February 28, 2025 • 38 mins
This week, Oz is on the road, at the Web Summit Qatar, and he’s not alone. Joining him from the iHeart pop-up studio in Doha is a very familiar figure, Jonathan Strickland. Oz and Jonathan sit down to discuss some of the highlights of the Web Summit, including the future of the AI chip race, advances in augmented reality and how news organizations are grappling with AI.
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 (00:00):
Welcome to tech Stuff, a production of iHeart Podcasts and Kaleidoscope.
This week, we're taking you on the road for a
special episode recorded in Doha, Cutter. Now, before you start
imagining sand.
Speaker 2 (00:12):
Dunes and pure blue coastlines, I should describe my surroundings.
Speaker 1 (00:17):
I'm currently sitting in a makeshift studio in a massive
airplane hangar like conference center called the Doha Exhibition and
Convention Center, So if this episode sounds a little bit.
Speaker 2 (00:28):
Different, you'll know why.
Speaker 1 (00:30):
For the last couple of days, I've been attending Web
Summit Cutter, an international conference where investors and entrepreneurs and
thought leaders.
Speaker 2 (00:37):
Gather from around the world to talk about the future
of technology.
Speaker 1 (00:41):
Our friends at iHeart are the official podcast partner's Web Summit,
which means that I'm here and so is a very
familiar figure, Jonathan Strickland.
Speaker 3 (00:51):
Well, hello, I'm so pleased to be a guest on
this historic, amazing podcast. I've heard so much about.
Speaker 1 (00:57):
It, Jonathan, Welcome back to tech Stuff. How does it
feel to be on the other side of the table.
Speaker 3 (01:01):
I'll be honest, it feels odd, but I'm so pleased
to be able to be a part of this. I
was been joking all week about you know, just when
I thought I was out, they pulled me back in.
Speaker 2 (01:15):
Absolutely. Well, I'm very pleased to see you.
Speaker 1 (01:17):
It's been almost seven years, I think since we saw
each other in person, So who knew the next time
would be in Doha?
Speaker 3 (01:22):
And yeah, here I am here we are.
Speaker 1 (01:24):
Based on the listener emails I've been receiving, I think
the listeners are going to be very, very very happy
to hear your voice today.
Speaker 2 (01:30):
You're already missed.
Speaker 3 (01:31):
Well just for you guys, Hey there, welcome to tech stuff.
There you go, what the tech? The tech?
Speaker 1 (01:38):
So nice have you vation? And I hope, I hope
this will be the first of many visits. Absolutely, I'm
not just in Doha. Yes, I would love to do
this on East time. So there've been some fascinating speakers
already at the conference, and I wanted to talk to you,
Jonathan today about some of what I've heard, hoping that
you can contextualize it kind of from a bird's eye
view of having covered technology as a journalist for two
(02:00):
decades sure and attended a bunch of these conferences.
Speaker 3 (02:04):
Yeah, conferences in the tech world in general play an
interesting role. So I've been to lots of different kinds.
I've been to trade shows like CEES or E three,
I've been to ones that have been held by specific
companies such as Intel or IBM and the web Sumit
of course is more of a general tech with Internet
(02:25):
focus on it, and there have been some pretty phenomenal speakers. Actually,
they've really impressed me.
Speaker 1 (02:32):
And when you come to these types of events, like
what do you look out for, both in terms of
kind of getting a peek around the corner to what
may be coming in the world of tech, but also
in terms of applying some kind of critical lens to
what you're hearing.
Speaker 3 (02:45):
That is an excellent question. So let's take CES for example.
Everyone is looking for the quirky thing that is a
standout from the standard stuff you would run into. So
an example would be from a couple of years ago,
the Little Rabbit AI, you know, pocket sized computer device,
which ended up not doing very well when it finally
(03:05):
came out. That's often the case where something that's quirky
gets your attention because it's different, but it turns out
it's not really practical or maybe it doesn't work as
well as was presented. So that's one thing you have
to look out for is yes, it's exciting and that
it's different, but at the same time you have to
ask the question, Okay, does this actually solve a problem.
One of the speakers we had yesterday actually talked about
(03:28):
this about how innovation for innovation's sake is a fool's
errand and that really innovation needs to be applied to
looking at real world problems, not inventing one. So that's
one thing, But I'd say another is you look to
see what people are talking about, what they're excited about,
and how they're talking about it, so that you can
(03:51):
kind of get a feel for what is going to
be the next push, and keep in mind that not
all those pushes are going to be successes. So as
an examp back in two thousand and eight when I
went to my first CEES, the big push back then
was from companies like Panasonic and Sony, where it was
all about three D television. Yeah, now, you and I
know that three D television never really took off. People
(04:14):
balked at the idea of having glasses that they had
to wear at home, and the expense and the lack
of content. All of these things contributed to the failure
of that technology. So when you're being exposed to these
new ideas and people are really excited about them, you
have to temper that a little bit with the reminder
that not everything that's being talked about is going to manifest,
(04:37):
or if it does, it won't manifest in the way
we anticipate. But if you do that and you proceed
with a kind of cautious optimism, I think great things
can be achieved.
Speaker 1 (04:48):
Yeah, with that in mind, I mean, there are a
bunch of kind of smaller booths where people are presenting
on the floor, largely on laptops. I guess it's like
AI applications, but then as a main stage where there's
been some really interesting speakers.
Speaker 2 (05:01):
So I've into three talks so far.
Speaker 1 (05:03):
I kind of want to talk about each of them
briefly to get your view of how they exist in
the context of the history of technology.
Speaker 3 (05:10):
Sure.
Speaker 1 (05:11):
The first one was about the global chip race and
the continuing ways from deep seeg the second was about
the augmented reality presence in future, and the third was
AI and journalism. A fun time, a fun top, especially
for US journalists to come of technology.
Speaker 3 (05:26):
Yeah, especially for those of us who worked for a
company that famously eliminated its editorial board in order to
replace it with AI.
Speaker 2 (05:34):
Wait wait which company?
Speaker 3 (05:35):
How stuff Works? Really? Yeah, back in twenty twenty three.
People have heard me talk about this on the show before,
but if you're new here, I worked for HowStuffWorks dot
com and I was a writer there, and in twenty
twenty three the company decided to stop working with freelance writers.
They had a freelance writing crew, and then they had
an editorial crew in house, and the idea was that
(05:58):
they were going to go with AI general rated articles
from that point forward. The editors would have to do
a full edit pass to make sure that everything was
accurate and correct, and then correct anything that wasn't right,
and as you know, generative AI can sometimes confabulate and
make things up. The editors protested, and then they were eliminated,
and everyone I knew who worked at how Stuff Works
(06:20):
was let go.
Speaker 2 (06:21):
Wow.
Speaker 3 (06:21):
Yeah, So I say this not so people will take
up torches and pitchforks and yell or anything, but rather
to explain that I am very much aware that I
have a very strong bias here, like I cannot be objective,
is what I'm trying to get.
Speaker 1 (06:36):
At I mean, I think if you work as a
journalist and you know you're constantly hearing about the idea
of being replaced by AI, it h it's not the
most appealing seat to sit in.
Speaker 2 (06:46):
This for sure.
Speaker 3 (06:47):
And we've been through similar shifts before, like the pivot
to video being the famous one, right, we saw editorial
departments decimated because the idea was that the written word
was no longer the way to deliver and it was
going to be video only. But it didn't last very long,
and then by that time you had all these journalists
who were out of work. So obviously, as journalists, we
(07:11):
have a sensitivity to these things.
Speaker 1 (07:14):
So we'll come back to that, but let's start with chips.
The talk I went to yesterday was from Andrew Feldman,
who is the CEO and co founder of Cerebras, which
is an AI chip manufacturer that has big ambitions to
take on in video. They manufacture the world's largest AI
chips and they're hoping to IPO in the US this
year to add fuel to their ambition to kind of
take on in video, and the company recently started offering
(07:36):
deep Seek running on its servers house in the US.
Deep Seak obviously being the Chinese AI model company that
managed to create this high performing reasoning model despite US
export controls on in Nvidia's most advanced chips, and the
CEO of Feldman talked about being quote crushed by demand
ever since offering deep seak on these servers. But first
(07:57):
of all, let's take a couple of steps back, because
I think last year you did a great episode on
tech Stuff with the title what are AI chips? And
I think that a brief refresher while the world talks
about it would actually be super helpful.
Speaker 3 (08:10):
Oh. Absolutely so. If you think about processors in general,
if we're talking about your classic processors for computers, you
really have three major types at this point. You have CPUs,
the central processing unit that's the BOG standard basic processor.
Typically they're very good at performing high speed arithmetic operations
(08:35):
in sequence, right. And then you have GPUs or graphics
processing units. These we associate with things like gamers, like
if you're a real gamer, you got yourself a killer GPU.
