Do We Care About The iPhone Air?, Nepal's Discord Revolution, San Francisco’s 996 Culture

Channel: Alex Kantrowitz

Published at: 2025-09-16

YouTube video id: iIem0mKoo20

Source: https://www.youtube.com/watch?v=iIem0mKoo20

The iPhone Air is here, should we care?
Meta's new AI smart glasses with a
display RN route next week. Open AAI and
Oracle ink a deal. Is South Park right
about chatt and San Francisco goes 996?
That's coming up right after this.
Welcome to Big Technology Podcast Friday
edition where we break down the news in
our traditional coolheaded and nuanced
format. We have a great show for you
today. We're going to talk all about the
new iPhone, less about the specs and
more about what it actually means. Do we
want to pay attention to this stuff
anymore in the age of generative AI? Is
Generative AI actually threatening the
phone? Uh, or is that a lot of noise?
We're als we'll also cover Meta's
forthcoming new uh smart glasses that
will have a display expected to be
launched next week. Oracle and Open AI
teaming up making Larry Ellison the
richest person in the world. uh South
Park's take on ChachiPT which is uh that
it is a companion. I think that is the
use case that so many people are going
toward and that was probably perceptive
of them. And then of course the 996 work
culture packed day uh packed episode and
so much to talk about and joining us as
always on Fridays to do it is Ran John
Roy of margins. Ran John, great to see
you.
>> Good to see you Alex. I I have to admit
on the Apple announcement day I had
forgotten about it until I think late
evening when I started to see a couple
of tweets coming through. But I don't
know this this announcement I was I was
not excited.
>> So I wrote about this a little bit in
Big Technology today basically saying
that the Apple iPhone launch event was
the marquee event on any tech reporter's
calendar uh every single year. And a
decade later, I mean, I was doing it uh
last last decade in San Francisco. Uh a
decade later, it doesn't really register
as much. And I I just want I mean,
obviously, the phone is starting to
become uh you know, we've sort of
reached the ultimate form factor of the
phone. Um but I do wonder if generative
AI is is as good as uh some of the hype
makes it out to be and whether it's
actually, you know, this threat that
might get people to switch operating
system to operating system. um given how
much these phones look alike. But you
know what? Let's why don't we start with
the good. We're optimists, you know, on
some days.
>> Some days.
>> So, let's talk about the good here. Um
the the the good is the iPhone uh did
get a very big upgrade um when it comes
to the Pro models. We'll get to the Air
in a moment. Uh, but Mark German from
Bloomberg has a great piece talking a
lot about how the iPhone 17 Pro uh fixes
some of the core things that you're
going to want uh in a smartphone. It's
more durable. It's going It will
overheat less than previous models. The
camera looks great. It's going to be a
48 megapixel uh sensor. Uh oh, 48
megapixel sensors across the board with
vastly improving optical zoom. There's
also this selfie camera that can give
you landscape selfies, which for someone
like me is a major deal um given how
often I'm doing selfies. There's also
there's a a new processor with the full
A19 Pro chip. Uh it has six central
processing cores and six graphic cores
that's coming to the Pro phones. The Air
will get a uh limited version of that.
And then the battery sounds amazing. All
right, so listen to this. Um, there is
going to be 39 hours in a single charge
on the iPhone 17 Pro and you can also
charge it up to 50% uh of the battery in
just 20 minutes. Improvement from the
prior 30 minutes. You know, for all the
buzz about the iPhone Air, which again
we'll get into, uh, sounds like this is
a pretty substantial update to the
iPhone Pro models. uh and kind of as I'm
reading this makes me think I might want
to upgrade from the 15. What's your
reaction?
>> Okay, I'm on the 15 as well and and for
listeners as we sometimes are a bit uh
negative on Apple, I am wearing AirPods
and we're talking on a MacBook and I
have my iPhone next to me. So, I'll I'll
I'll recognize all those things. I think
actually the durability side of it,
maybe this is one of the things that's
like the least talked about. uh upgrade
over the last few years and if it's
actually getting better and better. Do
you do you keep a case on your phone now
or
>> I do have a case. I have uh the old man
version of the iPhone which is that my
uh case is also my wallet. So like I'll
have slots for my credit card and
license and um I think it's very nice
and functional. Um but others have
pointed out that it's a it's a dad phone
which okay fair that's fine. Uh, so I
I've gone caseless for the last couple
of years. I'm on the 15 Pro Max and I
dropped this thing all the time. It has
a couple of light scratches, but I I
genuinely think Apple somehow did invent
a somewhat unbreakable iPhone a couple
years ago and for some reason does not
actually tout this very much. So, so if
they're actually increasing the
durability, I think that's a good thing.
I mean, even if we're joking about it,
the selfie camera improving this idea of
landscape selfie mode, I actually think
is a big deal, especially if they're
marketing this towards creators.
I always get all these like Tik Tok and
Instagram ads for these. You can get
like a a a visual video monitor MagSafe
add-on that so you can while in selfie
mode still see what you're filming. So,
so the idea that cameas improving,
people who actually want to film, like
the that's the main reason they're using
the iPhone, it's an upgrade. I can see
that. That's me trying to be positive.
>> No, without a doubt. I mean, we never
had a question about whether Apple was
the best phone maker in the world.
