don't catch on to what you're doing,
Overwatch does a lot of that work for you so.
Another of watched profile the police departments can use is
child trafficking AI persona, also known as Jason. Jason is
fourteen years old, likes video games, and has a hard
(12:17):
time interacting with girls.
Speaker 3 (12:19):
He's an insult or just a normal.
Speaker 1 (12:20):
Fourteen year old boy. To be fair to him, I
wasn't such a big video gamer, but I certainly had
those other two characteristics. In the presentation obtained by four
or for Media, there is a screenshot of a sample
text exchange between Jason and someone who is presumably a
predatory adult, asking these alone, Jason replies to the message, quote,
just chillin by myself. Man. My mom's at at Symbol
(12:43):
work and my dad's out of town, so it's just
me and my VID games video game controller emoji.
Speaker 4 (12:50):
I would have already been under a dictionary. I wouldn't
even be able to talk to this. You know, we're
not clear if this is a real exchange. We do
know that Massive Blues signed a three hundred and six
d thousand dollars contract with Panal County in Arizona, and
the county use an anti human trafficking grant to pay
for this exchange. The Panal County Sheriff's office did tell
(13:10):
four A four media that Overwatch has so far not
been used for any arrests.
Speaker 3 (13:14):
It is used for a lot of laughs.
Speaker 1 (13:17):
Yeah, I'm curious if the fact it hasn't been used
for any arrests it's considered to be a good thing
or a bad thing.
Speaker 3 (13:22):
That's what I was going to say. Yeah, exactly.
Speaker 1 (13:24):
So, Karen, while we're on the topic of what kind
of information can we gathered about people online, You've got
a headline for us about removing yourself from the Internet.
Speaker 3 (13:33):
Yeah.
Speaker 4 (13:33):
So the Wall Street Journal ran an article with the headline,
go delete yourself from the Internet, and it talks about
how Google recently updated its results about You feature and
you can plug in details like your name, various email addresses,
phone numbers, or street addresses, and Google will literally show
you what kind of things about you or out there
on the internet.
Speaker 1 (13:52):
Yeah. I looked at this and I was like, him
is basically asking me for all of my personal data
to run a search on what personal data exists on
the Internet. So, of course I I did it anyway,
And indeed, I think this is this is less about
a kind of cursory Google search and more of them
doing like a deep web search to reveal what exists
about you online? Is that true?
Speaker 4 (14:10):
Yes, it's things like where you live, if you've ever
gotten a speeding ticket, magazine subscriptions, which is very risky
for me. And these little bits of data can add
up to a pretty substantial amount of information about you
that's just floating around. You know, it can expose you
to annoyances like getting more junk mail or scarier situations
like identity theft or doxing or receiving unwonted flowers.
Speaker 1 (14:32):
Has that ever happened to you? I can't talk about
it and not to mention. Data brokers scrape this stuff,
package it and then sell it on as a dossier.
We're talking about very personal stuff here, like license plate
numbers of your vehicles, list of your family members, and
the dossiers are getting more and more detailed as time
goes on and more data becomes available. Yes.
Speaker 4 (14:51):
So just to be clear, like Google's results about you
only provides information, there's a separate request process for removing
things from search results, and there are are also other
services you can pay for to proactively remove personal data
from the Internet.
Speaker 1 (15:05):
One thing I can't stop thinking about is that deleting
yourself from the Internet is not something you can do
once and then it's done forever. I mean, it's like
a never ending task.
Speaker 4 (15:13):
It is a never ending story, and it's ongoing for
a couple of reasons. First, you know, if you don't
live in a state with data privacy laws like California
that require, for example, people search sites to take down
your data upon request, that information might just stay up there.
And second, data that's been taken down can reappear. And
I think this is an unfortunate part of being in
(15:35):
the modern world, is that like your data can just
be out there.
Speaker 1 (15:39):
I mean, I feel bad for people who live in
states where mugshot databases automatically uploaded to the Internet. I mean,
I find that so deeply unfair that you're kind of
haunted by your mugshot forever. Even if you know Sheriff's
department takes it down subsequently, it will never never leave
the Internet. Also, it's one of those things where we know,
you know, I know we have to be much more
(15:59):
careful with day to privacy, but it's just so hard
to actually do it. I mean, it's one of those
things where it seems I until something bad happens to you,
it's almost impossible to motivate yourself to take more care.