These are really good at parallel processing, where they can
take multiple threads of operations and run processing on them simultaneously,
which can sometimes depending on the type of computer problem
(08:58):
work faster than a CPU could for specific types of processes. NPUs,
or neural processing units, are the new hotness, and they're
kind of like GPUs, but on steroids. They're even more
about parallel processing. They're optimized to run the types of
processes that your typical AI operations require these days. So
(09:20):
the laptop I have in front of me right now
has an NPU in it. So there's a neural processing
unit in my little laptop I've got that will run
those kind of operations natively on the computer that don't
require you to have a cloud connection, so that you're
not shipping all your data off to the cloud to
get processed and then sent back to you once it's done. Obviously,
that brings up questions of privacy and security. So one
(09:43):
of the big attractive features of having NPUs is you
can run those processes natively on your own devices and
not have to depend upon some third party being able
to access the information, especially in a world where we
worry about the information being used to train future model
of AI.
Speaker 2 (10:01):
Had you come across this company before.
Speaker 3 (10:03):
Cerebras only in Greek mythology, where the three headed hound
of Hades is guarding the Gates.
Speaker 2 (10:10):
No.
Speaker 3 (10:10):
Actually, I have heard of it before, but I had
not really looked into it. One of the downfalls of
my era of tech stuff, and I often said this,
is that being in the United States and being an
American meant that it had frequently a very strong American perspective,
and it meant that companies that were operating outside of
(10:32):
America often got less focus on my show. So one
of the things that I think is cool about the
Web Summit is I'm encountering companies that maybe I've heard
the name, but I really didn't know much about, and
I'm starting to learn a lot more about them and
have a greater appreciation for them.
Speaker 1 (10:49):
It was interesting that kind of deep seek was such
a big part of this talk, I mean not surprising.
When did this narrative of like chip Wars or the
chip race start to kind of you know, Chris to
your when is when you were hosting takes stuff.
Speaker 3 (11:02):
So a few years back in video was starting to
get incredibly popular, not through the AI world, but because
of cryptocurrency. Because for cryptocurrencies like ethereum, you would still
use parallel processing in order to attempt to mine a
block first and get that reward for things like bitcoin.
(11:26):
That value had gone so high that you were looking
at purpose built chips for that that we had gone
beyond GPUs. But in Video was riding high because of that,
and it meant the gamers were really upset because the
chips themselves were in very short supply and often where
you would find them in the aftermarket marked up to
ridiculous prices, and they're already quite expensive. So in Video
(11:48):
was already in the news then and their stock price
was soaring already because of the popularity of the chips.
Once the AI industry started to really take off toward
the end of twenty twenty two, that's when in Vidia
really flourished, and in an incredibly savvy move, they began
(12:08):
to reposition themselves as a company that made AI chips
and not just the GPU company.
Speaker 1 (12:15):
Right, they sort of forced themselves into being considered a
national security or a kind of critical natural company.
Speaker 3 (12:22):
Yeah, like they like they it was them or no one,
it was kind of the view. And so that's when
in Nvidia went from being really a pretty powerful company
to like in the top three gets often battling in
the top three spaces for most valuable company in the world.
Speaker 1 (12:39):
Because of this chip race or chip wars narrative that
they were able to both profit from it also drive.
Speaker 2 (12:46):
But also I mean it's true, right, I mean that yeah,
has been true.
Speaker 3 (12:49):
Yeah, yeah.
Speaker 2 (12:50):
What happened next?
Speaker 3 (12:50):
Yeah, it's it's not boasting if you can back it up, right,
And you certainly could argue that there was exploitation going on,
like they were exploiting the the narrative, and in some
part I'm sure that's true because as I'm certain you
have noticed the craze around AI, there's a nugget of
truthfulness at the center, but there's a lot of hype
(13:13):
around it.
Speaker 1 (13:14):
Yeah, yeah, yeah, yeah, yeah, Well it was interesting presentation yesterday.
I mean, I would say taking on in video is
a lofty goal, absolutely point.
Speaker 3 (13:23):
But we've seen that happen before too, right, Like if
you go.
Speaker 2 (13:27):
Back, I guess the way Vidio disrupted Intel, which is
on the floor.
Speaker 3 (13:29):
That's exactly what I was going to say, Like if
you if you go back to the nineties, you could
ask someone like do you think n video is going
to overtake Intel? And you'd be left all the building. Yeah,
so there is precedence.
Speaker 1 (13:40):
What did you I mean, you were out of the
chair when when deep Seak happened, that was like my
second week. Yeah, there's a text down host and I
was like, oh my goodness, this is a this is
for me. I think probably the biggest text story I've
seen since the chetchipt released in November twenty twenty two.
I mean it just the reaction from the world, from
the market, like everyone wants to know what's going on,
(14:02):
But what did you think in the moment, and how
have you digested it since then?
Speaker 3 (14:06):
So I had two very strong reactions when the deep
seek information started to become mainstream news. One was how
interesting because we've just been through a year of the
biggest fuss ever being made over TikTok, and here we
(14:26):
have an artificial intelligence system from China where the perceived
thread of TikTok is you could argue minuscule compared to
an AI company that is completely dependent upon devouring as
much information as possible and then making use of it,
Like that should be the big national security concern. If
(14:50):
in fact that's the real reason why you're concerned about TikTok,
we could go into a crazy conversation about whether or
not that's true. And I'm not going to do that here.
Speaker 2 (14:58):
But the second, especially because the TikTok booths is right
next to it.
Speaker 3 (15:04):
Well, listen, every time I go by, I give reverence
and I floss a little bit. But no, the the
other reaction I had was that, golly, people have really
been waiting for a challenger to chat GPT. And I
think that tells you that as excited as people are
about Chad GPT and open ai in general, the need
(15:28):
for there to be competition and for perhaps there to
be a check on open AI's otherwise dominance in the
space is one that a lot of people were feeling,
even if they weren't able to articulate it.
Speaker 2 (15:39):
Yeah, I think.
Speaker 1 (15:40):
I mean that the point that the surrebrass guy was
making yesterday was that there they run these you know,
servers with their with their chips that run the open
source version of deep Seek's model, but in the US,
so that la is some of the he was basically,
he said explicitly like, don't use the deep seak product
like you. I mean, that's that's kind of interesting. I
(16:02):
think the second point you make about people wanting a
challenger to open ai is very true. I mean, emotionally,
like nobody likes, you know, one big bully in the playground, right,
But I think when you're in a place like Doha,
you also understand how the idea of a competing ecosystem
of different AI models and maybe a cheaper way to
use AI is an incredibly attractive thing in a world
(16:25):
where you know, the concept of American hegemony is is
fast receding and people want to know what's coming next.
So this this tech element of geopolitics or something which
is super fascinating.
Speaker 3 (16:37):
Yeah, and again that's something that that's probably the most
valuable takeaway I'm going to have from this I say
that it's only the second day, but the most take
a powerful takeaway I'm going to have on this trip
is that opening of my perspective.
Speaker 1 (16:53):
Coming up, Jonathan Strickland and I dig into the future
of augmented reality.
Speaker 2 (16:58):
Stay with us.
Speaker 1 (17:06):
So there's a lot to say about chips, but but
I wanted to move on to another technology that you've
covered closely over the years, augmented reality. I also went
to a talk given by a gentleman called Keyppan of
snap and he came out wearing these very futuristic RoboCop
like ar Snapchat glasses, and his view of the audience
(17:26):
was supposed to be broadcast on the screens behind him. Unfortunately,
the WIFEI didn't want, so poor guy came out wearing
these you know, spectacles and then.
Speaker 3 (17:39):
No, no first person's view. Oh that hurts.
Speaker 2 (17:43):
I know when you especially when you're giving it. I mean,
you're like a thousand.
Speaker 1 (17:45):
People in the audience and and you've got your big
razzle dazzle opening moment doesn't work.
Speaker 3 (17:50):
I don't know about you, os, but I get I
get secondhand embarrassment really easily. How ridiculous did the glasses look?
Speaker 2 (17:58):
I mean they looked.
Speaker 1 (18:00):
There was a there was a famous Greek shipping billionaire
in the sixties called Aristotle Nassis who married Jackie Kennedy. Yeah,
and he wore these great big sunglasses that come about
half his face. These look literally like they'd been pulled
off Aristotle Aristotle and assets his face.
Speaker 3 (18:19):
I mean, I was at Cees the year that Lady
Gaga came out and promoted her polaroid glasses, and I'm like, okay.
Speaker 1 (18:26):
I've we've seen big, big, big glasses, but anyway, keep
handed a valiant job. And I was thinking, probably, you know,
as you are also feel quite a little empathy and
thinking in his head, you know, his whole mind is screaming,
this is.
Speaker 2 (18:39):
A failure, like blah blah.