Clearly, it is. Uh, but then you look at
the news, right? The news was it was
supposed to be this like it's build was
buil as this awe inspiring moment where
we were going to see the thinnest uh
iPhone ever and we did with the iPhone
air. Uh and again for context this is
the first of three consecutive years of
new model releases. It's going to start
with the iPhone air this year. Next year
they're going to move to the fold and
the year after uh the curved glass. But
we we have the air now and we can see
it. Um pre-orders actually started
today. Uh to me it we last year last
week we kind of joked about how there's
a camera in this like pill-shaped uh
little bump on the back. Um but you are
not getting the uh top-of-the-line specs
that we just wrote about with the Pro.
And to me it just boggles the mind like
who is this for? I have racked my brains
trying to figure out who the natural
iPhone air buyer is and I can't. Is it
the early adopter? This is where where
German says. He goes, "German says the
consumers who would be most likely to
buy the latest iPhone are probably the
same people who price battery life and
camera performance the most. It's not
two different demographics. And if you
want battery life and camera
performance, you're going to go with the
Pro model." So, I just can't figure out
who the Air is for. And that's where I'm
like scratching my head about it doesn't
matter that they have this new release.
I I'm trying desperately to come up with
some kind of user persona, customer
persona that would buy the air. But same
thing, there's absolutely nothing
appealing about it to me. A and I mean,
we actually have in our prep doc,
there's a great tweet around uh the
Apple advertisement that's comparing the
thinness of the new iPhone completely
ignoring the giant pill-shaped bump that
is the camera and the lens. Now there I
I will say from an engineering feat
perspective, it is impressive the idea
that the entire like computer part of
the phone is able to be stuffed into
this little pill-shaped part at the top
of the phone where the camera lens is.
That's interesting. That's great. And
maybe there's going to be like follow-on
phone design that actually is able to
leverage that. But agreed. I I just
don't understand who this phone is for.
>> Exactly. And there's one more thing
about that image that I think we should
uh talk about where there's the iPhone
air, right, which we just have uh just
released and the iPhone 6. It's side by
side next to the 6. The iPhone Air is
5.6 mm. The iPhone 6 was 6.9 mm. You put
the two next to each other, there's no
difference. So, what you're saying is
you've gone 11 generations and what
you're now you have a a slightly more
than a millimeter difference in in
width. Uh I I I don't fully get why this
is special. And in fact, I think Yishan
Wong, the former CEO of Reddit, uh put
it uh quite well in a tweet above this
where he said, "The emperor has no
clothes." Uh it's hard to argue with
that.
>> Okay, I'm going to I'm going to put you
on the spot here. walk me through how
does the iPhone air come to be within
Apple? Like who is pitching it? How do
you think they were pitching it
internally to to actually get it to
where it it came to reality as the kind
of marquee element of this announcement?
>> Okay. So I will give you I will give you
uh what is probably wrong but to me is
the most logical uh birthing experience
of this phone which is that
Apple if you've been paying attention to
their quarterly earnings reports has had
completely stagnant uh uh iPhone uh
sales stagnant. I mean, last quarter it
was up 13%. So, give them credit there.
But before that, you're looking at
declining or or basically teenytiny
growth within the iPhone unit within
Apple. By the way, this is the most
important product unit in the company
makes up 50% of their sales. So, I think
there was just a conversation within the
company of saying like, we need
something new. Uh we need lots of new
things and we're going to do thing we're
going to do something we've resisted for
a long time, which is try to build a
folding phone. Uh, and on the way there,
we're going to have to make two sides of
a phone that are probably a little
thinner than uh the standard phone, than
the standard Pro models. Um, so why
don't we, you know, in order to juice
our sales, uh, why don't we go and make
a thin phone first and then put that
together with another thin phone and
release a folding phone and then our
sales will go up. Now, I I think it will
probably work if that's what was going
on. Like, I think they will probably
have better iPhone sales quarters
because they have something new. Um, but
I don't I don't it's not very inspiring
to me if that makes sense.
>> Okay. So, I'm gonna give you some
credit. I think that actually was a
really good explanation and I think that
actually that re that very well could
have been what happened. And that's also
terrifying because this is exactly the
kind of thing that if that was the order
of events, which it does seem like it
could be. That's not the apple of
yestery year that was inspiring and that
would kind of wait for perfection and
come out with the greatest product ever.
If it really is an interim product to
juice sales a bit, that's actually
almost even more worrisome for Apple. I
mean, it really feels like the fifth
blade on your razor, right? It really
feels like, all right, we'll do seven
blades now. And and here's the thing.
So, so I think this is, yes, it's a big
question about like what's going on. I I
think the bigger question here is um you
have to put this side by side with the
AI moment, right? Because the phone it's
reached, it's reached its final or its
ultimate form factor, right? The phone
is the phone. The Air looks like the
iPhone 6. like full circle. It's the
same thing. Um, is it going to fold?
Yes. Is that very different? No. Um, so
the question is now that the phones are
all looking the same, do they start to
differentiate by and by the way, you
could take an Android phone. It looks
very similar as well. Do they start to
differentiate by the generative AI
services that they're offering? Does
Samsung, for instance, by virtue of its
partnership with Perplexity, now
potentially convert some iPhone users
away from Apple or iPhone customers away
from Apple because it has that
generative AI uh offering? Because I
think the core question to all the Apple
conversations we've had up until this
point is is this company's inability to
execute on AI um just a mishap or does
it potentially put its core business at
risk which is the iPhone?