Speaker 4 (16:11):
And also, you and I always talk about this, like
what is to be gained by giving away my data
quite a bit? What is lost by giving away my
data quite a bit? But I don't care as much.
Speaker 1 (16:21):
Yeah, I think there are things you can do at
the margins, like make a fake email address or a
burner email address for every time you sign up for
a site online, have a different password every time you know.
Those are steps you can take that will be substantially protective.
It's time for a quick break now. When we come back,
we'll run through the short headlines and speak with Jeff Rosenthal,
(16:43):
co founder of the VC firm SIEV about investing in
energy infrastructure. Carrow it back, and it's time to get
into the shorties.
Speaker 3 (17:06):
I like that.
Speaker 1 (17:07):
I'll start with one of my favorite publications, The Aviationist.
Speaker 3 (17:13):
Do you really read that?
Speaker 1 (17:14):
You read it online? I read it online. They have
a story with the headline British Army radio wave weapon
cooks multiple drones swarms simultaneously. If you can write headlines
like that, you deserve to be anybody's favorite publication.
Speaker 3 (17:29):
Peabody, Babe Peabody.
Speaker 1 (17:31):
So the British Army is testing a new air defense
weapon that can simultaneously destroy more than one hundred drones
in one deployment. The so called rapid Destroyer we Brits
to God with names, damages or disrupts critical electronic components
inside drones using just radio waves at high frequencies. This
(17:51):
causes the drones to crash or malfunction during flight, and
it could actually be a transformative defensive weapon for conflicts
like the War in Ukraine. To the UK's Defense Intelligence,
Ukraine had to defend against attacks from more than eighteen
thousand drones in twenty twenty four.
Speaker 4 (18:07):
So bad news for members of jen Alpha trying to
pose his adults on Instagram. The Verge reports that Meta
is taking its AI driven age detection to the next level.
If a user claims to be an adult, meaning they
lie about their birth year, but then an AI tool
finds evidence that they are in fact a teenager, evidence
like messages from friends saying happy sixteenth birthday, smoking that's
(18:29):
a big smoking gun, Meta will automatically switch that user
to a teen account, So watch out who your friends are.
Speaker 3 (18:35):
These accounts are more restrictive.
Speaker 4 (18:37):
For example, they're private by default and Instagram limits the
kind of content teen c.
Speaker 1 (18:42):
Here's one for all the people still ignoring cryptocurrency. The
Wall Street Journal has a story under the headline Crypto
knocks on the door of a banking world, shut it out,
and it reports that some crypto firms plan to apply
for bank charters or licenses, allowing them to operate more
like traditional lenders that can take deposits and make loans.
This comes just a few years after the FTX collapse,
(19:04):
when many major banks cut ties with crypto firms. But
where a bank license comes striccher regulation. At least for now,
all eyes are on the crypto friendly Trump administration to
see where the crypto companies will have to play by
the same rules in order to get these licenses.
Speaker 4 (19:20):
And lastly, I wanted to highlight a new startup with
an unfathomable mission to achieve quote the full automation of
all work, so to basically achieve humans having to do
no work. This story is from tech Crunch, and the
company called Mechanize intends to research the limitations of AI
because they believe quote the explosive economic growth likely to
(19:43):
result from completely automating labor could generate vast abundance, much
higher standards of living, and new goods and services that
we can't even imagine today. Unsurprisingly, the announcement has received
considerable disdain and backlash on social media, but the company,
in its affiliate nonprofit AI research organization, actually has significant
backing from industry insiders, and they want you to know
(20:07):
that they are hiring humans.
Speaker 1 (20:09):
Not fully mechanized yet not yet. It is interesting, though,
how quickly things which seem like pipe dreams of the
tech industry become the bread and butter of our daily lives.
I mean, think no further than generative AI chat GPT.
Speaker 3 (20:25):
It's all I can hear about.
Speaker 4 (20:26):
It's all people focus on, and even people who don't
know I host a tech podcast are talking to me
about it. I'm hearing stories at book club, at my passoversator,
in all of my group chats. Everyone wants to tell
me ways in which they are using generative AI.