Speaker 1 (18:40):
But he had to just keep going, soldiering on, and
to be fair, he did and actually it was a
good talk. He kind of broke down the three phases
of augmented reality at Snapchat. Phase one was the phone
capturing the face and the computer vision basically knowing.
Speaker 2 (18:57):
Those eyes, that's the nose, that's the mouth when you want.
Speaker 1 (18:59):
A puke rain, but that's where it should come from.
The second phase was the computer vision understanding the human
bodies and feet, legs, so that you could basically put
gloves on, you could wear a costume where you would
know what your arms, it.
Speaker 3 (19:17):
Could map properly to the right location on the body exactly.
Speaker 2 (19:21):
And then phase three is basically capturing the world.
Speaker 1 (19:24):
So these spectacles that are kind of looking out and
I keep calling spectacles because that's what the product is called.
They're called spectacles. They don't normally say spectacles. I normally
say glasses, right, just FYI. But basically, you know, the
big leap forward is that rather than just understanding the
human body, which is hard enough, This actually understands the
whole world around you in terms of being able to
interpret what objects are, knowing what a surface is, knowing
(19:46):
how far away things are, so that you can basically
create a virtual overlay on the physical world. And one
of the really cool demos was actually somebody wearing these
glasses and looking at a bunch of ingredients, and then
the glasses sent the picture of the food in real
time to the cloud, which bat back at JENNYI suggested
(20:07):
recipe cool.
Speaker 2 (20:09):
Yeah, which is fun.
Speaker 1 (20:09):
I mean it's not I can't imagine it's going to
be a huge consumer use case, but it's fun.
Speaker 3 (20:14):
It's making me think of I don't know if you
remember this, but Google had Google Chef for a while
where you could do something similar. But obviously you're typing
the things. It could not, you know, do image processing
and do this, but you could type in the things
you had and it would suggest different recipes for you.
But it had the same foibles i'll say as generative AI,
(20:35):
and that occasionally you would get something where you're like,
well that sounds inedible, but all right, I.
Speaker 1 (20:40):
Hope raid if I just looked in my fridge, the
glasses might just say get a breeze again.
Speaker 2 (20:45):
Exactly, it's time to take out Exactly. But you mentioned Google.
Speaker 1 (20:52):
I think you were the proud owner of some of
the very good specs.
Speaker 3 (20:56):
Yeah, so not only was I an owner of Google Glass.
Speaker 2 (21:00):
You removed the way proud. Yeah.
Speaker 3 (21:02):
Well, I was not going to call attention to it,
but it's true. I think Google Glass was a noble effort.
I think it was nowhere close to being ready to
be a consumer product, which I think Google actually understood.
I mean, they never really marketed it as a mainstream
consumer product. But when I had them, I could see
the potential, and I thought that it was an incredible
(21:24):
potential that certainly was nowhere close to being realized. Yet.
Let's say you're walking around Doha and it's the first
time you've ever been here. Having the ability to see
directions in front of your eyes are or at least like
in a way that's not going to obscure your vision,
so that you can seamlessly navigate a city you've never
been to, that's incredible. Or being able to look at
(21:45):
a building and get a listing of the different businesses
that are in there, so that when late at night,
when you're craving Nando's. You don't spend forty five minutes
walking ground a mall wondering where it is. You know
where it is. Immediately.
Speaker 1 (21:58):
In twenty seventeen, you didn't have of tech stuff which
asked the question is augmented reality ready for prime time?
Speaker 3 (22:05):
Yes?
Speaker 2 (22:05):
That was seven years ago.
Speaker 3 (22:06):
Yes I did, I did ask that, and no it's not. Well,
there's a chicken and egg problem that's going on, and
that is the classic chicken and egg problem of hardware
versus software. Basically, the idea is that you have people
who are making the hardware, and when they create something,
even if it's really compelling, if there's not enough applications
(22:27):
for that hardware, there's not enough reason for people to
buy it. On Moss. The flip side, if you're a
software developer and you're looking at you see this really
exciting technology come out and you're thinking, Wow, that's really cool,
But I don't want to start dedicating resources to building
assets for this hardware until there's enough of a user
base to justify it. We talk about this with game
(22:48):
consoles all the time. A new game console comes out
and you're like, the console's incredible, but there's nothing to
play on it, so I'm going to wait, And so
it's this ongoing circular issue and it could be really
hard to break out of that.
Speaker 2 (23:01):
I mean, that's that.
Speaker 1 (23:02):
I think the software side is pretty interesting, right, I
mean the advances in AI on the computer vision side
have been pretty amazing.
Speaker 2 (23:10):
Yes, yes, And you know.
Speaker 1 (23:12):
If this deep Seek story plays out and it becomes
way way, way cheaper to run AI models everywhere, the
idea of real time computer vision could potentially unlock this
air reality. And what do you think needs to happen
for the situation you just described of walking down the
street in Doha Jonathan Stricklan was wearing his glasses.
Speaker 3 (23:34):
I think probably the biggest issue, honestly is the battery issue,
right right, because you can only manaturize batteries so much
before you get to a point where you don't have
enough juice to provide power to a sophisticated device for
more than maybe an hour. And if that's all you
need it for, that's fine. But I think for a
lot of people, the thought of a device that they
can wear for an hour and then they need to
(23:55):
recharge it is quite frustrating. I think it's one of
the reasons why things like Active three glasses, we're a
non starter because even though they provide an incredible experience,
having to recharge your glasses every couple hours. Like you know,
if you're watching a Peter Jackson movie, you might have
to stop in the middle so that you can recharge
your glasses to watch the rest of the film. So
I think that's really a big issue, is that how
(24:17):
do you miniaturize the technology in such a way that
the glasses are something that you want to wear. They
look cool, but there's still enough power capacity there to
provide a good experience for more than a short time.
And I don't know the answer to that, because batteries
are dependent upon chemistry, and we can hack technology really
(24:38):
quite effectively, but chemistry you start to run against fundamental
laws of the universe and it starts to get a
lot trickier.
Speaker 1 (24:46):
So do you have a whole I haven't had a
chance to attend any of it, but there's a whole
kind of new energy track at this conference where they're
talking about new battery technology. Right, it's interesting to be
here where you know, natural gas is coming out of
the ground the tune of trillions of dollars. Yes, there's
a lot of interest here in investing some of the
proceeds in these chemistry issues.
Speaker 3 (25:07):
Yes, right, Yes, And it may be that there's some
breakthroughs that can come through that makes it more of
a practical application, right, And I think that's what really
is needed. It ends up being it transcends the issues
of hardware and software and it starts to get to
we've got all the technical capability here apart from where's
the power source coming from? Right, And I think that's
(25:29):
going to be the big thing, honestly. Obviously Apple was
trying to get there where they wanted to have a
really kind of lightweight pair of glasses that had incredible
augmented reality capabilities to it. Instead we got something that
like futuristic robotic ski goggles. You know. And if you think, well,
Apple being a multi trillion dollar company, if they're having
(25:50):
they're starting to hit a wall there. Then we're probably
at a point where it's just going to be a
while before we start seeing something that people think of
as being oh, this is something I want to wear
on a regular basis, as opposed to I have a
specific application in mind, I don't mind wearing it for
that application, whether it's like industrial or educational or military
(26:12):
of course, military or gaming, something like that where you're thinking,
all right, I'm gonna wear this for forty five minutes
to an hour or whatever, and then I put it aside,
Like I think the real dream of augmented reality is
you have something that you can wear pretty much all
the time and activate it whenever you need to.
Speaker 1 (26:35):
When we come back, jonathan' strictly and I continue to
unpack what we've learned at Websomitic Cutter stay with us.
So finally, the last talk that I went to I
wants to discuss with you, which you've already touched on,
was from al mah Latour, the CEO of Dow Jones,
(26:56):
which is the parent company at the Wall Street Journal,
and he spoke about the future of journalism in the
age of AI. Unsurprisingly, he was very focused on how
IP creators and owners such as news publishers, can hold
onto the value of their work, and he mentioned there
are two ways of achieving this, commercial partnerships and litigation.
Speaker 3 (27:18):
Yeah, yeah, I'm not surprised. So obviously this is a
very tricky topic, right, with a lot of different components
to it, one of which is that as people who
generate content. People who generate news content. We are very
well aware that the things we're putting out, which are
meant to inform, maybe entertain, maybe open up people's eyes
(27:40):
to new perspectives on certain topics, is also being used
to train AI. So one component that scares me about
AI and journalism is just the idea of AI benefiting
from the work of journalists and the journalists see no
benefit in return, right, So it's the copyright issue essentially
is what comes down to. And then the flip side
(28:02):
is the concern, like what I saw at hell Stuff works,
where a company and perhaps it's a company that's actually
in dire straits and they're really looking like, how can
we reach a point where we're still able to provide
the services that we're trying to provide without bleeding ourselves
dry because we're in a business that it just does
(28:22):
not have a huge return on investment, and sadly, the
media often can be that. So you have these companies
that have an incentive to say, well, you know, human
resources are staff. That's a really big expense, and if
we could just cut them and use this tool to
do the same thing they were doing and perhaps just
have a few people left behind to massage whatever is
(28:46):
made into something that the general public can consume, and
maybe they don't even notice the difference. Why don't we
do that? And we've seen why you don't do that.