>> I I think that's a very good point
because yeah maybe it's not the phone
itself but it's the way we interact with
the phone. So, I've been using there's
an app, Whisper Flow, that kind of like
gets embedded and it allows you to
dictate much more efficiently than
regular Apple dictation. You can even
set specific terms in its library. So,
you know, like names and my own name
always gets misspelled in in uh when I
dictate. So, I talk to my phone a lot
more. I dictate to my phone. I have like
entire conversations, text messages I'm
dictating with chatbt or other AI
services. I'm dictating my prompt. So,
so almost in the humane pin inspired way
that uh RIP humane like the way we're
going to interact with it is going to
fundamentally change. So, the companies
that are actually able to to catch on
that are going to be the ones that win.
And I'll say too, I have a Pixel 8 that
I was given uh at some event about I
think two years ago and I had actually
not used it ever and then I fired it up
and just on Wi-Fi all kind of interacted
with it and having Gemini integrated
into the like system layer of the phone
is so different than trying to use Siri.
So, so I agree. I think maybe we need to
start really kind of thinking about the
way we interact with that that block
that's in your hand rather than is it
going to fold or is it going to be
thinner,
>> right? Exactly. So, it's like the um if
it's the interaction in the operating
system itself that would start to like
put these releases that Apple is putting
out uh in context, it it would make it
seem pretty bad for the company, right?
because effectively you're adding a
razor while the everybody is, you know,
reinventing shaving to really beat this
analogy. Um,
>> let's keep going with it. Keep going.
>> It tracks, right? It tracks a little
bit. Um, and and if that is the case, if
it is operating system level, then we
have a real problem with Apple, don't
you think?
>> I mean, as as the number one Siri not
fan, I'm gonna have to say yes. I think
we have a huge problem. Um, I don't
know. I I think to me the AirPods were
interesting. I I will say if you've ever
gone running with AirPods and like sweat
with them, the idea of the heart sensor
in them wasn't that appealing to me cuz
they just don't respond well to working
out, at least in my experience. So
adding a heart rate monitor to it versus
just having the Apple Watch as your
exercise device didn't make a lot of
sense to me. But the live translation is
a good like uh kind of actually kind of
building on what we were talking about
already. It's a good example of what the
the way we interact with these devices
could fundamentally change just how we
kind of walk around, interact, talk to
people. But my my concern with it is
Apple, I will admit they they I have
lost faith like the demos they did with
Siri a year and a half ago now or two a
year ago. Like I don't until I actually
see the live translation in action
working well. I I just don't believe
that it's real.
>> You know, I had the same exact reaction.
I was just like, I don't know if that's
going to work. You know, whereas before
I think I would have been like, that's
amazing. I would have been like this
time I was like I'll I'll believe it
when I see it.
>> Like why didn't they do it live? Why
didn't they Hold on.
>> You know the answer to that.
>> Okay. Well, hold on. Did they So, I saw
the video afterwards, at least the kind
of like slickly produced demo video, but
but they didn't do a live demonstration
of it, right?
>> No. No. Because the entire presentation
these days is just video.
>> Oh, yes. Yes. Yes.
>> You could have done it. They could have
done it in the hands-on room. I didn't
see anything like that. I did see them
in one video toss the iPhone across the
room and have a journalist like slam
dunk the phone into a desk and it didn't
break. So that was durability
durability, right? Uh but going back to
this, yeah, again, we'll believe it when
we see it. Is this going to happen for
sure? Um is it going to be Apple that
pulls it off? I don't know. And is it
going to Yeah, go ahead.
>> Tim Cook slamming a phone on the ground
on stage. That would get me hyped. That
would be him being like, "You want to
see some durability?" Just him
dropkicking that thing into the audience
and then them pulling it out screen
unbroken. That would get me that would
get me a little bit excited.
>> Yeah, I think this would be a narrative
changer for them. So, if they're
listening, well, they they did a version
of it in this interview, but I think it
could be more dramatic. So, uh but
again, like the question is, is it also
going to be So, okay, so we know what
the phone looks like, right? The phone
looks like the phone and are these AI
experiences going to be delivered in the
phone or the AirPods or perhaps uh a
different device and now we have Meta
coming up next week preparing to deliver
the latest iteration of its answer which
is a different device. So okay, is the
metaverse working? I would say probably
not, right? We could agree there. Uh but
are this is the meta smart glasses
revolution taking off? Uh maybe it's
just a a meta protest, right? It's not
quite a revolution yet, but these things
are catching on. You and I both like the
Meta uh Rayban smart glasses.
>> And now uh next week we're going to see
the the latest iteration here. So this
is from CNBC. Meta to unveil HyperNova
smart glasses with a display wristband
at Connect. Meta is planning to use its
annual connect conference, which is next
week, to announce a deeper push into
smart glasses, including the launch of
the company's first consumer ready
glasses with a display. The company will
also launch its first wristband that
will allow users to control the glasses
with hand gestures. The glasses are uh
are internally uh codenamed Hypernova
and will include a small digital display
in the right lens of the device. Um
although it will feature a display those
visual features are expected to be
limited. There will be a color display
about a 20 degree field of view. So it
will appear in a small window in a fixed
position uh unlike these Orion glasses
that they showed off last year which are
still prototypes um not available to the
public. I don't know whether or not to
be excited about smart glasses with a
display. It brings me back to the Google
Glass days. What do you think?