Speaker 1 (20:41):
Have there been any particularly striking usages that kind of
stuck with you?
Speaker 4 (20:46):
I actually was able to do a palm reading on
my mother. I'm sorry, mom for giving chat GPT your palm.
We're gonna have to look into that on Google.
Speaker 1 (20:54):
Chatchept will say yes to a request to do a
palm reading.
Speaker 4 (20:57):
Oh absolutely, and then it gives you all of your
different life lines and your bumps on your hand and
all of that. And then there's the lowest hanging fruit,
which is what do I text back to the guy
from Hinge, which seems to be the most popular use.
Speaker 1 (21:09):
Yeah. I love the idea of AI palmistry. I mean,
all of these industries that would have thought when the
writers were striking in Hollywood two years ago, this will
never happen to me.
Speaker 3 (21:19):
Of all of it, I.
Speaker 1 (21:19):
Should have laughed. But literally, really, a palm with AI
is insane.
Speaker 4 (21:23):
We're going to see all those storefronts in the West
Village and Greenwich Village just shutting down because chatgvt has
taken their jobs.
Speaker 1 (21:30):
Just computer screens and palm readers. Yeah. Something I think
back too often is my conversation with azeemas are Back
in March, he writes the Exponential View newsletter, and he
talks about the adoption of AI. Back in March, he'd
just given a presentation at south By Southwest about the
intersection between AI and energy demand, specifically how much energy
(21:50):
is needed to power all this new usage, and he
says something that really stuck with me, which is that
USAI dominance could be challenged by how quickly this country
can bring new power onto the grid.
Speaker 4 (22:02):
Yeah, and the tech companies are addressing this head on.
Microsoft made a deal with the owners of the Three
Mile Island nuclear plant, which suffered a meltdown in the
nineteen seventies, to reopen it, and in return, Microsoft committed
to paying almost double market energy rates over a twenty
year contract.
Speaker 1 (22:19):
So obviously energy is absolutely critical to the future of
the tech industry, and as we know US private money
increasingly plays a huge role in tech buildout. And just
look at Stargate, which are both fascinated by.
Speaker 4 (22:31):
I don't know if i'd say fave, but I'm definitely
interested in Stargate.
Speaker 1 (22:35):
This week, per Bloomberg quote, a new venture capital firm
called SIV has raised an inaugural fund of two hundred
million dollars to invest in startups tackling projects like nuclear
energy and manufacturing, joining a broader movement in Silicon Valley
to back physical world companies with national implications. Hits tell
us more is Jeff Rosenthal, the co founder and managing
(22:58):
partner of SIV. Jeff, thanks so much for being here today.
Speaker 2 (23:01):
Thanks so much for having me Oz. It's great to
be here.
Speaker 1 (23:03):
So tell us a little bit about SIV.
Speaker 4 (23:04):
So.
Speaker 2 (23:04):
SIV is a firm built with a very singular purpose
in mind. It's to back and build businesses that are
focusing on critical infrastructure and industry domestically. Here in the US.
There are these three multi decade megatrends that are really
meeting us quickly in this moment, and those are AI
and compute and the power needs behind them in the
digital infrastructure required. As you mentioned, the rev limitter to
(23:28):
AI really isn't GPUs, it is electrons. It's the reshoring
and reindustrialization of global industry after decades of offshoring and
globalization really following COVID and Ukraine and now accelerated with
the new administration. And then the third is the electrification
of everything and really just the need for energy abundance domestically,
not only because of economic prosperity, but also as a
(23:51):
pertinent national security requirement. So we see that as a
potential seventy trillion dollar market opportunity over the next fifteen years.
To put that in perspective, that would be like building
the rail system and the highway system every six weeks
for fifteen years here in the United States.
Speaker 1 (24:07):
And what's the name SIEV means SIV is short for civilization,
So big, big goals.
Speaker 2 (24:12):
I mean, you know, shoot for the stars, land on
the moon.
Speaker 1 (24:15):
I'm especially interested in energy, and we've spoken about Microsoft's
investments in nuclear power. Do you at SIV have a
kind of ten year view on how America's energy needs
are going to be addressed?