It's because the tools that are being made, they are fallible,
and sometimes to a point that is disturbing. I told
you this OZ, and there's an episode of tech stuff.
(29:08):
I think it's called something like. AI wrote this episode
sort of and in that I had it write an
episode of tech stuff. All I gave it was very
simple instructions. I used chat GPT and I said, write
an episode of tech stuff about the technology of airbags.
That was it, like. I didn't give it any further instruction,
and as part of what it regurgitated to me, it
(29:31):
gave me statements from three supposed experts, but none of
those people existed, which means automatically you cannot trust the information.
And when I would ask for things like I followed
up I said, could you give me a source for
this information? And it wouldn't be able to. So these
things are going to be things that improve over time.
(29:51):
But I think the problems we're seeing now concern me
because whether the technology actually gets better, so it's being
more accountable, or it just gets better at obfus skating
when it's making stuff up. I don't know that we'll
be able to tell the difference.
Speaker 1 (30:07):
Yeah, I mean, Latour was talking about how basically, how
do we at Dow Jones use these tools to make
sure we had compensated for our work so that we
can make more money so that we can fund more journalism.
Speaker 3 (30:20):
It's a very pragmatic sort of look, but it makes sense, like, yeah,
the good balancing art and commerce is always a struggle,
even no matter what outlet you're looking at.
Speaker 1 (30:31):
It's interesting as well, though, because it's not He wasn't
just talking in abstraction, Like right now, Dow Jones is
in partnership with open Ai and suing Perplexity.
Speaker 3 (30:41):
Wow.
Speaker 1 (30:41):
So with open Ai they made this deal last year
that The Wall Street Journal reported on its parent company
could be worth more than two hundred and fifteen million
dollars over the next five years in both cash and
credits for open ai technology.
Speaker 2 (30:56):
But last year Dow.
Speaker 1 (30:57):
Jones also sued Perplexity the AI search engine, and the
suit's not public and al Maesa didn't want to go
into details about what obviously, despite the best efforts of
Sarah Fisher from Axios on stage, but why it did
report on the lawsuit and they basically, you know, the
headline was that dow Jones alleges Perplexity is hallucinating fake
(31:17):
news and attributing it to real papers and that's illegal.
Speaker 3 (31:21):
Oh yeah, so very similar to what I was just saying,
exact right, Yeah, this is this is like someone writing
a term paper and having a citation for a source
that doesn't exist.
Speaker 1 (31:31):
It's like somebody making something up and then citing it
to a professor when the professor never said.
Speaker 3 (31:35):
It, right, right, right, right, And yeah.
Speaker 2 (31:37):
That's what that's what this suit is about.
Speaker 1 (31:39):
Absolutely, yea, references in a way that's even more egregious
to your point about AI getting better at covering its
tracks that it makes up up when you're doing fake citations,
that's pretty.
Speaker 3 (31:49):
I mean, it's it's terrifying because like obviously this is
this ends up becoming like a political term in some cases,
but like the whole fake news thing which often was
being used to try and delegitimize real news sources that
were just saying things you didn't like, Yea, you call
it fake news. So that you can dismiss it. But
we're actually talking about actual fake news, the news about
(32:12):
stuff that didn't happen or at least didn't happen the
way that you're being told it did, but it's being
put down in a record as if that, in fact
is what happened. How do you then get to the
point where you can determine reality from computational fiction.
Speaker 1 (32:30):
Yeah, yeah, I mean the litigation stuff is very interesting.
Obviously if you look back to the Victorian times and
the aftermath of the First Industrial Revolution, that was when
copywriting laws took effect. And so you know, as the
New York Times law suit against open Ai potentially comes
to trial, as Dow Jones lawsuit against Perplexity comes to trial, potentially,
(32:51):
we're going to see new president and it'll be really
interesting to see how that shakes out and how that
shapes our future. I mean, the power of law to
shape reality is so huge, and copyright law has been
such a fundamental of our society and who we are
for so long.
Speaker 3 (33:06):
Yeah, absolutely fascinating. It Also it goes back to that
old saying. It's almost a cliche, but you know, necessity
is the mother of invention and in this case, necessity
is the mother of the need for new laws. So
with copyright, before you get to things like mass printing,
copyright really wasn't that big of a concern, just because
(33:27):
it was such a pain in the butt to make
a copy of a work. But then you get to
a point where technology is capable of giving people opportunities
to do things on a much larger scale. That's where
you start to see the need for new law. And
I suspect we're going to see a lot of changes
to law in the next few years as a result
(33:48):
of the rise of technologies like AI.
Speaker 2 (33:51):
Yeah, I also thought it was interesting.
Speaker 1 (33:52):
I mean, obviously, dow Jones is, you know, a Murdock company,
so perhaps no surprise that they're pugnacious bunch.
Speaker 3 (34:02):
Yes that's pretty good, kindly, but you.
Speaker 1 (34:06):
Know, but the point that he made was basically, when
the Internet emerged, you know, twenty five years ago ish
as a mass adoption technology, the news organizations, the music
rights holders, you know, film and TV industry essentially allowed
the consumer expectation to develop the content was free, yes,
and that essentially hollowed out those industries to to you know,
(34:30):
to the extent we're seeing today where all of them
are really struggling. So his point was, how do we
make sure we we'ret a new technological inflection point that
we don't roll over again? And you know, so I
thought that was pretty interesting.
Speaker 3 (34:41):
You know, that's an incredibly good point. I mean, we
are where we are today in part because we've all
had the expectation of the content that we want to
access at any given moment and any point of the
day should be easy and free to access. That's a
great eight thing in many ways for the individual, but
(35:02):
it does put an incredible burden on the entities that
are actually creating that, whether you are a solo content
creator and you're just trying to make your passion your occupation,
or you're a big media company and you're trying to
put out really incredible content, which costs a lot of
money to do, and then you're getting tiny little pennies
(35:24):
in return. It's really hard in today's world to have
the same kind of media output that we would expect
from say the early two thousands and have it be
a profitable business. And if there is no profit, there
is no business. If there's no business, there's no content.
So ultimately it does come back down to us to say,
(35:44):
how do we create a world where this technology can
be practical and useful, not dangerous, and not something that
bankrupts any company that actually gets into the business of
doing it.
Speaker 1 (35:59):
And that was the other thing Alma was talking about
in terms of some of the positive use of AI.
Speaker 2 (36:04):
He talked about.
Speaker 1 (36:05):
Helping Dow Jones do better research. He talked about scaling
the output across multiple languages. Apparently Dow Jones wire service
is now available in Korean, largely translated by AI at
such a good enough level that Korean traders are able
to trade off it and also using I to make
their content more interactive.
Speaker 2 (36:24):
How do you feel about some of these uses of AI.
Speaker 3 (36:26):
I think that that's incredible. Like I think the idea
of using AI to improve accessibility to me is a
no brainer. Use for AI, making something that has got
value and usefulness to someone's life accessible to them is transformational.
It doesn't do me any good If the information I
(36:47):
need is in a format that I cannot access, it
might as well not exist. So I look at how
transformative the Internet has been over the last couple of decades.
I think I'm old enough where I remember when I
first saw the web like which I famously dismissed at
the time because it was so slow. But when the
(37:07):
web started to really take on properties that were undeniably
useful and transformative, I just took it for granted that
I was able to access all of it because it
largely came out of the United States. Now, if I
had lived anywhere else where I didn't know English, knowing
that that existed might have been interesting to me, but
(37:28):
it wouldn't be practical. I wouldn't be able to do
anything with it. So using AI to do things like translate,
whether it's a podcast or written piece of work or
a movie, that to me is one of the best uses,
the most incredible uses of AI. And obviously it's not perfect,
but it's really good. The issues we tend to run
(37:49):
into are things like idioms and cultural references that are
not easily translatable from one language to another. Those are
problems that are going to just remain for quite a while.
I don't know how long it's going to take us
to teach robots what these idioms mean. That's going to
be an issue that's ongoing. But I think that's a
very noble use of AI.
Speaker 1 (38:12):
Thank you so much, Jonathan. Thanks for joining today in Doha,
and I hope we'll do it again soon.
Speaker 3 (38:17):
Absolutely, I'll head up to New York if it's a
much shorter plane.
Speaker 2 (38:21):
Ride, that's for sure.
Speaker 1 (38:24):
That's it for this week for Tech Stuff, I'm mos Voloshin.
This episode was produced by Eliza Dennis. It was executive
produced by me Karen Price, and Kate Osborne for Kaleidoscope
and Katrina Norvelle for iHeart Podcasts.