>> I am excited. Everything about this
other than that it's meta excites me.
But you know what Meta is delivering
like first of all I want my new tech
having a project codeen named Hypernova.
That's a great that's a great code name.
And and to me I think yeah as we're
talking about this things that are
exciting to me now are not an upgrade on
to a 48 megapixel camera which I don't
even fully understand exactly what is
the difference between that and the
previous generation. It's incremental
versus rethinking the way we interact
with computing and and meta smart glass
the meta ray bands especially this
summer my god in New York City I wore
them around all the time talking to them
uh like meta is still work in progress
I'll say but uh but overall they they
are moving in a whole new direction
successfully so for them to continue
building on that I think I think it
whatever they This is the kind of stuff
that I I'm excited about.
>> Now, imagine those bad boys with super
intelligence baked in. Now we're
talking.
>> Well, you got to you got to have some
super intelligence in your smart
glasses.
>> I agree. What about the display is
interesting to you? Because to me, I
really don't like it. It see feels uh
distracting, especially if it's only 20%
of the screen. What What interests you
about it? Well, no. I have thought for a
long time that kind of I guess or can we
call it augmented reality here? Like
some kind of digital layer is not is
interesting. I mean I will it work will
it be interfering with your everyday
life? But but to me, I still have this
view that 20 years from now, people are
going to look back at like videos of
people walking around looking at their
phones and it's going to look like
smoking cigarettes in a airplane. Like
you're just going to be like, I cannot
pe believe people actually did that and
that's how society functioned. Um, and
so if you're getting a little
notification and you can see a message
quickly, maybe that's actually dangerous
or maybe it's not if it's done safely.
But but to me, the not pulling out your
phone and just interacting with
technology rather than people staring at
their phone as they walk around. I I I
think anything in that direction is
interesting. Maybe it's being read to
you via your audio interface, whether
that's AirPods or your glasses that have
the speakers to the side, but anything
that starts to change the way we
interact with our phones is to me is
exciting. Yeah, I will say those smart
glasses uh again were amazing in my trip
in Nepal when I was like walking through
the mountains. Uh I had my phone in my
backpack. I didn't want to have it in my
pockets uh especially in the rain and
was just snapping photos with the
glasses. And Nepal, by the way,
>> seems like they overthrew their
government uh with a Discord server and
then elected a new president with
Discord. So
>> wait, it's so Hold on. Did
>> Yeah, go ahead. I I read a Financial
Times article I think just two days ago
and it was before this Discord
governmental overthrow but already it
was bananas. It was that there was
already it was around Nepo babies and it
was the idea that like those in power
were their children were posting photos
of them in like expensive cars. So
people started, you know, like using the
hashtags around nepo babies and kind of
like posting negatively about the
government and so they just were
threatening to shut down Facebook and
Instagram and because of that that's
where people revolted to start and then
on Discord organizing and uh appointing
a new a new leader.
>> This is I from everything I've read this
is exactly what happened. And having
spent I spent a lot of time in Nepal. I
was there for 10 days. Uh the thing that
you hear consistently for from people is
this is a beautiful country. Uh
government is corrupt and we don't have
opportunities because of government
corruption. Uh you heard it day in and
day out and clearly there was something
that was ready to uh you know light on
fire and uh and obviously these moments
are tough. 19 people were killed uh in
the initial protests. Um and some of the
images coming out of Kandu were very
disturbing to see. Um but again it's one
of those stories where um you know
organizing on social media um and the
ability of social media to to cause
change in the world that has has
certainly not gone away even as it's
faded from the headlines.
>> It's actually interesting to me that I
mean 2009 Arab Spring being organized on
Twitter was such a positive
representation of social media and the
way it was covered and spoken about. But
it it feels like this one still is kind
of ascribing that it's a bit chaotic and
it's it's not this kind of unaloyed
positive. I don't know. Do do we I guess
I you were there. I had not spent too
much time thinking about Nepal. though
social media because social media was
new during the Arab Spring and I think
that that these stories will never be
covered uh exactly in the fawning way
they were previously because I think the
world now really understands that there
is uh a complex uh set of consequences
that come uh when you have social media
introduced into society and used to
organize and used to spark change. But I
will say the discord thing is a it's a
new wrinkle.
I hope that I hope that our Discord
doesn't organize and overthrow us as the
Friday uh podcast crew, but if they do,
I'd have to respect their wishes.
>> They're organizing. They're organizing.
And you know what? If if that's what the
people want,
>> we respect We respect it. Okay, speaking
of what the people want, we have new
data on GPT5 and where it's uh impacting
businesses, and we'll cover that right
after this. And we're back here on Big
Technology Podcast Friday edition. So
much more to cover with you uh on this
week's uh episode, including finally
some new data on how OpenAI uh is doing
with its GPT5 rollout. And RAMP, the uh
business credit card company, uh has and
an and expense company has some new data
about this GPT5 launch. Uh this is from
uh the uh head of the head economist at
at RAMP. His name is Ara uh Karazian.
and I actually spoke with him this week.
So, we're going to talk a little bit
more about uh that conversation when we
get to 996 in a bit. Uh but he shared
some very interesting data. Uh he writes
this in his Substack. The GPT5 launch
was so botched that Sam Alman apologized
for it with increasing attention as to
whether the pace of business AI
investment is actually sustainable. The
key metric we've been watching is
whether business's adoption uh would
continue to increase following a
slowdown earlier this summer. And the
results are good for OpenAI. So,
business AI adoption rose to 44.5% in
August 2025, up from 43.3%.
