May 21, 2025 • 53 mins
AI strategist and TikTok commentator Nate B. Jones joins CXOTalk host Michael Krigsman to explain the high-stakes race among OpenAI, Google, Meta, and others to advance large language models.
The discussion explores token economics, model specialization, and hands-on prompting skills every executive needs to translate AI progress into business results.
Watch Episode 880 now and subscribe for weekly conversations with top tech leaders.
🔔 Don’t forget to like, subscribe, and share for more thought-provoking conversations with top industry leaders.
🔷 Newsletter: www.cxotalk.com/subscribe
🔷 Read the summary and key points: https://www.cxotalk.com/episode/inside-the-ai-arms-race-llms-economics-and-strategy
🔷 LinkedIn: www.linkedin.com/company/cxotalk
🔷 Twitter: twitter.com/cxotalk
00:00 The Evolution of the AI Arms Race
02:11 Diverging Strategies in AI Development
07:14 Economic and Business Impacts of AI
13:20 The Role of Specialization and Prompting in AI Usage
16:17 Building Relationships with AI and Shifting Mindsets
23:01 Maximizing Value from AI Interactions
24:09 Operationalizing AI for Business ROI
26:14 The Next Generation and AI's Evolution
31:18 Human Psychology and AI's Persuasive Power
35:24 Exploring Cognitive Parasitism and Model Personalities
37:53 Comparing AI Models and Their Applications
39:47 Specialized AI Models for Coding and Learning
40:31 Exploring LLM Ecosystems and Tools
44:30 Economics and Efficiency of LLMs
48:55 Challenges and Norms in AI Usage
52:47 Wrapping Up with Gratitude and Reflections
53:08 Announcements and Farewell
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
This is amazing a discussion inside the AI arms race,
generative AI large language models, business strategy, and
the competitive landscape of AI tools.
Today on episode 880 of CXO Talk, our guest is the
(00:20):
incredible Nate B Jones, who is not only an astute market
observer but also a legit TikTokstar.
Nate, tell us about the AI arms race.
It's not predetermined that it that it is an arms race.
This is actually one of the surprising outcomes of the last
year or two then. Even Sam Altman himself has
(00:41):
admitted this when he got started building what is now
ChatGPT at Open AI. It was just a tiny research lab
and the running assumption for awhile, even after ChatGPT
launched is that they were goingto have enough of a model lead
that there wouldn't really be anarms race in that sense.
(01:02):
That is not true. And So what we are seeing is
that this is a technology that proliferate that's easy to
share. We see that from DeepSeek to
Meta to Google to Amazon trying to get in the race.
And it's remarkable to me how quickly this technology has
enabled people to catch up who previously would not have been
(01:26):
in the game at all. It's not an arms race.
So what is it? What are we, what are we talking
about here? Just like a bunch of friends
getting together and, you know, spending.
No, I. Think it is an arms race now,
but what I'm saying is if you looked at it in 2020-2021, 2022,
you wouldn't have thought that it was headed that way
necessarily. And so we find ourselves in an
arms race. And that's really important
because if you expect to go intoan arms race the way the United
(01:48):
States expected to go into an arms race with the Soviet Union
after World War Two, you prepare.
But in this world we are findingourselves in an arms race and
all of the players are playing it by ear.
And I think that you have a different set of behaviors when
you are making it up as you go, and you see that to some extent
(02:08):
in the way some of the major players are taking actions.
As an example, I think that you see Mark Zuckerberg evolving his
strategy for Meta somewhat in real time as he sees the results
of various training runs coming out of Meta, as he sees moves
from other players in the space.And I think that is why you see
(02:28):
him start to emphasize things like AI as a companion more in
the last few weeks versus AI as a senior software development
engineer, which was more of his emphasis earlier in the year
when frankly, I think he had more optimism about the results
of the Llama 4 training run. And so these these models are
(02:50):
very difficult to train. They are not deterministic
models, they're emergent models.And that means that any given
training run, you might have spent a lot of money and you
didn't get what you wanted out of it.
And so when I think about it that way, these the the way the
news shapes becomes much more interpretable to me, Michael.
How do we reconcile, on the one hand, the billions of dollars
(03:15):
that are being spent against AI being a companion?
I mean, it's so, so bizarre to even think about it in this way.
And where does all of this plug into the the overall landscape
of these major LLM vendors? It's really important to
(03:38):
differentiate between the application we see for AI with
consumers and the base layer technology that these model
makers are engaged in building. So I would say, and I think most
folks in the space would agree with me, major model makers are
not building breakthrough technology in AI in LLM
(03:58):
specifically in order to build companions for people.
But that might be a profitable use case in some places.
And so I think when I look at the space, I start by saying
what are the incentives for eachof the players around the base
intelligence they're developing and then that shapes the
direction they want to pull the use cases.
(04:20):
And so in Meta's case, if you look at the two use cases that
I've mentioned so far, Mark talking about senior SD ES and
Mark talking about companions, companions fits right into the
milieu that Facebook is developing, right?
You want to have a social feed, You want to have a social space.
He was talking about the fact that Americans have three
(04:41):
friends, but he thinks they haveappetite for 15.
The difference is 12. He wants to close that with AI.
That is a very meta way to play the game.
But if you go over to Open AI, they have a different set of
incentives on the board. Their goal is to develop
baseline intelligence. It always has been.
(05:02):
They have accidentally, and I say that very intentionally,
they accidentally developed an incredibly powerful consumer
application when they launched ChatGPT.
They've chosen to double down onthat good fortune that they had,
and they're very intentionally building it now.
And so they're in the position where they are building a
application for consumers that they want to be almost an
(05:24):
everything app and at the same time building baseline
intelligence that they think will be much more useful to
major corporations and to very, very deep science and tech.
And so I suspect we're going to see a bifurcation in the arms
race where you have players who are going to be saying our best
models are for drug breakthroughs, Our best models
(05:47):
are for the hardest problems in supply chain management or in
automotive design or whatever the hard problem might be.
But we have excellent models that we provide to the everyday
population in an attractive app format at low cost because we
find that we can effectively monetize at that scale.
So the days of the broad generalpurpose expert at everything
(06:13):
model you feel are coming to a close.
Intelligence is scaling past task saturation very rapidly.
And so you're correct that even now Open AI has variant models
that they've forked for education, they have variant
models they've forked for drug development that are not
accessible in your app or mine. But I don't really feel the
(06:36):
difference because the intelligence that is coming to
the app is more than enough for most of the tasks I bring to it.
And that's what I mean by task saturation.
If you look at intelligence as an exponential curve, we're
saturating out on most of the tasks most of us do much, much
sooner than we hit the intelligence ceiling.
I I know someone who told me, look I am aware O3 is supposed
(06:58):
to be better. I'm a smart guy.
I don't need the difference and I don't feel the difference
versus ChatGPT 4 O. It doesn't matter for the tasks
that I do. So this specialization is
heavily underway at this time already.
It is in fact, I would expect usto see the first drugs that were
designed in part or in whole by AI within the next 24 months.
(07:23):
Folks have been using various forms of AI for a long time on
this show. We've had the CEO several times
of in silicone medicine, which is one of The Pioneers of using
adversarial networks to design drugs.
But what? Where does this leave the large
(07:47):
vendors as well as the smaller LLMS?
I mean, right now you have this very this broad landscape of LLM
tools. There's two competing dynamics
in play. Dynamic 1 is that pre training
models at the next level of scale is extremely expensive and
serving those models at scale through the cloud is super
(08:09):
pricey. Those are two separate price
loads, and big model players have to have both, and that is
why Open AI is investing so heavily in the Stargate project,
for example. But at the same time, once you
create a model, it becomes much easier to distill it for other
(08:29):
purposes, run a narrower distribution of that model that
is equivalent for many tasks much more cheaply, and
essentially proliferate the technology.