Speaker 2 (38:36):
Please rate, review, and reach out
Speaker 1 (38:38):
To us at tech Stuff podcast at gmail dot com.
Welcome to tech Stuff, a production of iHeart Podcasts and Kaleidoscope.
This week, we're taking you on the road for a
special episode recorded in Doha, Cutter. Now, before you start
imagining sand.
Speaker 2 (00:12):
Dunes and pure blue coastlines, I should describe my surroundings.
Speaker 1 (00:17):
I'm currently sitting in a makeshift studio in a massive
airplane hangar like conference center called the Doha Exhibition and
Convention Center, So if this episode sounds a little bit.
Speaker 2 (00:28):
Different, you'll know why.
Speaker 1 (00:30):
For the last couple of days, I've been attending Web
Summit Cutter, an international conference where investors and entrepreneurs and
thought leaders.
Speaker 2 (00:37):
Gather from around the world to talk about the future
of technology.
Speaker 1 (00:41):
Our friends at iHeart are the official podcast partner's Web Summit,
which means that I'm here and so is a very
familiar figure, Jonathan Strickland.
Speaker 3 (00:51):
Well, hello, I'm so pleased to be a guest on
this historic, amazing podcast. I've heard so much about.
Speaker 1 (00:57):
It, Jonathan, Welcome back to tech Stuff. How does it
feel to be on the other side of the table.
Speaker 3 (01:01):
I'll be honest, it feels odd, but I'm so pleased
to be able to be a part of this. I
was been joking all week about you know, just when
I thought I was out, they pulled me back in.
Speaker 2 (01:15):
Absolutely. Well, I'm very pleased to see you.
Speaker 1 (01:17):
It's been almost seven years, I think since we saw
each other in person, So who knew the next time
would be in Doha?
Speaker 3 (01:22):
And yeah, here I am here we are.
Speaker 1 (01:24):
Based on the listener emails I've been receiving, I think
the listeners are going to be very, very very happy
to hear your voice today.
Speaker 2 (01:30):
You're already missed.
Speaker 3 (01:31):
Well just for you guys, Hey there, welcome to tech stuff.
There you go, what the tech? The tech?
Speaker 1 (01:38):
So nice have you vation? And I hope, I hope
this will be the first of many visits. Absolutely, I'm
not just in Doha. Yes, I would love to do
this on East time. So there've been some fascinating speakers
already at the conference, and I wanted to talk to you,
Jonathan today about some of what I've heard, hoping that
you can contextualize it kind of from a bird's eye
view of having covered technology as a journalist for two
(02:00):
decades sure and attended a bunch of these conferences.
Speaker 3 (02:04):
Yeah, conferences in the tech world in general play an
interesting role. So I've been to lots of different kinds.
I've been to trade shows like CEES or E three,
I've been to ones that have been held by specific
companies such as Intel or IBM and the web Sumit
of course is more of a general tech with Internet
(02:25):
focus on it, and there have been some pretty phenomenal speakers. Actually,
they've really impressed me.
Speaker 1 (02:32):
And when you come to these types of events, like
what do you look out for, both in terms of
kind of getting a peek around the corner to what
may be coming in the world of tech, but also
in terms of applying some kind of critical lens to
what you're hearing.
Speaker 3 (02:45):
That is an excellent question. So let's take CES for example.
Everyone is looking for the quirky thing that is a
standout from the standard stuff you would run into. So
an example would be from a couple of years ago,
the Little Rabbit AI, you know, pocket sized computer device,
which ended up not doing very well when it finally
(03:05):
came out. That's often the case where something that's quirky
gets your attention because it's different, but it turns out
it's not really practical or maybe it doesn't work as
well as was presented. So that's one thing you have
to look out for is yes, it's exciting and that
it's different, but at the same time you have to
ask the question, Okay, does this actually solve a problem.
One of the speakers we had yesterday actually talked about
(03:28):
this about how innovation for innovation's sake is a fool's
errand and that really innovation needs to be applied to
looking at real world problems, not inventing one. So that's
one thing, But I'd say another is you look to
see what people are talking about, what they're excited about,
and how they're talking about it, so that you can
(03:51):
kind of get a feel for what is going to
be the next push, and keep in mind that not
all those pushes are going to be successes. So as
an examp back in two thousand and eight when I
went to my first CEES, the big push back then
was from companies like Panasonic and Sony, where it was
all about three D television. Yeah, now, you and I
know that three D television never really took off. People
(04:14):
balked at the idea of having glasses that they had
to wear at home, and the expense and the lack
of content. All of these things contributed to the failure
of that technology. So when you're being exposed to these
new ideas and people are really excited about them, you
have to temper that a little bit with the reminder
that not everything that's being talked about is going to manifest,
(04:37):
or if it does, it won't manifest in the way
we anticipate. But if you do that and you proceed
with a kind of cautious optimism, I think great things
can be achieved.
Speaker 1 (04:48):
Yeah, with that in mind, I mean, there are a
bunch of kind of smaller booths where people are presenting
on the floor, largely on laptops. I guess it's like
AI applications, but then as a main stage where there's
been some really interesting speakers.
Speaker 2 (05:01):
So I've into three talks so far.
Speaker 1 (05:03):
I kind of want to talk about each of them
briefly to get your view of how they exist in
the context of the history of technology.
Speaker 3 (05:10):
Sure.
Speaker 1 (05:11):
The first one was about the global chip race and
the continuing ways from deep seeg the second was about
the augmented reality presence in future, and the third was
AI and journalism. A fun time, a fun top, especially
for US journalists to come of technology.
Speaker 3 (05:26):
Yeah, especially for those of us who worked for a
company that famously eliminated its editorial board in order to
replace it with AI.
Speaker 2 (05:34):
Wait wait which company?
Speaker 3 (05:35):
How stuff Works? Really? Yeah, back in twenty twenty three.
People have heard me talk about this on the show before,
but if you're new here, I worked for HowStuffWorks dot
com and I was a writer there, and in twenty
twenty three the company decided to stop working with freelance writers.
They had a freelance writing crew, and then they had
an editorial crew in house, and the idea was that
(05:58):
they were going to go with AI general rated articles
from that point forward. The editors would have to do
a full edit pass to make sure that everything was
accurate and correct, and then correct anything that wasn't right,
and as you know, generative AI can sometimes confabulate and
make things up. The editors protested, and then they were eliminated,
and everyone I knew who worked at how Stuff Works
(06:20):
was let go.
Speaker 2 (06:21):
Wow.
Speaker 3 (06:21):
Yeah, So I say this not so people will take
up torches and pitchforks and yell or anything, but rather
to explain that I am very much aware that I
have a very strong bias here, like I cannot be objective,
is what I'm trying to get.
Speaker 1 (06:36):
At I mean, I think if you work as a
journalist and you know you're constantly hearing about the idea
of being replaced by AI, it h it's not the
most appealing seat to sit in.
Speaker 2 (06:46):
This for sure.
Speaker 3 (06:47):
And we've been through similar shifts before, like the pivot
to video being the famous one, right, we saw editorial
departments decimated because the idea was that the written word
was no longer the way to deliver and it was
going to be video only. But it didn't last very long,
and then by that time you had all these journalists
who were out of work. So obviously, as journalists, we
(07:11):
have a sensitivity to these things.
Speaker 1 (07:14):
So we'll come back to that, but let's start with chips.
The talk I went to yesterday was from Andrew Feldman,
who is the CEO and co founder of Cerebras, which
is an AI chip manufacturer that has big ambitions to
take on in video. They manufacture the world's largest AI
chips and they're hoping to IPO in the US this
year to add fuel to their ambition to kind of
take on in video, and the company recently started offering
(07:36):
deep Seek running on its servers house in the US.
Deep Seak obviously being the Chinese AI model company that
managed to create this high performing reasoning model despite US
export controls on in Nvidia's most advanced chips, and the
CEO of Feldman talked about being quote crushed by demand
ever since offering deep seak on these servers. But first
(07:57):
of all, let's take a couple of steps back, because
I think last year you did a great episode on
tech Stuff with the title what are AI chips? And
I think that a brief refresher while the world talks
about it would actually be super helpful.
Speaker 3 (08:10):
Oh. Absolutely so. If you think about processors in general,
if we're talking about your classic processors for computers, you
really have three major types at this point. You have CPUs,
the central processing unit that's the BOG standard basic processor.
Typically they're very good at performing high speed arithmetic operations
(08:35):
in sequence, right. And then you have GPUs or graphics
processing units. These we associate with things like gamers, like
if you're a real gamer, you got yourself a killer GPU.