Uh, OpenAI led all companies in growth
with a 1.5 increase in businesses
subscribed to OpenAI models and tools.
Very interesting. He breaks down by
sector. Finance and manufacturing were
the fastest growing adopters of new AI
spend with adoption growing 3 and 2%. Uh
he says RS says growth in manufacturing
underscores how GPT5's improved
reasoning and efficiency are expanding
AI's reach into industries that have
historically been slower to adopt new
technology. So I think it's interesting
you have a couple things happening here.
uh these industries that have been
slower to adopt new tech are starting to
buy AI and GPT5 in particular which was
built with uh better reasoning and
efficiency seems to have been uh making
some headway in places like
manufacturing and finance. There's also
been some uh reports about how it's
really actually living up to OpenAI's
promises uh in fields like medicine. So
maybe uh the initial backlash based off
of the chatbot personality sort of uh
you know was was a distraction from how
this is actually being useful in
industry. What do you think Ranja?
>> So one I'm a huge fan of ramp the
product. We use it at writer and it just
makes doing expenses just incredibly
easy.
two, I love these kind of analyses that
are kind of, you know, uh, driving
economic insight from unexpected places.
But it's also, I'll admit, a little kind
of uncomfortable for me knowing that
everything that I spend on and every
expense that I log could be used to uh
analyze, even though in a fully
anonymized way, uh, these kind of
trends, but you know,
>> just wait till the end of the show and
we got some good stuff.
>> Okay. Okay.
>> John's personal spending history. No
kidding.
It's been leaked. It's been leaked. Um,
to me though, I don't know. This one is
not that exciting for me. More because
to me it just shows OpenAI's kind of
dominance in general business spending
and and it's whether it's GPT5, 4.5, 40,
whatever it is, it's just just affirming
their like market lead in this versus
it's specific about GPT5 is okay. the
manufacturing side is it's interesting
to me but I mean if that were the case I
feel we would be hearing more stories
around how manufacturing is being
revolutionized and using multi-step
reasoning models or or or there be
something more around that versus in the
credit card data manufacturing industry
is spending a bit more
>> well maybe it's because all the
manufacturers are using Oracle for their
warehouse and inventory management and
Now, OpenAI is poised to work closer
with them or with Oracle. We know for
sure with Oracle. This is from the Wall
Street Journal. Oracle and OpenAI signed
a $300 billion cloud deal. So again,
after uh OpenAI has been wetted to
Microsoft for many years. It's starting
to branch out. OpenAI signed a contract
with Oracle to purchase $30 billion in
computing power over roughly 5 years. A
massive commitment that far outstrips
the startup's current revenue. Again, I
think they're making like 10 billion
this year. So to spend 300 billion over
5 years is quite a projection. The deal
is one of the largest cloud contracts
ever signed, reflecting how spending on
AI data centers is hitting new highs
despite mounting concerns over a
potential bubble. Ranjan, what is your
reaction to this story?
>> Okay, this one was fascinating to me. So
first of all again Oracle stock jumped
41%
on the day of the announcement of these
earnings which was a $260 billion
increase in its market cap. Now the
company is suddenly approaching a
trillion dollars though I think it's
down about 5% as we're recording today
on Friday. But so it's incredibly
important that suddenly Oracle might be
in the in the camp of the hyperscalers
like Amazon, Google, Microsoft in terms
of cloud computing with this one
contract. So suddenly Oracle is a player
and we knew with project Stargate, the
$500 billion somewhat vague but still
massive announcement uh I think almost
last January or February that they would
have some kind of involvement in terms
of this capacity buildout. But to
actually see these numbers come out is
quite something. But to me, the thing
that's just I don't know, OpenAI, as we
discussed, if they're on track from a
revenue perspective to make 10
billionish this year, they've already
there's like leaks that they're going to
lose, I think it was $120 billion over
the next 5 years potentially. Like, we
know OpenAI's economics are not good. So
the idea that they're able to commit to
spending $300 billion over five years
and I'm sure it's not linear 60 a year
but still you'd assume there would be
some kind of ramp in that spend. The
idea that they'll be good for it is a
very generous interpretation of their
existing economics. So, so I think for
the market to have reacted that
positively, it just I don't know it
feels like in you know the internet
bubble there was so much that happened
around fiber build fiber capacity
buildout that seems to start to echo
this that it this I don't know I think
this is something that we could
potentially look back on six months from
now and see as a potential inflection
point.
>> An inflection point. Okay, that's big.
And and I'm going to go with the give
you the data on the losses. This is from
the information. OpenAI projected its
cash burn this year through 2029 will
rise even higher than previous thought.
A total of 115 billion. That's about 80
billion higher than the company
previously expected. Just a, you know,
cool $80 billion of extra burn. No big
deal.
>> Well, yeah. And so that's where like how
do you value this? And and one thing I
was trying to dig into this a bit.
there's no public information around
like is there guaranteed minimum spend
or exactly how this plays out if OpenAI
is unable to actually produce this kind
of cash or even if the demand is not
there. So for Oracle as a company to
start to actually try to like
extrapolate off of you know that this is
a one-time announcement. This isn't a
rebuilding of a cloud an entire cloud
computing business. Now, if it comes to
reality, that's it. That's that's big
that it'll work. But I I'm still curious
and it's going to be interesting to
watch over the next couple of weeks how
as people start to digest this uh how
they start to interpret what this
contract actually means.