Speaker 2 (24:27):
Absolutely? Yeah. So we're saying we want to triple energy consumption,
you know, in a decade. Right. It's it's just the
United States. It's like a lot of these things that
people say. It's like, hey, if we're going to meet
the needs of the time, if we're going to meet
the demand curve of AI and reshoring and electrification, we
need to triple energy. It just doesn't really work that
way because we're talking about infrastructure. This is steel, it's rebar,
(24:48):
it's concrete, it's tens of thousands of people on work
sites every single day. And so it's not AI, it's
not software. You can't just say, like, we have a
million more users, let's just scale it. You know, when
we looked at this and we saw inflect demand with
relatively fixed supply. We thought that there was really nothing
that could completely cross that chasm except maybe the thing
that we invented seventy years ago, which is a scaled vision.
(25:10):
So you see plenty of money and venture going into
new reactors. And so when we look at the United States,
we have ninety four operational reactors today, we're the leading
nuclear superpower in the world. However, there's about fifty different
designs active, which means that they're all special snowflakes. And
when you build anything in of one as a singular project,
that's the most expensive way you could ever do things.
(25:31):
And so what we innovated on at the Nuclear Company,
which we co founded on our platform actually before raising
outside capital, is really a fleet scale approach to developing
and deploying approved nuclear.
Speaker 1 (25:41):
Reactors, a fleet of reactors.
Speaker 2 (25:43):
Yes, so instead of building one approved reactor gigawatt scale reactor,
we're actually building in of six. And the idea is
that you have huge cost declines and huge time declines
from reactor one to two to three to five. You
can build a real supply chain, you can build real
trades that roll from reactor to reactor, and this is
all through observed data.
Speaker 1 (26:04):
Power generation is one challenge, power storage is another. Onboarding
to the grid is another, Dealing with regulations is another.
What's the kind of total picture of the ways in
which you are addressing the opportunity of energy demand in
the US.
Speaker 2 (26:23):
Yeah, I think it's a very broad set of needs.
To your point, it's energy generation, it's storage, it's dealing
with more and more volatility on the grid which increases cost.
It's dealing with, as you mentioned in the opening, the
veritable arms race that the hyperscalers are going through for power.
This is a existential crisis for them to secure their
(26:44):
future growth. There is industrial efficiency technology. There's everything in
the vertical of digital infrastructure, cooling, high capacity lines, you know,
next generation racks and servers, innovation and acceleration in GPU,
and this is all Chapter one. That's the most exciting
part is like this is the worst that this technology
will ever be. And this is the very beginning of
(27:06):
the innovation cycle for all of these things. When you
use chat GPT's the worst it will ever be. The
way we really model what we personally back or build
as we say, hey, is this scale tech versus deep tech.
We think ninety five percent of the solutions are here today.
The chasm to cross is more one of execution than
one of pure intervention.
Speaker 1 (27:25):
Just just to clarify, scale tech means something that already
exists that you can accelerate by investing in, whereas deep
tech means essentially replicating the role of university or funding
for research that may only bear commercial fruits a decade
from now.
Speaker 2 (27:40):
That's right. I think it's too long duration to capital
intensive for a vehicle like ours, and we're also not PhDs,
so we really don't have much to offer you if
you're still in a lab, so we won't really take
binary science or technology risk. We then look for proven
customer demand, hence our LP base of true operators in
the spaces in which we invest. We want to understand
the appetite for adoption, the imminence of that adoption, when
(28:04):
would you adopt it, and then we want to back
into competitive unit economics. Businesses with those features really scale exponentially.
They can really redefine markets.
Speaker 4 (28:15):
After the break, OZ has more from Jeff Rosenthal, a
VC firmative that's on tech Stuff's tech support.
Speaker 3 (28:21):
Stay with us.
Speaker 1 (28:35):
On the nuclear side. This is kind of one of
the great debates I think going on in the tech
world will be how much of text amount of energy
is met by nuclear. Of course, one of the facts
in this is that people are scared of nuclear power.
I mean they're scared of how you suppose that the
wast they're scared of meltdowns. Did you have any concerns
about the first project in this space? You know, you're
(28:56):
putting your name to being nuclear.