These are really good at parallel processing, where they can
take multiple threads of operations and run processing on them simultaneously,
which can sometimes depending on the type of computer problem
(08:58):
work faster than a CPU could for specific types of processes. NPUs,
or neural processing units, are the new hotness, and they're
kind of like GPUs, but on steroids. They're even more
about parallel processing. They're optimized to run the types of
processes that your typical AI operations require these days. So
(09:20):
the laptop I have in front of me right now
has an NPU in it. So there's a neural processing
unit in my little laptop I've got that will run
those kind of operations natively on the computer that don't
require you to have a cloud connection, so that you're
not shipping all your data off to the cloud to
get processed and then sent back to you once it's done. Obviously,
that brings up questions of privacy and security. So one
(09:43):
of the big attractive features of having NPUs is you
can run those processes natively on your own devices and
not have to depend upon some third party being able
to access the information, especially in a world where we
worry about the information being used to train future model
of AI.
Speaker 2 (10:01):
Had you come across this company before.
Speaker 3 (10:03):
Cerebras only in Greek mythology, where the three headed hound
of Hades is guarding the Gates.
Speaker 2 (10:10):
No.
Speaker 3 (10:10):
Actually, I have heard of it before, but I had
not really looked into it. One of the downfalls of
my era of tech stuff, and I often said this,
is that being in the United States and being an
American meant that it had frequently a very strong American perspective,
and it meant that companies that were operating outside of
(10:32):
America often got less focus on my show. So one
of the things that I think is cool about the
Web Summit is I'm encountering companies that maybe I've heard
the name, but I really didn't know much about, and
I'm starting to learn a lot more about them and
have a greater appreciation for them.
Speaker 1 (10:49):
It was interesting that kind of deep seek was such
a big part of this talk, I mean not surprising.
When did this narrative of like chip Wars or the
chip race start to kind of you know, Chris to
your when is when you were hosting takes stuff.
Speaker 3 (11:02):
So a few years back in video was starting to
get incredibly popular, not through the AI world, but because
of cryptocurrency. Because for cryptocurrencies like ethereum, you would still
use parallel processing in order to attempt to mine a
block first and get that reward for things like bitcoin.
(11:26):
That value had gone so high that you were looking
at purpose built chips for that that we had gone
beyond GPUs. But in Video was riding high because of that,
and it meant the gamers were really upset because the
chips themselves were in very short supply and often where
you would find them in the aftermarket marked up to
ridiculous prices, and they're already quite expensive. So in Video
(11:48):
was already in the news then and their stock price
was soaring already because of the popularity of the chips.
Once the AI industry started to really take off toward
the end of twenty twenty two, that's when in Vidia
really flourished, and in an incredibly savvy move, they began
(12:08):
to reposition themselves as a company that made AI chips
and not just the GPU company.
Speaker 1 (12:15):
Right, they sort of forced themselves into being considered a
national security or a kind of critical natural company.
Speaker 3 (12:22):
Yeah, like they like they it was them or no one,
it was kind of the view. And so that's when
in Nvidia went from being really a pretty powerful company
to like in the top three gets often battling in
the top three spaces for most valuable company in the world.
Speaker 1 (12:39):
Because of this chip race or chip wars narrative that
they were able to both profit from it also drive.
Speaker 2 (12:46):
But also I mean it's true, right, I mean that yeah,
has been true.
Speaker 3 (12:49):
Yeah, yeah.
Speaker 2 (12:50):
What happened next?
Speaker 3 (12:50):
Yeah, it's it's not boasting if you can back it up, right,
And you certainly could argue that there was exploitation going on,
like they were exploiting the the narrative, and in some
part I'm sure that's true because as I'm certain you
have noticed the craze around AI, there's a nugget of
truthfulness at the center, but there's a lot of hype
(13:13):
around it.
Speaker 1 (13:14):
Yeah, yeah, yeah, yeah, yeah, Well it was interesting presentation yesterday.
I mean, I would say taking on in video is
a lofty goal, absolutely point.
Speaker 3 (13:23):
But we've seen that happen before too, right, Like if
you go.
Speaker 2 (13:27):
Back, I guess the way Vidio disrupted Intel, which is
on the floor.
Speaker 3 (13:29):
That's exactly what I was going to say, Like if
you if you go back to the nineties, you could
ask someone like do you think n video is going
to overtake Intel? And you'd be left all the building. Yeah,
so there is precedence.
Speaker 1 (13:40):
What did you I mean, you were out of the
chair when when deep Seak happened, that was like my
second week. Yeah, there's a text down host and I
was like, oh my goodness, this is a this is
for me. I think probably the biggest text story I've
seen since the chetchipt released in November twenty twenty two.
I mean it just the reaction from the world, from
the market, like everyone wants to know what's going on,
(14:02):
But what did you think in the moment, and how
have you digested it since then?
Speaker 3 (14:06):
So I had two very strong reactions when the deep
seek information started to become mainstream news. One was how
interesting because we've just been through a year of the
biggest fuss ever being made over TikTok, and here we
(14:26):
have an artificial intelligence system from China where the perceived
thread of TikTok is you could argue minuscule compared to
an AI company that is completely dependent upon devouring as
much information as possible and then making use of it,
Like that should be the big national security concern. If
(14:50):
in fact that's the real reason why you're concerned about TikTok,
we could go into a crazy conversation about whether or
not that's true. And I'm not going to do that here.
Speaker 2 (14:58):
But the second, especially because the TikTok booths is right
next to it.
Speaker 3 (15:04):
Well, listen, every time I go by, I give reverence
and I floss a little bit. But no, the the
other reaction I had was that, golly, people have really
been waiting for a challenger to chat GPT. And I
think that tells you that as excited as people are
about Chad GPT and open ai in general, the need
(15:28):
for there to be competition and for perhaps there to
be a check on open AI's otherwise dominance in the
space is one that a lot of people were feeling,
even if they weren't able to articulate it.
Speaker 2 (15:39):
Yeah, I think.
Speaker 1 (15:40):
I mean that the point that the surrebrass guy was
making yesterday was that there they run these you know,
servers with their with their chips that run the open
source version of deep Seek's model, but in the US,
so that la is some of the he was basically,
he said explicitly like, don't use the deep seak product
like you. I mean, that's that's kind of interesting. I
(16:02):
think the second point you make about people wanting a
challenger to open ai is very true. I mean, emotionally,
like nobody likes, you know, one big bully in the playground, right,
But I think when you're in a place like Doha,
you also understand how the idea of a competing ecosystem
of different AI models and maybe a cheaper way to
use AI is an incredibly attractive thing in a world
(16:25):
where you know, the concept of American hegemony is is
fast receding and people want to know what's coming next.
So this this tech element of geopolitics or something which
is super fascinating.
Speaker 3 (16:37):
Yeah, and again that's something that that's probably the most
valuable takeaway I'm going to have from this I say
that it's only the second day, but the most take
a powerful takeaway I'm going to have on this trip
is that opening of my perspective.
Speaker 1 (16:53):
Coming up, Jonathan Strickland and I dig into the future
of augmented reality.
Speaker 2 (16:58):
Stay with us.
Speaker 1 (17:06):
So there's a lot to say about chips, but but
I wanted to move on to another technology that you've
covered closely over the years, augmented reality. I also went
to a talk given by a gentleman called Keyppan of
snap and he came out wearing these very futuristic RoboCop
like ar Snapchat glasses, and his view of the audience
(17:26):
was supposed to be broadcast on the screens behind him. Unfortunately,
the WIFEI didn't want, so poor guy came out wearing
these you know, spectacles and then.
Speaker 3 (17:39):
No, no first person's view. Oh that hurts.
Speaker 2 (17:43):
I know when you especially when you're giving it. I mean,
you're like a thousand.
Speaker 1 (17:45):
People in the audience and and you've got your big
razzle dazzle opening moment doesn't work.
Speaker 3 (17:50):
I don't know about you, os, but I get I
get secondhand embarrassment really easily. How ridiculous did the glasses look?
Speaker 2 (17:58):
I mean they looked.
Speaker 1 (18:00):
There was a there was a famous Greek shipping billionaire
in the sixties called Aristotle Nassis who married Jackie Kennedy. Yeah,
and he wore these great big sunglasses that come about
half his face. These look literally like they'd been pulled
off Aristotle Aristotle and assets his face.
Speaker 3 (18:19):
I mean, I was at Cees the year that Lady
Gaga came out and promoted her polaroid glasses, and I'm like, okay.
Speaker 1 (18:26):
I've we've seen big, big, big glasses, but anyway, keep
handed a valiant job. And I was thinking, probably, you know,
as you are also feel quite a little empathy and
thinking in his head, you know, his whole mind is screaming,
this is.
Speaker 2 (18:39):
A failure, like blah blah.
Speaker 1 (18:40):
But he had to just keep going, soldiering on, and
to be fair, he did and actually it was a
good talk. He kind of broke down the three phases
of augmented reality at Snapchat. Phase one was the phone
capturing the face and the computer vision basically knowing.
Speaker 2 (18:57):
Those eyes, that's the nose, that's the mouth when you want.