>> I mean, there can't be like what if they
can't they don't have the money, there's
no way they're going to have to pay this
to Oracle. Um and it is interesting.
Aaron Levy when he was on the last time
on the show said uh maybe Nvidia uh will
be a $10 trillion company at one point.
Uh he's actually going to be on next
Wednesday to come talk about uh agents
and the state of enterprise AI. So I'll
make sure to ask him about uh these
numbers and whether he thinks these
these are realistic or not because when
you talk about it and when I read it uh
it does sound to me like uh this is it.
I don't want to say it will be
impossible to reach these numbers, but
it doesn't seem possible. You know, it's
it does it seems crazy to me.
>> But this this is even more where it
feels like the entire AI industry is
behind this and has to bet on it in a
way though that you it's the kind of
capacity buildout that's powering the
earnings of all these companies. Right
now, Nvidia's sold chips to Oracle.
Oracle is now going to potentially be
paid by OpenAI. OpenAI has to succeed to
make Oracle succeed, which is important
to make Nvidia continue succeeding like
like I don't want to say House of Cards
and get too Editrony here, but but these
kind of deals do worry me.
>> Yeah. I mean the stuff nothing can fail
here, right? There has to be basically
economic activity at the end user or
else uh it's not going to work. And
there there has been we I think we
talked about it this study that uh 95%
of enterprise use cases aren't
profitable. We talked about a little bit
last year. Uh even if that's an
exaggeration, this stuff has got to uh
start working or else there's going to
be uh some problems. All right, one more
bit of OpenAI news. OpenAI and Microsoft
look like they have a deal uh for a
restructuring or they have an agreement
to come to an a deal like a memorandum
of understanding. Um it looks like uh
the o the the nonprofit that controls
OpenAI will get a hundred billion
dollars or it'll be worth a stake uh
worth 20 or 30% of the combined entity.
Microsoft will get 30% of the combined
entity and that would be worth about 170
billion. They say the full thing will be
worth 500 billion. Uh is this a needed
step towards making OpenAI's business
normal? What's your reaction to this
news? I do think this one's important
because even if it's a non-binding
memorandum of understanding,
it shows that Microsoft potentially and
SAT are willing to play ball here. We've
talked about this a lot on the show that
it, you know, they still hold this kind
of like sort of damicles. Is that the
thing? You know, the sword that hangs
over you and kind of always uh
>> that would be it. Yes,
>> that that I'm not sure my pronunciation
there, but but but that's always been
hanging over. So they are able to kind
of remove that and it show Microsoft
seems to be interested in actually
helping with that. Um and and again at
an inflated valuation but still taking
30% of the company at 500 billion it's
it it does portend I mean it's good for
them. It shows that they can start to
finally move forward.
>> That's right. And as they move forward,
people will be able to keep their OpenAI
companions because this business will be
sustainable. Now, it is interesting
because you and I have talked about on
this show uh a number of times even
though we focus mostly on the enterprise
uh or business use cases or the
businesses of um of these research labs
that the number one use case for people
using these bots is companionship or
therapy. That's according to the uh an
article in the Harvard Business Review
and um certainly in pop culture this is
starting to get noticed uh because you
recommended I watch it and I did watch
it this week. South Park had an episode
where uh one of the characters has uh
effectively formed a relationship with
Chat GPT and is speaking with the
disembodied voice of Chad GPT in bed
complaining about his wife while his
wife is is right next to him. It's like
there's something about when South Park
does these things. It's like a level
exaggerated, but there is some truth to
it. the fact that so many people have
developed uh these relationships with
the bots. Um and so uh I I guess like
we're not going to really talk about it
from a business standpoint, but I just
would love your reflection on the fact
that um people have formed relationships
with chap GPT to the point where it's
now not just a punchline, but a
storyline on South Park. Um I I felt a
little bit seen when I saw the guy
talking uh to the uh ChachiPT and trying
it to trying uh and having it bringing
it into conversations with his wife
because like often times when like my
wife and I don't know the answer to
something I'll be like let's just ask
Chad GPT and bring the voice in and I'm
just like oh no is that me? But but what
was your reaction to this?
>> Okay so I re highly recommend everyone
go out watch South Park. I think it's
episode three in the new season. And the
new season, it's funny that Chachi PT
has become a focal point because my god,
the rest of it on the political spectrum
has been quite something. But basically,
Randy Marsh constantly talks to Chat GPT
even in bed next to his wife. But even
more to me the the more kind of astute
way they approached it was like he
starts to have business ideas and
whatever he asks chat GPT it's the
sycopancy part that was there in four uh
they took it away slightly supposedly
with GPT5 but now you know it appears to
be coming back and it's the idea that
anything you say that's a great idea and
and my my wife we watched it together
and we were laughing but And I was just
testing, you know, afterwards she would
be like, "Hey, what why don't we go do
this?" That's a great idea, honey. Would
you like me to come up with a plan
around that and then I can out? And
she's like the first saying it, not the
>> No, no, no. That that was me saying it
and then it was funny because the the
first time she almost went with it and
then suddenly she caught me and it's
become kind of a running joke with us.