And so in a sense, if you look at the idea that the cost to
deliver intelligence at a given level is dropping roughly 10 XA
year, it is actually not too surprising to see DeepSeek hit
(08:52):
when it did because it was roughly AGPT 4 level
intelligence that arrived about a year later that depended on
the presence of GPT 4 initially and was just delivered extremely
efficiently. And so I think we're going to
continue to see that 1-2 punch where you have the major model
makers investing lots of money in moving the frontier forward,
(09:15):
and you have a proliferation of that tech very rapidly.
And that is actually one of the interesting tensions in the
space because at the end of the day, these are not capital
assets. The depreciation on these is
crazy. And so if you think about it
from the investment perspective,you're essentially as a frontier
model maker buying the future cash flows that you think will
(09:36):
be there. If you can hit a particular
intelligence level. And then betting that that
distribution relationship that you form will be stable over
time so that they don't run to and also ran that has an
equivalent model from last year.Folks, a reminder we're talking
with Nate Jones and we have questions that you can we have
(10:00):
opportunity for you to ask questions.
If you're on Twitter, use the hashtag CXO talk.
If you're watching on LinkedIn, just pop your questions into the
chat and I urge you to take advantage of this opportunity to
ask Nate pretty much whatever you want.
And we have a question from LinkedIn, from Greg Walters, who
(10:21):
says, so you're talking Nate, about a real bifurcation, Greg
Walter says between heavy research and end users and he
asks almost like B to B versus Bto C, which then need two
different revenue models and business models.
(10:41):
What do you think? Any comments on that?
Very much the case. And I think one of the
interesting ways to break out the major players in the space
is to think about the underlyingrevenue and business models that
they bring to the table already.Because if you are already a
publicly traded company, you have a very different set of
incentives versus a company thatis a disruptor, a startup, a
(11:03):
smaller company. So for example, Google is
looking to defend search revenueand you can read a lot of the
way they've played the AI game over the past few months to a
year and playing catch up as essentially a very broad based
play to 1, defend search. And two, ensure that they can
defend the contracts they have with Google Cloud.
(11:24):
And those are the two major players driving.
Satya Nadella is in a similar place and he's kind of talked
openly about this where he's looking at AI as essentially a
play to drive Azure cloud. And if he can do that, he's
going to be successful. And so I think the more you
understand the existing businessmodel constraints, the more you
understand the direction these players are going to take.
(11:44):
Greg Walters comes back and he says he foresees a world where
we ditch spreadsheets, word processors and slide desks,
slide decks for a single AI agent.
It isn't AI in your app. It's your app inside your AI.
And he says which vendor today shows that they can own that
(12:05):
last app pivot, if at all that last mile.
I don't know that I agree, to behonest with you.
And I'll tell you I part of it is just the the grizzled Gray
hairs you get from being around customers in the business space
for a long time. The joke that I've had running
for decade plus now from a martech perspective is you
(12:27):
never, ever, ever, ever bet against marketers in Excel.
Excel will outlive every B to B SAS product on the market.
And I think that that goes for PowerPoint.
And I'm not saying like I personally use Figma slides, I
love it. But the point is these
technologies are surprisingly durable because humans form
durable habits around them. And so I think that when you
(12:48):
think about how time is allocated, if you go back to
Stewart Butterfield and writing about Slack as a disruptive
force because it changes how youspend time when he wrote we
don't sell saddles here, that concept is something that I use
to think about how AI is disrupting workflows.
And so to me, I think it's actually more useful to look at
(13:10):
AI the way people are actually using it now, which is that
they're using it and they're splitting their time into
microbursts. And so they're going into Chat
GPTI mean, I do this, everyone Iknows does this.
They do some work in ChatGPT. They pull a work product out,
they paste it somewhere else andthen they continue to modify it
over there. And am I willing to bet that
(13:33):
that is going to get somewhat easier?
Yes, I think you get hat tips toward that with the way Claude
has integrated into Gmail, into Google Calendar, etcetera.
Do I think that that will suddenly mean that we will all
have an app layer driven inside a particular like chat
application? I think that's unlikely, partly
(13:55):
because of the sophisticated nature of the usage patterns
that professional users bring tothese currently best in class
apps. Best in class Excel users have a
terrifying fluency in the way they put formulas together.
You are not likely to get that level of power out of an
(14:16):
everything app. Specialization still matters,
and I think that is partly wheresoftware builders can go for
defensive positioning for building modes.
In an age when intelligence continues to rise, you can still
bet on specialization. If we look at the large language
models and they're all coming out now with these deep research
(14:39):
modes, couldn't we say that deepresearch mode or agentic AI and
really very similar is this specialized tool?
And so users who really understand how these tools work
and can make use of them effectively, do they become like
(14:59):
the Excel power users, especially when you can now
customize it for your use case and for industries and so forth.
That part is really true. In fact, I think one of the
biggest differentiators for human talent over the next
decade is your fluency in power use cases of an AI application.
(15:19):
Deep research is a good example,but so is the power of prompting
itself. Prompting continues to evolve.
It's a moving target. But if you can get good at
prompting, good at learning, you're prompting, good at
evolving, you're prompting. That has tremendous alpha for
you. I expected that alpha not to be
durable, to be honest with you. Like a year and a half ago if
you'd asked me I would have saideveryone's going to catch up on
(15:41):
prompting or the models will getbetter at inferring our intent
and it won't matter. But what I found is that's just
not true. Even with the best models having
better high quality, coherent prompting still yields alpha and
there are not enough people learning it to commoditize it.
And I don't see that actually happening because I'm realizing
(16:01):
how hard it is to learn. So I think prompting is a key
use case. I think you're correct that deep
research is another one. When agents start to come along,
which I fully expect to happen in the next few months here,
learning to use your agent well is going to be another big
differentiator. Let's take another question from
LinkedIn and then let's jump over to some questions that are
(16:22):
coming in on Twitter and you guys are asking great questions.
I encourage you just ask, ask questions, take advantage of
this opportunity. So data Insta on LinkedIn says
he's excited or she is excited or they're excited for this
discussion. How can we best leverage these
(16:43):
insights? You know, after we're done,
we're going to do some light editing on this video, create a
transcript, get a summary, and it's all going to be on the CXO
Talk website. So next week, go there, get the
transcript and study it and watch the video again and again,
those sections that are relevantto you.
(17:04):
Nate thoughts on on how folks can take the best advantage of
this? When I teach AI, like I teach AI
in a in a variety of places, I write a sub stack on it.
And what I emphasize over and over again is this is a
technology that you learn by doing.
And I think that it's harder than people realize to flip a
(17:25):
fairly deep switch in our minds to go from human default
thinking, where we think inside our own heads first, to AI
conversational default thinking.And that is the switch that I'm
trying to articulate when I teach, where I say the meta
skill you need to learn in the age of AI is to switch to an
(17:46):
conversational AI default mode when I have a question or a
thought or an idea. Now I have pretty successfully
trained myself to stick it into AI 1st, and people often
misunderstand me when I say thatbecause they assume it means I
get the answer back. And I like to remind people,
moving to an AI default stance does not mean leaving your brain
(18:10):
at the door. In fact, it's the opposite.