Speaker 1 (18:59):
A puke rain, but that's where it should come from.
The second phase was the computer vision understanding the human
bodies and feet, legs, so that you could basically put
gloves on, you could wear a costume where you would
know what your arms, it.
Speaker 3 (19:17):
Could map properly to the right location on the body exactly.
Speaker 2 (19:21):
And then phase three is basically capturing the world.
Speaker 1 (19:24):
So these spectacles that are kind of looking out and
I keep calling spectacles because that's what the product is called.
They're called spectacles. They don't normally say spectacles. I normally
say glasses, right, just FYI. But basically, you know, the
big leap forward is that rather than just understanding the
human body, which is hard enough, This actually understands the
whole world around you in terms of being able to
interpret what objects are, knowing what a surface is, knowing
(19:46):
how far away things are, so that you can basically
create a virtual overlay on the physical world. And one
of the really cool demos was actually somebody wearing these
glasses and looking at a bunch of ingredients, and then
the glasses sent the picture of the food in real
time to the cloud, which bat back at JENNYI suggested
(20:07):
recipe cool.
Speaker 2 (20:09):
Yeah, which is fun.
Speaker 1 (20:09):
I mean it's not I can't imagine it's going to
be a huge consumer use case, but it's fun.
Speaker 3 (20:14):
It's making me think of I don't know if you
remember this, but Google had Google Chef for a while
where you could do something similar. But obviously you're typing
the things. It could not, you know, do image processing
and do this, but you could type in the things
you had and it would suggest different recipes for you.
But it had the same foibles i'll say as generative AI,
(20:35):
and that occasionally you would get something where you're like,
well that sounds inedible, but all right, I.
Speaker 1 (20:40):
Hope raid if I just looked in my fridge, the
glasses might just say get a breeze again.
Speaker 2 (20:45):
Exactly, it's time to take out Exactly. But you mentioned Google.
Speaker 1 (20:52):
I think you were the proud owner of some of
the very good specs.
Speaker 3 (20:56):
Yeah, so not only was I an owner of Google Glass.
Speaker 2 (21:00):
You removed the way proud. Yeah.
Speaker 3 (21:02):
Well, I was not going to call attention to it,
but it's true. I think Google Glass was a noble effort.
I think it was nowhere close to being ready to
be a consumer product, which I think Google actually understood.
I mean, they never really marketed it as a mainstream
consumer product. But when I had them, I could see
the potential, and I thought that it was an incredible
(21:24):
potential that certainly was nowhere close to being realized. Yet.
Let's say you're walking around Doha and it's the first
time you've ever been here. Having the ability to see
directions in front of your eyes are or at least like
in a way that's not going to obscure your vision,
so that you can seamlessly navigate a city you've never
been to, that's incredible. Or being able to look at
(21:45):
a building and get a listing of the different businesses
that are in there, so that when late at night,
when you're craving Nando's. You don't spend forty five minutes
walking ground a mall wondering where it is. You know
where it is. Immediately.
Speaker 1 (21:58):
In twenty seventeen, you didn't have of tech stuff which
asked the question is augmented reality ready for prime time?
Speaker 3 (22:05):
Yes?
Speaker 2 (22:05):
That was seven years ago.
Speaker 3 (22:06):
Yes I did, I did ask that, and no it's not. Well,
there's a chicken and egg problem that's going on, and
that is the classic chicken and egg problem of hardware
versus software. Basically, the idea is that you have people
who are making the hardware, and when they create something,
even if it's really compelling, if there's not enough applications
(22:27):
for that hardware, there's not enough reason for people to
buy it. On Moss. The flip side, if you're a
software developer and you're looking at you see this really
exciting technology come out and you're thinking, Wow, that's really cool,
But I don't want to start dedicating resources to building
assets for this hardware until there's enough of a user
base to justify it. We talk about this with game
(22:48):
consoles all the time. A new game console comes out
and you're like, the console's incredible, but there's nothing to
play on it, so I'm going to wait, And so
it's this ongoing circular issue and it could be really
hard to break out of that.
Speaker 2 (23:01):
I mean, that's that.
Speaker 1 (23:02):
I think the software side is pretty interesting, right, I
mean the advances in AI on the computer vision side
have been pretty amazing.
Speaker 2 (23:10):
Yes, yes, And you know.
Speaker 1 (23:12):
If this deep Seek story plays out and it becomes
way way, way cheaper to run AI models everywhere, the
idea of real time computer vision could potentially unlock this
air reality. And what do you think needs to happen
for the situation you just described of walking down the
street in Doha Jonathan Stricklan was wearing his glasses.
Speaker 3 (23:34):
I think probably the biggest issue, honestly is the battery issue,
right right, because you can only manaturize batteries so much
before you get to a point where you don't have
enough juice to provide power to a sophisticated device for
more than maybe an hour. And if that's all you
need it for, that's fine. But I think for a
lot of people, the thought of a device that they
can wear for an hour and then they need to
(23:55):
recharge it is quite frustrating. I think it's one of
the reasons why things like Active three glasses, we're a
non starter because even though they provide an incredible experience,
having to recharge your glasses every couple hours. Like you know,
if you're watching a Peter Jackson movie, you might have
to stop in the middle so that you can recharge
your glasses to watch the rest of the film. So
I think that's really a big issue, is that how
(24:17):
do you miniaturize the technology in such a way that
the glasses are something that you want to wear. They
look cool, but there's still enough power capacity there to
provide a good experience for more than a short time.
And I don't know the answer to that, because batteries
are dependent upon chemistry, and we can hack technology really
(24:38):
quite effectively, but chemistry you start to run against fundamental
laws of the universe and it starts to get a
lot trickier.
Speaker 1 (24:46):
So do you have a whole I haven't had a
chance to attend any of it, but there's a whole
kind of new energy track at this conference where they're
talking about new battery technology. Right, it's interesting to be
here where you know, natural gas is coming out of
the ground the tune of trillions of dollars. Yes, there's
a lot of interest here in investing some of the
proceeds in these chemistry issues.
Speaker 3 (25:07):
Yes, right, Yes, And it may be that there's some
breakthroughs that can come through that makes it more of
a practical application, right, And I think that's what really
is needed. It ends up being it transcends the issues
of hardware and software and it starts to get to
we've got all the technical capability here apart from where's
the power source coming from? Right, And I think that's
(25:29):
going to be the big thing, honestly. Obviously Apple was
trying to get there where they wanted to have a
really kind of lightweight pair of glasses that had incredible
augmented reality capabilities to it. Instead we got something that
like futuristic robotic ski goggles. You know. And if you think, well,
Apple being a multi trillion dollar company, if they're having
(25:50):
they're starting to hit a wall there. Then we're probably
at a point where it's just going to be a
while before we start seeing something that people think of
as being oh, this is something I want to wear
on a regular basis, as opposed to I have a
specific application in mind, I don't mind wearing it for
that application, whether it's like industrial or educational or military
(26:12):
of course, military or gaming, something like that where you're thinking,
all right, I'm gonna wear this for forty five minutes
to an hour or whatever, and then I put it aside,
Like I think the real dream of augmented reality is
you have something that you can wear pretty much all
the time and activate it whenever you need to.
Speaker 1 (26:35):
When we come back, jonathan' strictly and I continue to
unpack what we've learned at Websomitic Cutter stay with us.
So finally, the last talk that I went to I
wants to discuss with you, which you've already touched on,
was from al mah Latour, the CEO of Dow Jones,
(26:56):
which is the parent company at the Wall Street Journal,
and he spoke about the future of journalism in the
age of AI. Unsurprisingly, he was very focused on how
IP creators and owners such as news publishers, can hold
onto the value of their work, and he mentioned there
are two ways of achieving this, commercial partnerships and litigation.
Speaker 3 (27:18):
Yeah, yeah, I'm not surprised. So obviously this is a
very tricky topic, right, with a lot of different components
to it, one of which is that as people who
generate content. People who generate news content. We are very
well aware that the things we're putting out, which are
meant to inform, maybe entertain, maybe open up people's eyes
(27:40):
to new perspectives on certain topics, is also being used
to train AI. So one component that scares me about
AI and journalism is just the idea of AI benefiting
from the work of journalists and the journalists see no
benefit in return, right, So it's the copyright issue essentially
is what comes down to. And then the flip side
(28:02):
is the concern, like what I saw at hell Stuff works,
where a company and perhaps it's a company that's actually
in dire straits and they're really looking like, how can
we reach a point where we're still able to provide
the services that we're trying to provide without bleeding ourselves
dry because we're in a business that it just does
(28:22):
not have a huge return on investment, and sadly, the
media often can be that. So you have these companies
that have an incentive to say, well, you know, human
resources are staff. That's a really big expense, and if
we could just cut them and use this tool to
do the same thing they were doing and perhaps just
have a few people left behind to massage whatever is
(28:46):
made into something that the general public can consume, and
maybe they don't even notice the difference. Why don't we
do that? And we've seen why you don't do that.