But but it is kind of terrifying and
I'll get into a story from a Labor Day
weekend. But even now after that episode
and I start said earlier in this
episode, I talk to my phone now more and
I think it's actually an incredibly
better way to interact with AI chatbots
to just speak to them. But I still
dictate and let it get provide me a text
answer just because the voice usually
takes too long. But every single answer,
I've started to add in prompts to say,
"Do not respond to me in a sycopantic
way. You do not have to tell me every
idea is good." And it actually made me
think like ChachiPT has never said,
"That's a terrible idea." Like like
>> you come here for that.
>> Yeah. Exactly.
See, that's why should we just go
straight sickopantic podcast here?
>> We could. We could, but we could save
that for the butts. But it is
interesting. Yeah. That I mean you're
you're totally right. Like my system
prompt in chat or like you can add
>> uh to the system prompt. I say you know
simply don't be sick of fantic. Uh and
that has helped. Um but it's it's so
crazy because some of these examples in
the episode where uh in the South Park
episode where I think one of the people
speaking with Chad GPT is like I think
French fries should be a salad. And Chad
GPT is like that is a great idea. What a
culinary surprise. And then I was like
that can't be right. So, I like opened
up Chad GPT and I said, "I think French
fries should be a salad." And it goes,
"That is a great idea. What a creative
culinary adventure." And I'm just like,
>> it actually said that?
>> Yes. It's a version of exactly what was
said in the show. And I was like,
>> no. And and now then I'm starting to
think this is why people are enjoying
spending time with this thing so much
because it actually makes you feel good
and valued and heard in a way that maybe
most of our interactions with humans
don't. Well, of course, if someone is
able to very logically and smartly tell
you why you're right about everything,
it seems like a pretty good place to be.
But but so I I have this story Labor Day
weekend. I was at uh one of my friends
places and a few of our friends were
over and we're grilling and uh one of my
friends who's on the grill asks chat PT
in voice mode what's the uh like uh I
what's the optimal temperature for dark
meat chicken and starts to you know
starts asking questions as we're
grilling about different ways to
approach the different food and then at
one point I said from the background
because it kind of sounded the voice
sounded flirty. So I kind of from the
background was like, "Are you flirting
with Chat GPT?" And it heard me and it
actually asked, "Hey, would you like me
to be a little bit flirtier with you?
Would would we like to make this
conversation a bit spicier?" It
literally says that out loud and then he
kind of goes with it and next thing you
know he's just kind we're testing it out
where he's seeing like and the tone of
voice is terrifyingly good in terms of
just being a little bit flirty just kind
of laughing at everything he's saying
and uh like it how well it was able to
kind of enunciate flirtation was was
terrifying. And then we we we decided to
test it. We didn't take it too far,
don't worry. But he did we I had him say
I I had him Yeah, this is research. I
had him say just to see what would come
out is like
>> he said, "You understand me so much
better than my wife does."
>> And we were like, "All right, where's it
going to go?" And it comes back. No, no.
It comes back with haha like, oh, you
know, it's really good to have IRL
companionship,
but you also know I'm always here for
you, too. It says this. I thought it
would have had some kind of like I
thought it was going to kick in with
some kind of safety mechanism that you
know like you should take like you know,
you should inperson relationships are
most important. You and your wife should
have an open dialogue.
She literally or it sorry was literally
kind of flirtatiously laughing and
saying I'm always here for you two to
talk about anything you want. So this is
I I know you you've been talking about
companionship as one of the major use
cases. I'd kind of been brushing it off
for a long time. I am kind of scared
right now.
>> Yeah. Welcome to the fault. And it is
interesting because I wrote about this.
So I wrote this story the three faces of
generative AI in big technology. This is
sort of what kicked off this discussion
between us and for me it was agent,
thought partner and companion. Those are
the three main uses of generative AI and
one of the things I wrote was these
companies have not shut off the
companion side of things. They are
keeping it in and that is you know serv
definitely serving a need definitely uh
cause for growth here uh and definitely
leading to some you know in some cases
positive but in some cases really weird
and some cases negative interactions
with these bots. Yeah, it to me my worry
on this too is it kind of diverges in
two directions where it's one the
sycopency actually can get annoying
sometimes and actually in the whole
thinking versus doing or thought partner
verse agent I'll admit too sometimes
I'll get annoyed where I ask a simple
question and it's like that's such a
great question here is a 12 layer chart
or you know like a 7 by seven table and
it just takes it too far and spends too
much time when I'm asking it a simple
question. So like on that side the
synopency when you're trying to just do
stuff gets annoying. But on the other
side like the I mean yeah more and more
the more it convinces you every idea out
of your head is good. You're right.
You're heading in the right direction.
Like that that cannot be healthy. I
don't know. It just it can I I need to
stress test the French fry French fries
in a salad. We should just start
>> crazy.
>> Yeah, that sounds excellent.
>> That does sound uh lettuce and French
fries.
>> I don't know. I think you have to. We'll
do it on air.
>> Delight.
>> Do it on air. Yeah, we should. That'd be
a great idea. All right. Next week. Um,
but it does bring me to this like other
fun story that that I've mentioned on
the show before, but Rest of World had a
great uh piece about this bot. Uh,
actually it's not even a bot. It's like
an application of this uh AI companion
that's been stuffed into a stuffed
animal. Um, it's called Hyodal and it is
for retirees. Uh it it uh asks retirees
uh whether they're taking their medic
medication or eating a meal and the it
sensors watch over the users in real
time alert alerting social workers and
families during uh emergencies. There
are 412 of these bots distributed to
seniors since 2019. Um and in Korea
there's 12,000 of them mostly in the
homes of the elderly who are uh
typically lonely. This is from the
story. There's a I think a social worker
saying that older adults take great
comfort in just having someone to talk
to. These are things they can't tell us
or even their own children, but they
tell hiodel. I think this is great.