My brain works harder, I am sharper, I'm a better
communicator and I'm a more rigorous thinker because I have
to argue with AI all the time and I don't take what it says
for granted and I fight with it.But that sharpens my thinking in
a way that I wasn't getting whenI was just living up here inside
my own head all the time. This is a question from Arsalan
(18:30):
Khan. On the surface, it seems like
kind of a silly question, but there's a there's a very deeper
meaning here. And Arsalan says this being a
quote UN quote companion where he's referring to an earlier
part of this discussion require some type of relationship.
(18:51):
Are we saying that we need a relationship with AI and humans
are not just the master overlords for now, So we need
that relationship with AI to effectively use it?
When scientists do work in the lab and they study humans, some
of them study other intelligent animals like dolphins or like
chimpanzees. One of the things that's often
(19:12):
looked at is when do these creatures develop the ability to
have an imagined image or an imagined understanding of an
interlocutor, someone they're talking with?
When you and I are talking, we both have a sense of each other
and we have a sense and maybe ananticipation of what's going to
come next, and we shape the conversation accordingly.
(19:34):
And so when you talk to an AI, what's interesting is because
it's fluent in language, you geta lot of the same dynamic that
you get in a human conversation.And we form on our side that
similar mental image that we form when we are building a
relationship. We form an imagined community
(19:54):
with the AI. And I find this in myself.
And this doesn't need to get into a philosophical statement
about what AI is or isn't. It is just a statement about how
we humans tend to work in language.
I tend to look at a particular model, I sense a personality,
and I find it works better to engage with the personality I
see and I get. It's more productive.
(20:16):
And so just to be very concrete about it, I find that O3 from
ChatGPT is an excellent professional colleague.
It doesn't feel like a personal relationship to me, but boy, do
we get a lot done. Four O from the same model.
Maker is a more personal model. It's a little bit warmer.
Claude has a very distinct personality.
There's a lot of writing about it, and I think that they've
(20:39):
done a phenomenal job crafting Apersona with Claude that you can
feel like you can relate to. And so, yeah, I do think that
the idea of forming these relationships is important to
the way we connect with models because it's really the way our
human brains are wired, and the models in that sense mirror it.
That requires a real repetitive use of multiple models in order
(21:06):
to understand the nuances between one model and the next.
And frankly, most of us normal humans don't have the time to do
that. So how do you form a
relationship with something thatis essentially amorphous?
Which is to say, there's all of these different models and their
(21:27):
capabilities kind of blend and blur together unless you're
really an expert at it. As I sort of have worked with a
lot of people who are making this journey into AI, it's
actually a very organic practicewhen you get into it.
And so I've seen people move from having a couple chats a
week to having 10 chats a day over the course of a month or
(21:48):
two. And I think one of the big
unlocks is seeing something thatis surprisingly delightful out
of a chat. And so if you're talking about
something that's interesting to you, that's compelling to you,
and you get an insight or you get a thought from AI, your
brain is thinking, I get some cool stuff out of this and
you're more likely to come back.And I think one of the big
unlocks in driving that sort of gradual behavior shift is
(22:11):
memory. And that's something where I got
to give credit to Shed GPT for the way they launched that
product. Memory makes that app much more
sticky because it remembers something about past chats.
It's not perfect, but it remembers enough about past
chats that you then feel like you're invested, and that then
builds on itself. It's the sort of virtuous
(22:33):
flywheel of user engagement for them, where the more people use
it, the more it remembers and the more people want to use it
and so on. And So what I found in practice
is that if you get to a point invery casual usage where you get
some kind of useful insight, that's what starts to tip it.
Which suggests to me that if you're at the point where you're
(22:54):
at a couple chats a week, pick those chats and make them about
something that's really interesting to you.
Make them high value to you and see what you get.
The question is how? How can people do that?
How can users make these chats high value?
What's the entry point? My suggestion is that you get
(23:17):
curious about things in your life that feel stuck.
They might be personal, they might be professional.
Maybe you're talking about the promotion you feel like you're
missing out on. Maybe you're talking about a
business model question for the product that you're launching.
Maybe you're talking about an exercise routine you want to
start. But whatever it is, you need to
(23:38):
care enough about it it you needto be like thinking about it in
the back of your head already. And this becomes a way to engage
with it because that provides the motivation that makes it
worthwhile for you to continue that chat.
All right, let's jump to some more questions.
This is from Twitter from Gus Bechdash.
And he says insights are not real capital assets.
(24:02):
So how do you really evaluate AIassets or companies?
He's absolutely right. I was actually talking with a
CTOA few weeks ago. He manages footprint for about
A6000 person company and we talked pretty openly about this
because he has rolled out a chatproducts for their entire
(24:26):
company and everyone is reporting that they like it.
People are reporting that they're saving.
I think the estimate was 2 hoursa week, which sounds very
believable to me, but he was really honest with me.
He said, you know, Nate, they say 2 hours a week, but we don't
see it anywhere. And if you're at 6000 people and
it's 2 hours a week, that's a lot of hours that we are on
(24:49):
paper saving. But are they just going to the
coffee shop? Like what's going on with those
saved hours? And so I think when you talk
about ROI and you talk about howyou value artificial
intelligence in the business, one of the ways you need to talk
about it is how do you actually operationalize AI at pain points
and workflows so you can really rigorously measure the
(25:10):
difference. Because just rolling out a chat
app is useful, but putting AI into a particular workflow in
the business that is business critical.
Picking a leverage point where you think an LLM here can drop
10 steps out of this manual process and then measuring the
difference. Well, now you start to get an
ROI calculation and now you start to be able to talk about
(25:30):
the difference that installing AI makes.
And from a valuation perspective, look, I, I am very
much of the view that software these days is depreciating in
value very rapidly, but you havenever had more value in it.
And I'm going to leave it to theaccountants to figure out how
they want to handle that one. But to me, if you have a modular
architecture and you are installing it at high leverage
(25:53):
points in the workflow. And I go back to sort of my
training in Kaizen process management where you map out
processes and you find how processes actually work in
business to deliver value. And then you find which parts of
those processes you can hit withAI in a way that allows you to
drive dramatic efficiencies. You're going to unlock value
that way, and that will wash outinto the bottom line.
(26:14):
We need to have a discussion about the different models and
where they fit together. But an interesting question has
popped up on LinkedIn, and this is from Andrew Borg.
And he says everything that we're talking about right now is
true for current users, but muchof it will change for the next
(26:41):
generation. So we're we're kind of the the
babes in the woods learning about these new fascinating
tools. But what about the next
generation? What do you think about those
folks? I have some next generation in
my house. They're 9:00 and 7:00 and what I
tell them is things that I thinkwill be durable across longer
(27:02):
time horizon. So I think things like character
qualities like high agency, likeresilience, like curiosity,
rigorous critical thinking, those are likely to translate
well in part because those have always been valuable work skills
and in part because I am findingas I use AI that those are
extremely valuable in the age ofAII, do not think that we will
(27:24):
be expected to be lower agency workers when we are managing 10
AI agents, quite the opposite. And so I can't prognosticate a
straight line and say this is what the world is going to look
like in March of 2027. But I can say that I think we
are moving toward a world where intelligence is saturating tasks
(27:46):
very rapidly. We are gaining slower on the
ability to maintain intent over time with agents, but we're
getting there. And we're going to have agents
that can do useful work. And it's going to be up to us to
figure out how to actually applythem.
And there's going to be a massive decade, 15 year long
ragged edge around this technology where businesses are
(28:06):
going to be adopting it at different paces.
And we are going to need to be figuring out department by
department, team by team, personby person, how you actually pull
AI into the work that you do. On the subject of pulling AI
into the work that you do, Journey on Twitter says What do
(28:27):
you see as the future of Gen. AI versus agentic AI?
Their biggest issue with currentGen.