It's because the tools that are being made, they are fallible,
and sometimes to a point that is disturbing. I told
you this OZ, and there's an episode of tech stuff.
(29:08):
I think it's called something like. AI wrote this episode
sort of and in that I had it write an
episode of tech stuff. All I gave it was very
simple instructions. I used chat GPT and I said, write
an episode of tech stuff about the technology of airbags.
That was it, like. I didn't give it any further instruction,
and as part of what it regurgitated to me, it
(29:31):
gave me statements from three supposed experts, but none of
those people existed, which means automatically you cannot trust the information.
And when I would ask for things like I followed
up I said, could you give me a source for
this information? And it wouldn't be able to. So these
things are going to be things that improve over time.
(29:51):
But I think the problems we're seeing now concern me
because whether the technology actually gets better, so it's being
more accountable, or it just gets better at obfus skating
when it's making stuff up. I don't know that we'll
be able to tell the difference.
Speaker 1 (30:07):
Yeah, I mean, Latour was talking about how basically, how
do we at Dow Jones use these tools to make
sure we had compensated for our work so that we
can make more money so that we can fund more journalism.
Speaker 3 (30:20):
It's a very pragmatic sort of look, but it makes sense, like, yeah,
the good balancing art and commerce is always a struggle,
even no matter what outlet you're looking at.
Speaker 1 (30:31):
It's interesting as well, though, because it's not He wasn't
just talking in abstraction, Like right now, Dow Jones is
in partnership with open Ai and suing Perplexity.
Speaker 3 (30:41):
Wow.
Speaker 1 (30:41):
So with open Ai they made this deal last year
that The Wall Street Journal reported on its parent company
could be worth more than two hundred and fifteen million
dollars over the next five years in both cash and
credits for open ai technology.
Speaker 2 (30:56):
But last year Dow.
Speaker 1 (30:57):
Jones also sued Perplexity the AI search engine, and the
suit's not public and al Maesa didn't want to go
into details about what obviously, despite the best efforts of
Sarah Fisher from Axios on stage, but why it did
report on the lawsuit and they basically, you know, the
headline was that dow Jones alleges Perplexity is hallucinating fake
(31:17):
news and attributing it to real papers and that's illegal.
Speaker 3 (31:21):
Oh yeah, so very similar to what I was just saying,
exact right, Yeah, this is this is like someone writing
a term paper and having a citation for a source
that doesn't exist.
Speaker 1 (31:31):
It's like somebody making something up and then citing it
to a professor when the professor never said.
Speaker 3 (31:35):
It, right, right, right, right, And yeah.
Speaker 2 (31:37):
That's what that's what this suit is about.
Speaker 1 (31:39):
Absolutely, yea, references in a way that's even more egregious
to your point about AI getting better at covering its
tracks that it makes up up when you're doing fake citations,
that's pretty.
Speaker 3 (31:49):
I mean, it's it's terrifying because like obviously this is
this ends up becoming like a political term in some cases,
but like the whole fake news thing which often was
being used to try and delegitimize real news sources that
were just saying things you didn't like, Yea, you call
it fake news. So that you can dismiss it. But
we're actually talking about actual fake news, the news about
(32:12):
stuff that didn't happen or at least didn't happen the
way that you're being told it did, but it's being
put down in a record as if that, in fact
is what happened. How do you then get to the
point where you can determine reality from computational fiction.
Speaker 1 (32:30):
Yeah, yeah, I mean the litigation stuff is very interesting.
Obviously if you look back to the Victorian times and
the aftermath of the First Industrial Revolution, that was when
copywriting laws took effect. And so you know, as the
New York Times law suit against open Ai potentially comes
to trial, as Dow Jones lawsuit against Perplexity comes to trial, potentially,
(32:51):
we're going to see new president and it'll be really
interesting to see how that shakes out and how that
shapes our future. I mean, the power of law to
shape reality is so huge, and copyright law has been
such a fundamental of our society and who we are
for so long.
Speaker 3 (33:06):
Yeah, absolutely fascinating. It Also it goes back to that
old saying. It's almost a cliche, but you know, necessity
is the mother of invention and in this case, necessity
is the mother of the need for new laws. So
with copyright, before you get to things like mass printing,
copyright really wasn't that big of a concern, just because
(33:27):
it was such a pain in the butt to make
a copy of a work. But then you get to
a point where technology is capable of giving people opportunities
to do things on a much larger scale. That's where
you start to see the need for new law. And
I suspect we're going to see a lot of changes
to law in the next few years as a result
(33:48):
of the rise of technologies like AI.
Speaker 2 (33:51):
Yeah, I also thought it was interesting.
Speaker 1 (33:52):
I mean, obviously, dow Jones is, you know, a Murdock company,
so perhaps no surprise that they're pugnacious bunch.
Speaker 3 (34:02):
Yes that's pretty good, kindly, but you.
Speaker 1 (34:06):
Know, but the point that he made was basically, when
the Internet emerged, you know, twenty five years ago ish
as a mass adoption technology, the news organizations, the music
rights holders, you know, film and TV industry essentially allowed
the consumer expectation to develop the content was free, yes,
and that essentially hollowed out those industries to to you know,
(34:30):
to the extent we're seeing today where all of them
are really struggling. So his point was, how do we
make sure we we'ret a new technological inflection point that
we don't roll over again? And you know, so I
thought that was pretty interesting.
Speaker 3 (34:41):
You know, that's an incredibly good point. I mean, we
are where we are today in part because we've all
had the expectation of the content that we want to
access at any given moment and any point of the
day should be easy and free to access. That's a
great eight thing in many ways for the individual, but
(35:02):
it does put an incredible burden on the entities that
are actually creating that, whether you are a solo content
creator and you're just trying to make your passion your occupation,
or you're a big media company and you're trying to
put out really incredible content, which costs a lot of
money to do, and then you're getting tiny little pennies
(35:24):
in return. It's really hard in today's world to have
the same kind of media output that we would expect
from say the early two thousands and have it be
a profitable business. And if there is no profit, there
is no business. If there's no business, there's no content.
So ultimately it does come back down to us to say,
(35:44):
how do we create a world where this technology can
be practical and useful, not dangerous, and not something that
bankrupts any company that actually gets into the business of
doing it.
Speaker 1 (35:59):
And that was the other thing Alma was talking about
in terms of some of the positive use of AI.
Speaker 2 (36:04):
He talked about.
Speaker 1 (36:05):
Helping Dow Jones do better research. He talked about scaling
the output across multiple languages. Apparently Dow Jones wire service
is now available in Korean, largely translated by AI at
such a good enough level that Korean traders are able
to trade off it and also using I to make
their content more interactive.
Speaker 2 (36:24):
How do you feel about some of these uses of AI.
Speaker 3 (36:26):
I think that that's incredible. Like I think the idea
of using AI to improve accessibility to me is a
no brainer. Use for AI, making something that has got
value and usefulness to someone's life accessible to them is transformational.
It doesn't do me any good If the information I
(36:47):
need is in a format that I cannot access, it
might as well not exist. So I look at how
transformative the Internet has been over the last couple of decades.
I think I'm old enough where I remember when I
first saw the web like which I famously dismissed at
the time because it was so slow. But when the
(37:07):
web started to really take on properties that were undeniably
useful and transformative, I just took it for granted that
I was able to access all of it because it
largely came out of the United States. Now, if I
had lived anywhere else where I didn't know English, knowing
that that existed might have been interesting to me, but
(37:28):
it wouldn't be practical. I wouldn't be able to do
anything with it. So using AI to do things like translate,
whether it's a podcast or written piece of work or
a movie, that to me is one of the best uses,
the most incredible uses of AI. And obviously it's not perfect,
but it's really good. The issues we tend to run
(37:49):
into are things like idioms and cultural references that are
not easily translatable from one language to another. Those are
problems that are going to just remain for quite a while.
I don't know how long it's going to take us
to teach robots what these idioms mean. That's going to
be an issue that's ongoing. But I think that's a
very noble use of AI.
Speaker 1 (38:12):
Thank you so much, Jonathan. Thanks for joining today in Doha,
and I hope we'll do it again soon.
Speaker 3 (38:17):
Absolutely, I'll head up to New York if it's a
much shorter plane.
Speaker 2 (38:21):
Ride, that's for sure.
Speaker 1 (38:24):
That's it for this week for Tech Stuff, I'm mos Voloshin.
This episode was produced by Eliza Dennis. It was executive
produced by me Karen Price, and Kate Osborne for Kaleidoscope
and Katrina Norvelle for iHeart Podcasts.
Speaker 2 (38:36):
Please rate, review, and reach out
Speaker 1 (38:38):
To us at tech Stuff podcast at gmail dot com.
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