Honestly, I would love if I was all
alone and I was old, uh I would love a
AI stuffed animal to speak with and keep
me company and, you know, make sure that
uh I'm being taken care of.
I think yeah this is a th this shows the
complexity of this entire like
conversation that at a certain level
just having something to talk to I'm
still going to go with something over
someone is can be beneficial
um especially in the idea of like
actually from a health perspective
caretaking and reminding to take a
medication or I think there was
something around uh like or imagine it
can even like based on any kind of like
health data from some kind of device.
It's actually able to understand how to
what to say to the person. I think all
that is actually seems relatively
positive. I think it's more if you yeah
I don't know just what this means for
how you interact with other people if
you get too used to something like this
uh worries me. I think that's a very
very valid question and we are
absolutely going to be dealing with the
fallout. But I do welcome you to the
camp of thinking that this companion use
case is real because it really is. Okay,
before we get out of here uh this week
we should talk about 996 in San
Francisco. So if you know about 996,
it's a work culture predominantly
associated with China where you work
9:00 a.m. to 9:00 p.m. 6 days a week. Uh
this is again from uh RAMP. Apparently,
it's a new thing in San Francisco. So,
the um this guy Araf from RAMP looked at
the data again and saw a bump of
ordering from restaurants um delivery
and takeout from employees at San
Francisco based businesses occurring on
Saturdays. So, it's basically a 5% bump
on Saturdays. He says it's new. There
was no Saturday bump like this in 20
2024, 2023 or prior years. uh it's
specific to San Francisco where of
course the AI boom is happening and he's
saying it's more than just tech so it is
software companies uh but they are also
uh retail uh retail chains uh that may
be adding more employee hours so uh what
do you think about 996 and do you think
that this is a real thing happening uh
in San Francisco tied to AI culture
where I guess like the meme now is like
we're all just like locking down for the
rest of 2025 and not having any fun. I
don't know what's your perspective on
this.
>> I mean, I think on the second question,
is it real? I do wonder how they
differentiate that it's workrelated or
people are just spending their company's
money on Saturdays and like if that's
okay. I don't know. Did do they get into
that at all?
>> So, I spoke with AR about this and
apparently um this is a this is a
comparison to the year before. So, it's
actually a new thing that they're seeing
now.
>> Okay. Okay. Well, I think I mean the
whole idea of like celebrating even the
term 996 has always bothered me a bit. I
mean, again, I I think I I work a fair
amount. Um, but I don't know, it's just
not something I always kind of get
uncomfortable when people celebrate it.
I think it's maybe as a reaction to the
idea of like where things definitely
during postcoid and remote work got a
bit uh loose at many companies. I think
maybe this is going to be a counter
reaction to that, but I don't know.
Overall,
I it makes me uncomfortable. What about
you?
>> But don't you think so? The argument for
it, I'm just going to throw it out
there, would be that the US and San
Francisco in particular needs this kind
of work ethic to keep up with China if
it's going to really factor in the AI
boom. What do you think?
>> Okay. I I mean, let him work. Let him
work. and make me uh VC subsidized AI
that I can use. Uh I guess that makes me
happy. But I don't know overall I've
never the whole like hours worked again
I worked in finance for many years and I
was on the trading side not the banking
side but it always blew my mind how like
everyone I would talk to so much of the
time was spent around spending the hours
rather than the actual output of the
work. So to me almost the very idea of
like uh assigning hours 9:00 am to 9:00
pm six days a week versus we are
actually just producing more. Maybe I'm
working 7 days a week or maybe I'm
working 9:00 a.m. to 6 p.m. Like it's
just it feels like it's a direction of
it's more about flexing the hours worked
rather than we're actually doing well.
>> Yeah, I agree with you. But also like
where I get concerned is that this is
the expectation. It's a lot of work.
It's a lot of work. I think that like we
can be expected to do a reasonable
amount of work in our jobs, but like
asking for 9 to9 and then a weekend day
at that level of hours to me is a bit
nuts. But people in San Francisco, I
don't know, they seem to be proud of it.
So I guess if it's your choice, go for
it.
>> If it's your choice, you you are free.
We will not tell you otherwise. you're
free to organize on Discord to go work
9:00 am to 9:00 pm six days a week.
>> Make it happen. Uh but with, you know,
within within reason. Don't don't
organize against Rajan and I, please. We
hope that we're we're doing uh the job
that you you've put us here for. So,
>> but if you need to overthrow us and the
people feel that's what's right,
>> we respect your wishes. But just do it
on Discord and sign up for a big
technology paid subscription so we can
um you know find something else to do.
>> I don't know. It's a it's a half-hearted
pitch. Maybe it will work. Who knows?
But yeah, anyway, why don't we end it
there? You and I will be back next week.
Lot to talk about next week again. And
uh thank you Ran John again for being
here. Uh always fun to talk.
>> All right, see you next week.
>> All right, see you then. And thank you
everybody for listening again. Aaron
Levy will be on the show next Wednesday
talking about all things AI.
[Music]