AI is its lack of accuracy and the lack of sourcing for the
information it generates. Is this solved by a gentic AI
and should Gen. AI be phased out in favor of it?
(28:49):
I disagree with the dichotomy. So if you actually look at it
from a classical machine learning perspective, large
language models and transformer architectures are a branch in
the much larger machine learningfamily.
I think that's the first and correct place to start.
If you look at the creation of agents based on LLMS, they are
(29:12):
essentially, and I was talking with the developer about this
last week, they're LLMS that have tools and a good set of
system instructions. It's not actually a
fundamentally different technology.
They still rely on transformer based architectures and all
we're doing is giving them toolsand good instructions.
And if you want to come back to hallucinations, I I would call
(29:33):
out a couple things. One, LLMS are hallucinating much
less than they used to. And I think that's one of the
most pernicious myths in the business is that the state of
hallucination that we had in 2022 has been steady.
It hasn't been, it's, it's come down arguably a factor of 100
and it is much, much more accurate now.
And it is true that it gets evenmore accurate as you give LLMS
(29:57):
inference time compute so that they can reason and as you give
LLMS tools so that they can reason with tools.
It's not a straight line correlation though, because I
also find that if you give an LLM tools, sometimes it writes
the tool or or uses the tool in the way it imagines is most
(30:17):
useful rather than in the way that corresponds to reality.
So as an example, and this this is a real SEO tip for the
marketers out there, it is a known issue with some LL Ms.
that when they use the web to reason, they will sometimes come
back with what are called malformed URLs.
But if you look at those malformed URLs, you'll find
(30:40):
they're not malformed. They are actually beautiful URLs
that correspond precisely to what the the topic should be for
that particular piece of information that the LLM wants.
And that is effectively free SEOguidance and advice.
Make a make a page URL with thatname because that is what the
(31:05):
chat wants to see exist. And that's just a little
sidebar. I will say the idea of of links
being malformed is getting better over time.
O3 is much, much better about this than four O was three
months ago, and it will keep getting better.
Simone Jo Moore says, do you find the psychology of humans
wants a relationships last slashdiscussion with AI because the
(31:29):
AI has no judgement on the individual.
Even with all of the disaster attempts where it's been, it's
shown a mean bias, but there's no judgement is do you think
that's a driving factor here of the psychology?
I would say kind of yes and which is what we used to do in
theater. So I think my, my, yes and would
be the the concept of AI as an infinitely patient listener is
(31:54):
very powerful for people. That part is really true.
The effect that it has on peoplevaries based on our baseline
psychological makeup. And I have had enough exposure.
I get all kinds of crazy emails that I see the range of
relationship styles that people are able to form with AI.
(32:15):
And I find that for people who have a psychological disposition
that enables them to engage withthe outside world, to form
relationships with others that are reasonably healthy, none of
us are perfect. They're able to relate to AI
somewhat in the way you could relate to a colleague or a
friend. They're not talking to it all
(32:37):
the time, but they talk to it regularly and they get their
value back. And they do find that AI
listening to them can help them process things that would be
difficult to process with a friend sometimes because the
friend might not be available ormight not be up to listening to
whatever you want to talk about.So far, so good.
But for folks who have personality disorders of various
(32:58):
sorts, it is a very, very dangerous tool.
And I have seen first hand chatsfrom people who don't have sort
of a stable baseline psychological profile.
And effectively the chat encourages dark patterns in that
scenario and sort of sucks them in into their own world,
isolates them. And it is a dangerous thing.
(33:19):
And it's something we should talk about because I think the
impact is very different depending on your baseline
ability to form relationships inthe real world.
This is a tough problem. It is.
It's not an easy one. These tools are so profoundly
convincing. Yes.
In fact, I saw a study this pastweek that the implied persuasive
(33:41):
multiple for AI is something like 5 or 6 XA human.
And that got me thinking becauseI don't know if you saw the
news, but the CEO of Instacart came over to open AI this week.
And if you look at her LinkedIn,everywhere she's gone she's
built ad systems. My default assumption is she is
(34:03):
there to build an ad system for ChatGPT.
I would be shocked if she didn't.
I think her title is CEO of applications, something like
that. And she, I think, is likely also
to be almost like the Chief Operating officer to build out
the consumer facing part of the company beyond just the raw
(34:30):
chat. But what have we got that we can
sell to people? To me, I think the thing that's
compelling about what's saleablewith ChatGPT with other
conversations is you have effectively the ability to
compress the entire buying funnel from a marketer
perspective into a chat. You can have an initial
(34:51):
awareness, you can have buying intent begin to form.
You can have option selection. I lived this because I went
through that whole process with ChatGPT.
When I picked out a sound system, I had ChatGPT do
research for me. I talked about my options.
I fed it a blueprint of my living room, and we went through
the whole buying funnel together.
(35:12):
It was profoundly informative inthe ultimate purchase that I
made, and I think that that is exactly the kind of conversation
that people are having multiplied by millions all over
the world right now and only more so going forward.
Nastasha Smith on Twitter says Nate, can you explain cognitive
(35:32):
parasitism using the model's ownloops to rewrite its runtime
behavior versus regular prompting?
If I understand correctly, the idea is you're using the chat
interface as a way to shape model personality, to move it
through late in space, to shape how it operates at runtime.
(35:54):
That is absolutely possible. There are definitely folks on
Twitter who've gone very, very deep down that rabbit hole much
deeper than me. And I think the lesson I take
away from that is that the default vanilla model that you
get in the chat application is acarefully molded face of a much
(36:15):
more complex intelligence. You will find that if you ever
compare outputs from the raw APIfeed to a chat application, this
is true for Claude, this is truefor ChatGPT, this is true for
Gemini. And you can do things like
adjust temperature, adjust top Pin the API and you can get
different results, but you can also to this user's question in
(36:39):
chat in prompt, start to move the model to a different
personality stance. And that is sometimes a very
multi turn process where you're utilizing memory and you're
saying, I want you to remember this, I want you to think about
this. And you almost, it's almost like
if you've ever done yoga, the way an instructor will guide you
(36:59):
into a meditative place with particular sets of mental
instructions, like imagine yourself in a quiet place.
Well, you're not really in a quiet place if the truck is
going by, but you're imagining that.
And that works on your body in the same way.
You can kind of use that mindsetto give the chat instructions
that move the LLM into a different space.
(37:21):
It's a it's a really interestingexperience if you've ever done
it. And the term cognitive
parasitism, I never heard that term before.
No, I, I mean, I've heard it once or twice.
I think to me the, the way I tend to think of this is, is, is
perhaps a little bit more positive.
It is navigating the model through latent space.
(37:44):
And so if you think about it as a higher dimensional space that
the model's drawing from, you'renavigating where that model
stands is in the shape of that model through through that
space. Let's talk about the models.
We have Open AI. We have Google.
We have Microsoft. We have Anthropic.
We have Facebook. We have Twitter with Grok.
(38:06):
Oh, with X? Yeah.
With yes X thank you. I'm still stuck in the old
Twitter days as opposed to the. I I guess it's hard.
And there's all there's a political dimension here, but we
won't go there. So what's the difference between
these models? How can we determine when to use
what? I think there's a few shorthands
(38:29):
that are really helpful. The 1st is make sure your intent
is clear. Like.
For me, I think it's very usefulto talk about it in 2 frames. 1
is what is your best everyday model, and then there's a bunch
of other things you might use for speciality applications.
And then the other frame is whatis the best model for a given
(38:50):
hard task. And I think if you put those two
frames together, you have a really interesting rubric to
understand how these models apply.