Speaker 2 (28:58):
I think it gives us solid that zero people have
died from nuclear in the United States. You know, like
hundreds of fallen off of windmills. If you think about
coal plants or all the other ways in which we
consume power, there's trade offs to everything. I personally feel
that splitting atoms to create chain reactions to power giant
turbines that give us the energy and prosperity for our
(29:19):
cities in our society is almost the most amazing thing
we've ever done as a species. And we're not handling
nuclear waste ourselves. It's not like I'm taking it out
in a bag in the back of my house. We
work with nuclear operating utilities, and so this is constellation
or floor to power in light. They handle the nuclear
waste today without issue. We do believe that this is
an essential part of the power mix of the future.
(29:40):
It won't be all of it, you know, Like today,
the most efficient way to really run a data center
on power is two gas turbines because that achieves a
capacity factor, meaning they can stay on close to one
hundred percent of the time. The problem with intermittent energy
powering data centers is you really can't do training or.
Speaker 1 (29:57):
Inference intimatum being basically solar.
Speaker 2 (30:01):
Wind, et cetera. Yeah, so the value of twenty four
x seven clean base power is exponential for the entire system.
It's really really important. So it's not all just apples
to apples.
Speaker 1 (30:12):
There is I mean, there are cool storage solutions around,
you know, capturing wind and solar energy to make it
on nine twenty four seven or investments in.
Speaker 2 (30:20):
That space explicitly true. There are emergent technologies for large
long duration battery storage for wind and solar. There's things
like form energy and in Tora energy, there's interesting startups
that are looking to store thing and it's store power
and heat. So these things are amazing, they're so inspiring,
but they are not really here today.
Speaker 1 (30:41):
We talked a lot on this show about how some
of these things that seemed like pipe dreams or seemed
like science fiction, uh, you know, becoming science fact in
our lifetimes.
Speaker 2 (30:49):
These stories have been all around us for a long time.
It's more an awareness issue than it is like an
invention issue. So people are like, where's my iPhone? You know,
I got my first iPhone? Change my life, life Gmail
and change my life. Where are these things? And so
I would argue that this is one of the spaces
that you just see the future manifest regularly, like the
impossible become possible. It's such a fun innovation cycle.
Speaker 1 (31:13):
You know.
Speaker 2 (31:13):
There's things like Cigar Lake. Cigar Lake is in eastern Saskatchewan, Canada.
It's owned by Camico. They discovered it in the early eighties.
It's underneath a snow melt lake fourteen hundred feet deep.
It's a billion three year old deposit of uranium. Most uranium,
say in Africa point one two point two percent uranium
per ounce of war, right, and so these massive sulphuric
(31:35):
acid leaching fields. This is like twenty four percent uranium
war It's the second most concentrated uranium deposit in the world.
So they spent decades figuring out how to freeze the lake.
They free on this entire lake. Okay, then they rebarred it,
and then they drilled it fourteen hundred feet down and
they've been mining it with robots for three decades. And
it's where something like a quarter of our uranium comes
(31:56):
from in the United States. And so all I'm saying
the reason I offer you this, this is like a
wonder of the world. There are one hundred of these.
We just don't really know about them because we haven't
been excited about this space for a long time.
Speaker 1 (32:07):
Bloomberg used the phrase physical world companies with national implications.
I thought was a good phrase because the world you're
investing in and building, of course, intersects very heavily with politics.
But on clean energy, it's kind of interesting because President
Trump has talked about beautiful coal, beautiful, clean coal, beautiful beautiful,
(32:28):
and has certainly stepped away from the Biden administration's commitment
to like keleen energy and green energy. At the same time,
I think Q one of twenty twenty five was the
biggest bonanza quarter for investment in climate tech the US
has ever seen. Five million dollars invested in one quarter
up nearly sixty five percent year on year. What's going
on there is private capitals seeing an opportunity in public
(32:51):
drawback or how do you know?
Speaker 2 (32:52):
I mean, this is the Golden Age. So first and foremost,
there's this thing called Geben's paradox, and with Jebens paradox,
exis ammons is that for you know, one hundred and
fifty years, humans have essentially gotten one percent more efficient
every single year. We invent LEDs where we figure out
a more efficient electric motor, and then we use that
gain an efficiency which drops pricing to increase consumption by
(33:15):
one point three percent. So every time we increase efficiency gains,
we actually use more of what we have.