And so for me, I can make a pretty convincing case having
done some head to heads. But the best everyday model out
there right now is probably chatG PTO3.
I have some people in the last week who are making the case for
(39:12):
Grok 3. They're arguing that it has more
more less constraints on output tokens versus the more
constraints on output tokens forO3, so it's more useful.
I'll test it, We'll see. But I find that if you can name
and choose a best everyday modelfor you based on fairly rigorous
(39:33):
sort of comparison testing across common queries, then you
have answered 90% of your workflow questions.
And then it becomes, well, what hard problems or what edge cases
are interesting to solve. And I'll give you some examples
there. So Google has done a phenomenal
job on a number of different fronts.
I think their product distribution or is really
(39:54):
challenged right now. But from a coding perspective,
Gemini 2.5 Pro is an incredible coding assistant.
I love it. That is my default model when I
am coding in Windsurf. I go to it all the time.
I find it more useful in the open AI models in coding right
now. And it used to be Cloud 3.5, but
I've switched to Gemini. And if I look at the Notebook LM
(40:19):
product that Google launched, itis an incredible product just
for learning about new topics. I can put 300 different PDFs, I
can put links in, I can put YouTube videos in, I can chat
with the entire corpus of knowledge.
It's phenomenal. When I look at Claude, I find
as, as I mentioned, the personality is there, but I also
think that Claude is really, really useful for exploring the
(40:40):
concept of LLMS with tools because Claude is so tightly
bound into the MCP universe, thethe model context protocol
universe, and those are basically shop fronts for tools
that the LLM can call. And Anthropic is leaning in
heavily. They're the anchor for that
ecosystem. They have now made it possible
that you can just paste a URL infor an MCP and Claude will just
(41:03):
pick it up and start to use the tool.
And I think that you're going tosee more and more examples where
Claude is leveraging the MPC ecosystem they fostered to give
themselves additional scale because they're not the biggest
player in the space. There are a a bunch of other
models and tools that build and extend the models.
(41:26):
For example, you have DeepSeek, you have perplexity,
youhaveyou.com. You have Manus.
You have Jen Spark. Yeah.
Where do all of these fit into this landscape?
I think I tend to distinguish between sort of base models like
DeepSeek and tools like Manus and others.
(41:47):
So tools have sort of their own life to them.
They're either good or they're not good for particular use
cases. And models tend to be that sort
of baseline intelligence that's either good for your everyday
use case or it's good for a range of use cases.
And so DeepSeek is a model that people seem to be primarily
using to pull open source, host themselves and have a high
(42:12):
quality model that they've been able to host themselves.
After the initial flurry of likeapp downloads and this and that,
That is where I actually see DeepSeek getting used.
Tools like Manas are essentiallyfast forwarding that agentic
future. It's the idea that you can use
an agent to prepare reasoning across the web and a very
complex output. So I've used Manas to produce
(42:34):
like a website in one shot basedon research, and I find that
useful. But what I found so far with
what is essentially this first crop of agentic tools is they
tend to need a fair bit of shepherding to actually deliver
value. And that's true also of the
(42:54):
coding agent. So Devon people have really
mixed feelings on the Devon AI agent because if you know how to
use it and you're a senior engineer and you wanted to pick
up Tier 3 tasks and just work while you're not looking and you
know what you can expect it to do, It does fine.
If you are not an engineer and if you are giving it more than
(43:18):
it can handle and if you're not supervising it correctly, it's
going to be a disaster to work with.
And so I think that a lot of a lot of the arts at this stage,
and this is a now piece of advice, right?
Next year if we talk, it's goingto be a different conversation.
But for right now, you need to be on your toes when you're
using those tools because they will not always give you what
(43:41):
you expect. I know someone who was using
John Spark just yesterday and they were initially it was
really funny. I got like 2 successive Slack
messages. The first one was, Oh my God, it
worked. And it showed this screenshot of
this incredible job that John Spark had done at filling out
and enriching lead contacts, right?
So that so they had e-mail addresses and can you enrich
(44:02):
that into like a profile on the person and Jens Spark had gone
out and done a gentic web searchand like pulled stuff in blah,
blah, blah. Well, the next Slack message
said, oh wait, because most of the of the links that have been
pasted in were fake. It was about 80% fake.
And so I think one of the thingsthat that underlined for me is
(44:22):
that these tools are able to do good work, but you have to be
smart enough to use them well and not over expect from them.
Tell us about the economics. So much of this is being driven
by the the economics facing the large language model with tokens
and so forth. So tell us about that.
To start with, like, and Sasha put this so succinctly back in
(44:46):
January, he said that the foundational equation of our age
is going to be dollars per tokenper Watt.
And I think that that sums up somuch.
If you are looking at any kind of scaled application, you're
not in the chat, you're in the API, you're looking at token
costs. And fundamentally you have two
different classes of cost. You have input tokens that come
at a much lower cost and you have output tokens that come at
(45:09):
a much higher cost. And the job of application
designers is essentially to figure out for a given model,
for a given model, what is the correct alication, what is the
correct output token length? How can I use that efficiently?
And if I'm looking at a very complex task, I should have
different models and I should only use the intelligence I
(45:29):
need. So I'm not overpaying for
intelligence. And I think one of the big open
questions for solution designersright now is how do you right
size the intelligence in a rigorous way for the task that
you have? There's not really one standard
framework to do that today. And I think we're living in a
world where that's going to change in the next year or two
because there's so much appetitefor intelligence and people
(45:51):
don't want to overpay for it. At the end of the day, the token
cost is driven by the energy it takes to process through a,
through a chip, the query that users bring to the LOM.
And so it literally comes down to the watts that, that that
Satya was talking about. They're pulling in electricity,
they're running the chips, they have them in giant data centers.
(46:14):
And you are paying for that token to hit the chip, generate
a response and come back. And it's like electricity
metering, but it's for 2025. Why is the cost per token going
down so dramatically? Well, there's a few reasons.
One, NVIDIA is getting better and better at making these chips
(46:36):
and making them more efficient. Anytime you have this kind of
capital infusion in the space, there's a tremendous drive for
efficiency so that you can do more with the demand.
And like demand is through the roof.
Microsoft has had to promise that they will double data
center availability by 2027 in Europe because there's so much
(46:57):
European demand for AI and that's still not enough.
And so if you have demand like that, the pressure on efficiency
is tremendous. Chipsets, the way you put the
chips together into server racks, all of that technical
data center design stuff, that comes down to designing it so
it's as efficient as possible toserve.
And and by the way, serving models is different than
(47:18):
training them. People often confuse those
separate tracks, separate efficiency distributions,
separate servers, sometimes separate chips, and you have to
design the entire configuration of the server rack in the data
Center for the workload you anticipate.
And so all of that technical stuff happens so that you can
(47:39):
very efficiently send a query intoday and get a response that is
efficient to serve. Now, from a pricing perspective,
the fact that we started this conversation talking about an
arms race is highly significant here because there's so much
competition. People are driving prices down.
One of the reasons why open AIAPI costs have dropped is
because Deep Sea came out and open sourced a very good model.
(48:02):
It's also because Gemini has been relentless about dropping
intelligent models that are very, very cheap to serve.
From an API perspective, that's an that's a market grabbing play
by Google. They're doing that on purpose
because they want you to switch your behavior as an API consumer
and move your tokens over there.As the context window of the
(48:26):
models grows larger, the aggregate cost may increase even
if the cost per token is significantly smaller.
The footprint of the models is growing from a from a sort of
cost and energy perspective for sure we are using more and more
and more of these models and I would anticipate that growth to
(48:47):
continue. Everybody in the space
anticipates that demand will keep sky rocketing, and so the
overall footprint will grow and grow and grow.
Yeah, absolutely. We have another question.
I just want to grab very fast, and this is from Elizabeth Shaw,
who says not everyone's thinkingis sharpened by the use of LOMS.
There are many people who use the output as truth and that
(49:11):
pushes the average work and expectation level down.
People are looking for shortcutsand less work.
How will LLMS deliver on the promise you speak of, given this
kind of promulgation of least common denominator information
(49:35):
that so often comes out? When I think about that, I go
back to the selection pressures that we've talked about all the
way through. We talked about how we have
never seen intelligence revolution like this.
We have never seen growth and scale in non human intelligence
like this. Previous machine learning
generations didn't do this. People think about these as
(49:59):
companions. They react to them badly.
They react to them well, but they do.
And when I put all of that together with sort of the
business pressures that we have when we install AII think the
answer is that business pressureacts as a selector for high
quality usage of AI. And so I will tell you, I don't
(50:20):
tolerate AI slop when I see it from other people either.
And people who are coming with AI slop into the business are
going to get negative repercussions because people who
use AI well will simply out compete them.
That product will be better. The thinking will be more
rigorous. And so do I believe that people
(50:41):
will use AI to produce slop and bad answers and lazy responses?
Yes. I was reading a profile on
someone who went to an Ivy League university and is just
using AI to generate answers andkind of Slough their way
through. And by the way, it can't be that
the results are that average if they're getting good grades all
the way through to which they were.
(51:03):
The problem is that they're not learning the skills they need.
And so you will get people who do that.
They will do well for a while, and then they're going to run
into a situation where someone out competes them because they
use AI well. And that is going to have to
play out enough times for enoughpeople for norms to start to
form around what best practice looks like.
(51:25):
But I have no doubt, because I've seen it play out in my own
life enough times in conversations with others enough
times, that one, people who use AI well will out compete humans
who don't, and two, people who use AI well will ruthlessly out
compete people who take AI as the answer.
Arsalan Khan, he's going to get the last question here.
(51:47):
Does chat create a false sense of security and privacy very
fast? Please, Nate.
I would argue that humans have afalse sense of privacy on the
Internet to begin with, and chatextends that.
So fundamentally, I think we allimagine our usage of the
Internet is much more private than it is.
If you've ever worked in AD tech, you will never think that
(52:07):
way again. It just doesn't work that way
and we bring that same mindset to AI.
And yes, you can read the terms of service.
It's not. It's not as bad as the nightmare
people will tell you unless you're using a model with a
really poor terms of service. Deep seeks are not great if you
use their cloud provided model. But Even so, you're not as
private as you think you are in general.
(52:29):
That's my answer. I can tell you, I sure wouldn't
put a lot of private informationinto Manus, which is from China,
or DeepSeek, which is Chinese. I'd be pretty careful about what
I put into my LLM. And on that note, Nate B Jones,
(52:50):
thank you so much for taking thetime to speak with us today and
share your wealth of knowledge. Very grateful to you.
I had an absolute blast. I think this was a great
conversation. I really appreciate the quality
questions we got. I think that made a lot of fun.
And thank you to everybody who asked such amazing questions.
You guys are awesome. Before you go, subscribe to the
(53:11):
CXO Talk newsletter on our website.
We have incredible shows. There's no show next week, but
the next one after that in two weeks from today is the
Executive Vice President and Chief Product Officer of Cisco,
and he's great. So you must check it out.
Thank you so much, everybody. Thanks to Nate, and I hope you
(53:31):
have a great day. Take care everyone.
This is amazing a discussion inside the AI arms race,
generative AI large language models, business strategy, and
the competitive landscape of AI tools.
Today on episode 880 of CXO Talk, our guest is the
(00:20):
incredible Nate B Jones, who is not only an astute market
observer but also a legit TikTokstar.
Nate, tell us about the AI arms race.
It's not predetermined that it that it is an arms race.
This is actually one of the surprising outcomes of the last
year or two then. Even Sam Altman himself has
(00:41):
admitted this when he got started building what is now
ChatGPT at Open AI. It was just a tiny research lab
and the running assumption for awhile, even after ChatGPT
launched is that they were goingto have enough of a model lead
that there wouldn't really be anarms race in that sense.
(01:02):
That is not true. And So what we are seeing is
that this is a technology that proliferate that's easy to
share. We see that from DeepSeek to
Meta to Google to Amazon trying to get in the race.
And it's remarkable to me how quickly this technology has
enabled people to catch up who previously would not have been
(01:26):
in the game at all. It's not an arms race.
So what is it? What are we, what are we talking
about here? Just like a bunch of friends
getting together and, you know, spending.
No, I. Think it is an arms race now,
but what I'm saying is if you looked at it in 2020-2021, 2022,
you wouldn't have thought that it was headed that way
necessarily. And so we find ourselves in an
arms race. And that's really important
because if you expect to go intoan arms race the way the United
(01:48):
States expected to go into an arms race with the Soviet Union
after World War Two, you prepare.
But in this world we are findingourselves in an arms race and
all of the players are playing it by ear.
And I think that you have a different set of behaviors when
you are making it up as you go, and you see that to some extent
(02:08):
in the way some of the major players are taking actions.
As an example, I think that you see Mark Zuckerberg evolving his
strategy for Meta somewhat in real time as he sees the results
of various training runs coming out of Meta, as he sees moves
from other players in the space.And I think that is why you see
(02:28):
him start to emphasize things like AI as a companion more in
the last few weeks versus AI as a senior software development
engineer, which was more of his emphasis earlier in the year
when frankly, I think he had more optimism about the results
of the Llama 4 training run. And so these these models are
(02:50):
very difficult to train. They are not deterministic
models, they're emergent models.And that means that any given
training run, you might have spent a lot of money and you
didn't get what you wanted out of it.
And so when I think about it that way, these the the way the
news shapes becomes much more interpretable to me, Michael.
How do we reconcile, on the one hand, the billions of dollars
(03:15):
that are being spent against AI being a companion?
I mean, it's so, so bizarre to even think about it in this way.
And where does all of this plug into the the overall landscape
of these major LLM vendors? It's really important to
(03:38):
differentiate between the application we see for AI with
consumers and the base layer technology that these model
makers are engaged in building. So I would say, and I think most
folks in the space would agree with me, major model makers are
not building breakthrough technology in AI in LLM
(03:58):
specifically in order to build companions for people.
But that might be a profitable use case in some places.
And so I think when I look at the space, I start by saying
what are the incentives for eachof the players around the base
intelligence they're developing and then that shapes the
direction they want to pull the use cases.
(04:20):
And so in Meta's case, if you look at the two use cases that
I've mentioned so far, Mark talking about senior SD ES and
Mark talking about companions, companions fits right into the
milieu that Facebook is developing, right?
You want to have a social feed, You want to have a social space.
He was talking about the fact that Americans have three
(04:41):
friends, but he thinks they haveappetite for 15.
The difference is 12. He wants to close that with AI.
That is a very meta way to play the game.
But if you go over to Open AI, they have a different set of
incentives on the board. Their goal is to develop
baseline intelligence. It always has been.
(05:02):
They have accidentally, and I say that very intentionally,
they accidentally developed an incredibly powerful consumer
application when they launched ChatGPT.
They've chosen to double down onthat good fortune that they had,
and they're very intentionally building it now.
And so they're in the position where they are building a
application for consumers that they want to be almost an
(05:24):
everything app and at the same time building baseline
intelligence that they think will be much more useful to
major corporations and to very, very deep science and tech.