Speaker 1 (33:21):
And this was why when deep Sea came out, the
Nvidia stock price went down, a lot of small people said, well,
actually this might be a good thing ultimately because although
they've done it without the very very advanced chips, this
will ultimately drive more demand for chips one.
Speaker 2 (33:34):
Percent and more demand for AI. I believe we're pretty
energy blind in our society to a degree. In the
case of these clean energy companies and these efficiency technologies.
For the twenty years leading up to this we were
in an abundance market. You just did not need these things.
There was no market driver, there was no demand driver
for innovation in these spaces, and so that you can
see the returns being pretty abysmal.
Speaker 1 (33:56):
Historical energy demand was flat until Jerry, we.
Speaker 2 (33:59):
Were solving problem that didn't exist yet. And so it's
not just AI that's a major contributor. It's AI. It's
the reshoring and reindustrialization of global industry. If you want
to smelt steel in America, it takes a lot of
power building flush rocks in America if you want to
I mean, let alone, refinement of critical minerals and then
the mining of critical minerals like these are all things
that are nonpartisan in my opinion. Like this is a
(34:21):
movement that's been happening for quite some time. And we see,
you know, the tariffs more as a canary in the
coal mine.
Speaker 1 (34:26):
So the energy sector has been particularly punished by the
swings and roundabouts on the tariffs. I think abound sixteen
percent this year. That's not something that concerns you as
somebody investing in this space, I.
Speaker 2 (34:38):
Think it certainly concerns me. I think that you know,
you talk to anybody in retail markets are softening, consumers
are spending less money. But I think that these trends
that we're talking about are multi decade mega trends that
you know, you're going to need these things if you
want to have a prosperous society in the future.
Speaker 1 (34:58):
When people think about the sa they tend to think
about consumer like you know, will by Parker or Goop. Well,
I think about SaaS, I think about you know, Slack
or Shopify or whatever else. But you're not the only
person to be looking addressing this kind of infrastructure opportunity
through VC. Andrewson, I think, announced a six and roe
(35:18):
million dollar fun last year to address American dynamism.
Speaker 2 (35:22):
What's striving this necessity is the mother of innovation first
and foremost. You know, you have these mega trens we've identified,
I think many other people have. I think that there's
also a more durable moat in physical products in a
sense now, so while you know, hardware the first four letters,
and hardware is hard and so Benuura is always like, yeah,
you know, I'll stick to enterprise SaaS. That is way
(35:44):
less defensible than it used to be. You know, now
with windsurf and your own time, no staff, you can
code and develop, you know, your own apps or platforms
or technologies. Suites that used to take millions of dollars
in years and dozens of people on front end back
in nearing development. None of that is necessary anymore.
Speaker 1 (36:03):
So.
Speaker 2 (36:03):
The defensibility of what was really like the center of
the Venn diagram for venture deals a decade ago has
kind of disappeared to a degree. I think people are
conscious of that, and so we'd rather work with companies
that are utilizing the tools of the day to really
reimagine and reinvent critical industries versus just being like yet
another digital service provider.
Speaker 1 (36:25):
Jeff, thank you for joining us on tech Stuff.
Speaker 2 (36:27):
Of course, thank you for having me.
Speaker 3 (36:32):
That's it for this week for tech Stuff.
Speaker 1 (36:34):
I'm Karra Price and I'm as Valoshin. This episode was
produced by Eliza Dennis, Victoria Dominguez and Adriana Tapia. Is
Executive produced by me Cara Price and Kate Osborne for
Kaleidoscope and Katrina Norvel for iHeart Podcasts. The engineers are
Bahiit Fraser and Graham Gibson. Kyle Murdock makes this episode
and he also rid of theme.
Speaker 4 (36:54):
Song Join us next Wednesday for tex Stuff the Story,
when we will share an in depth conversation with journalist
and podcast host Evan Ratliffe about a radical group of
young tech people concerned by the existential threat of AI,
who are on trial for murder.
Speaker 1 (37:09):
Please rate, review, and reach out to us at tech
Stuff podcast at gmail dot com. We really want to
hear from you.
TechStuff News
Hosts And Creators
Oz Woloshyn
Karah Preiss
Popular Podcasts
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com