And so I suspect we're going to see a bifurcation in the arms
race where you have players who are going to be saying our best
models are for drug breakthroughs, Our best models
(05:47):
are for the hardest problems in supply chain management or in
automotive design or whatever the hard problem might be.
But we have excellent models that we provide to the everyday
population in an attractive app format at low cost because we
find that we can effectively monetize at that scale.
So the days of the broad generalpurpose expert at everything
(06:13):
model you feel are coming to a close.
Intelligence is scaling past task saturation very rapidly.
And so you're correct that even now Open AI has variant models
that they've forked for education, they have variant
models they've forked for drug development that are not
accessible in your app or mine. But I don't really feel the
(06:36):
difference because the intelligence that is coming to
the app is more than enough for most of the tasks I bring to it.
And that's what I mean by task saturation.
If you look at intelligence as an exponential curve, we're
saturating out on most of the tasks most of us do much, much
sooner than we hit the intelligence ceiling.
I I know someone who told me, look I am aware O3 is supposed
(06:58):
to be better. I'm a smart guy.
I don't need the difference and I don't feel the difference
versus ChatGPT 4 O. It doesn't matter for the tasks
that I do. So this specialization is
heavily underway at this time already.
It is in fact, I would expect usto see the first drugs that were
designed in part or in whole by AI within the next 24 months.
(07:23):
Folks have been using various forms of AI for a long time on
this show. We've had the CEO several times
of in silicone medicine, which is one of The Pioneers of using
adversarial networks to design drugs.
But what? Where does this leave the large
(07:47):
vendors as well as the smaller LLMS?
I mean, right now you have this very this broad landscape of LLM
tools. There's two competing dynamics
in play. Dynamic 1 is that pre training
models at the next level of scale is extremely expensive and
serving those models at scale through the cloud is super
(08:09):
pricey. Those are two separate price
loads, and big model players have to have both, and that is
why Open AI is investing so heavily in the Stargate project,
for example. But at the same time, once you
create a model, it becomes much easier to distill it for other
(08:29):
purposes, run a narrower distribution of that model that
is equivalent for many tasks much more cheaply, and
essentially proliferate the technology.
And so in a sense, if you look at the idea that the cost to
deliver intelligence at a given level is dropping roughly 10 XA
year, it is actually not too surprising to see DeepSeek hit
(08:52):
when it did because it was roughly AGPT 4 level
intelligence that arrived about a year later that depended on
the presence of GPT 4 initially and was just delivered extremely
efficiently. And so I think we're going to
continue to see that 1-2 punch where you have the major model
makers investing lots of money in moving the frontier forward,
(09:15):
and you have a proliferation of that tech very rapidly.
And that is actually one of the interesting tensions in the
space because at the end of the day, these are not capital
assets. The depreciation on these is
crazy. And so if you think about it
from the investment perspective,you're essentially as a frontier
model maker buying the future cash flows that you think will
(09:36):
be there. If you can hit a particular
intelligence level. And then betting that that
distribution relationship that you form will be stable over
time so that they don't run to and also ran that has an
equivalent model from last year.Folks, a reminder we're talking
with Nate Jones and we have questions that you can we have
(10:00):
opportunity for you to ask questions.
If you're on Twitter, use the hashtag CXO talk.
If you're watching on LinkedIn, just pop your questions into the
chat and I urge you to take advantage of this opportunity to
ask Nate pretty much whatever you want.
And we have a question from LinkedIn, from Greg Walters, who
(10:21):
says, so you're talking Nate, about a real bifurcation, Greg
Walter says between heavy research and end users and he
asks almost like B to B versus Bto C, which then need two
different revenue models and business models.
(10:41):
What do you think? Any comments on that?
Very much the case. And I think one of the
interesting ways to break out the major players in the space
is to think about the underlyingrevenue and business models that
they bring to the table already.Because if you are already a
publicly traded company, you have a very different set of
incentives versus a company thatis a disruptor, a startup, a
(11:03):
smaller company. So for example, Google is
looking to defend search revenueand you can read a lot of the
way they've played the AI game over the past few months to a
year and playing catch up as essentially a very broad based
play to 1, defend search. And two, ensure that they can
defend the contracts they have with Google Cloud.
(11:24):
And those are the two major players driving.
Satya Nadella is in a similar place and he's kind of talked
openly about this where he's looking at AI as essentially a
play to drive Azure cloud. And if he can do that, he's
going to be successful. And so I think the more you
understand the existing businessmodel constraints, the more you
understand the direction these players are going to take.
(11:44):
Greg Walters comes back and he says he foresees a world where
we ditch spreadsheets, word processors and slide desks,
slide decks for a single AI agent.
It isn't AI in your app. It's your app inside your AI.
And he says which vendor today shows that they can own that
(12:05):
last app pivot, if at all that last mile.
I don't know that I agree, to behonest with you.
And I'll tell you I part of it is just the the grizzled Gray
hairs you get from being around customers in the business space
for a long time. The joke that I've had running
for decade plus now from a martech perspective is you
(12:27):
never, ever, ever, ever bet against marketers in Excel.
Excel will outlive every B to B SAS product on the market.
And I think that that goes for PowerPoint.
And I'm not saying like I personally use Figma slides, I
love it. But the point is these
technologies are surprisingly durable because humans form
durable habits around them. And so I think that when you
(12:48):
think about how time is allocated, if you go back to
Stewart Butterfield and writing about Slack as a disruptive
force because it changes how youspend time when he wrote we
don't sell saddles here, that concept is something that I use
to think about how AI is disrupting workflows.
And so to me, I think it's actually more useful to look at
(13:10):
AI the way people are actually using it now, which is that
they're using it and they're splitting their time into
microbursts. And so they're going into Chat
GPTI mean, I do this, everyone Iknows does this.
They do some work in ChatGPT. They pull a work product out,
they paste it somewhere else andthen they continue to modify it
over there. And am I willing to bet that
(13:33):
that is going to get somewhat easier?
Yes, I think you get hat tips toward that with the way Claude
has integrated into Gmail, into Google Calendar, etcetera.
Do I think that that will suddenly mean that we will all
have an app layer driven inside a particular like chat
application? I think that's unlikely, partly
(13:55):
because of the sophisticated nature of the usage patterns
that professional users bring tothese currently best in class
apps. Best in class Excel users have a
terrifying fluency in the way they put formulas together.
You are not likely to get that level of power out of an
(14:16):
everything app. Specialization still matters,
and I think that is partly wheresoftware builders can go for
defensive positioning for building modes.
In an age when intelligence continues to rise, you can still
bet on specialization. If we look at the large language
models and they're all coming out now with these deep research
(14:39):
modes, couldn't we say that deepresearch mode or agentic AI and
really very similar is this specialized tool?
And so users who really understand how these tools work
and can make use of them effectively, do they become like
(14:59):
the Excel power users, especially when you can now
customize it for your use case and for industries and so forth.
That part is really true. In fact, I think one of the
biggest differentiators for human talent over the next
decade is your fluency in power use cases of an AI application.
(15:19):
Deep research is a good example,but so is the power of prompting
itself. Prompting continues to evolve.
It's a moving target. But if you can get good at
prompting, good at learning, you're prompting, good at
evolving, you're prompting. That has tremendous alpha for
you. I expected that alpha not to be
durable, to be honest with you. Like a year and a half ago if
you'd asked me I would have saideveryone's going to catch up on
(15:41):
prompting or the models will getbetter at inferring our intent
and it won't matter. But what I found is that's just
not true. Even with the best models having
better high quality, coherent prompting still yields alpha and
there are not enough people learning it to commoditize it.
And I don't see that actually happening because I'm realizing
(16:01):
how hard it is to learn. So I think prompting is a key
use case. I think you're correct that deep
research is another one. When agents start to come along,
which I fully expect to happen in the next few months here,
learning to use your agent well is going to be another big
differentiator. Let's take another question from
LinkedIn and then let's jump over to some questions that are
(16:22):
coming in on Twitter and you guys are asking great questions.
I encourage you just ask, ask questions, take advantage of
this opportunity. So data Insta on LinkedIn says
he's excited or she is excited or they're excited for this
discussion. How can we best leverage these
(16:43):
insights? You know, after we're done,
we're going to do some light editing on this video, create a
transcript, get a summary, and it's all going to be on the CXO
Talk website. So next week, go there, get the
transcript and study it and watch the video again and again,
those sections that are relevantto you.
(17:04):
Nate thoughts on on how folks can take the best advantage of
this? When I teach AI, like I teach AI
in a in a variety of places, I write a sub stack on it.
And what I emphasize over and over again is this is a
technology that you learn by doing.
And I think that it's harder than people realize to flip a
(17:25):
fairly deep switch in our minds to go from human default
thinking, where we think inside our own heads first, to AI
conversational default thinking.And that is the switch that I'm
trying to articulate when I teach, where I say the meta
skill you need to learn in the age of AI is to switch to an
(17:46):
conversational AI default mode when I have a question or a
thought or an idea. Now I have pretty successfully
trained myself to stick it into AI 1st, and people often
misunderstand me when I say thatbecause they assume it means I
get the answer back. And I like to remind people,
moving to an AI default stance does not mean leaving your brain
(18:10):
at the door. In fact, it's the opposite.
My brain works harder, I am sharper, I'm a better
communicator and I'm a more rigorous thinker because I have
to argue with AI all the time and I don't take what it says
for granted and I fight with it.But that sharpens my thinking in
a way that I wasn't getting whenI was just living up here inside
my own head all the time. This is a question from Arsalan
(18:30):
Khan. On the surface, it seems like
kind of a silly question, but there's a there's a very deeper
meaning here. And Arsalan says this being a
quote UN quote companion where he's referring to an earlier
part of this discussion require some type of relationship.
(18:51):
Are we saying that we need a relationship with AI and humans
are not just the master overlords for now, So we need
that relationship with AI to effectively use it?
When scientists do work in the lab and they study humans, some
of them study other intelligent animals like dolphins or like
chimpanzees. One of the things that's often
(19:12):
looked at is when do these creatures develop the ability to
have an imagined image or an imagined understanding of an
interlocutor, someone they're talking with?
When you and I are talking, we both have a sense of each other
and we have a sense and maybe ananticipation of what's going to
come next, and we shape the conversation accordingly.
(19:34):
And so when you talk to an AI, what's interesting is because
it's fluent in language, you geta lot of the same dynamic that
you get in a human conversation.And we form on our side that
similar mental image that we form when we are building a
relationship. We form an imagined community
(19:54):
with the AI. And I find this in myself.
And this doesn't need to get into a philosophical statement
about what AI is or isn't. It is just a statement about how
we humans tend to work in language.
I tend to look at a particular model, I sense a personality,
and I find it works better to engage with the personality I
see and I get. It's more productive.
(20:16):
And so just to be very concrete about it, I find that O3 from
ChatGPT is an excellent professional colleague.
It doesn't feel like a personal relationship to me, but boy, do
we get a lot done. Four O from the same model.
Maker is a more personal model. It's a little bit warmer.
Claude has a very distinct personality.
There's a lot of writing about it, and I think that they've
(20:39):
done a phenomenal job crafting Apersona with Claude that you can
feel like you can relate to. And so, yeah, I do think that
the idea of forming these relationships is important to
the way we connect with models because it's really the way our
human brains are wired, and the models in that sense mirror it.
That requires a real repetitive use of multiple models in order
(21:06):
to understand the nuances between one model and the next.
And frankly, most of us normal humans don't have the time to do
that. So how do you form a
relationship with something thatis essentially amorphous?
Which is to say, there's all of these different models and their
(21:27):
capabilities kind of blend and blur together unless you're
really an expert at it. As I sort of have worked with a
lot of people who are making this journey into AI, it's
actually a very organic practicewhen you get into it.
And so I've seen people move from having a couple chats a
week to having 10 chats a day over the course of a month or
(21:48):
two. And I think one of the big
unlocks is seeing something thatis surprisingly delightful out
of a chat. And so if you're talking about
something that's interesting to you, that's compelling to you,
and you get an insight or you get a thought from AI, your
brain is thinking, I get some cool stuff out of this and
you're more likely to come back.And I think one of the big
unlocks in driving that sort of gradual behavior shift is
(22:11):
memory. And that's something where I got
to give credit to Shed GPT for the way they launched that
product. Memory makes that app much more
sticky because it remembers something about past chats.
It's not perfect, but it remembers enough about past
chats that you then feel like you're invested, and that then
builds on itself. It's the sort of virtuous
(22:33):
flywheel of user engagement for them, where the more people use
it, the more it remembers and the more people want to use it
and so on. And So what I found in practice
is that if you get to a point invery casual usage where you get
some kind of useful insight, that's what starts to tip it.
Which suggests to me that if you're at the point where you're
(22:54):
at a couple chats a week, pick those chats and make them about
something that's really interesting to you.
Make them high value to you and see what you get.
The question is how? How can people do that?
How can users make these chats high value?
What's the entry point? My suggestion is that you get
(23:17):
curious about things in your life that feel stuck.
They might be personal, they might be professional.
Maybe you're talking about the promotion you feel like you're
missing out on. Maybe you're talking about a
business model question for the product that you're launching.
Maybe you're talking about an exercise routine you want to
start. But whatever it is, you need to
(23:38):
care enough about it it you needto be like thinking about it in
the back of your head already. And this becomes a way to engage
with it because that provides the motivation that makes it
worthwhile for you to continue that chat.
All right, let's jump to some more questions.
This is from Twitter from Gus Bechdash.
And he says insights are not real capital assets.
(24:02):
So how do you really evaluate AIassets or companies?
He's absolutely right. I was actually talking with a
CTOA few weeks ago. He manages footprint for about
A6000 person company and we talked pretty openly about this
because he has rolled out a chatproducts for their entire
(24:26):
company and everyone is reporting that they like it.
People are reporting that they're saving.
I think the estimate was 2 hoursa week, which sounds very
believable to me, but he was really honest with me.
He said, you know, Nate, they say 2 hours a week, but we don't
see it anywhere. And if you're at 6000 people and
it's 2 hours a week, that's a lot of hours that we are on
(24:49):
paper saving. But are they just going to the
coffee shop? Like what's going on with those
saved hours? And so I think when you talk
about ROI and you talk about howyou value artificial
intelligence in the business, one of the ways you need to talk
about it is how do you actually operationalize AI at pain points
and workflows so you can really rigorously measure the
(25:10):
difference. Because just rolling out a chat
app is useful, but putting AI into a particular workflow in
the business that is business critical.
Picking a leverage point where you think an LLM here can drop
10 steps out of this manual process and then measuring the
difference. Well, now you start to get an
ROI calculation and now you start to be able to talk about
(25:30):
the difference that installing AI makes.
And from a valuation perspective, look, I, I am very
much of the view that software these days is depreciating in
value very rapidly, but you havenever had more value in it.
And I'm going to leave it to theaccountants to figure out how
they want to handle that one. But to me, if you have a modular
architecture and you are installing it at high leverage
(25:53):
points in the workflow. And I go back to sort of my
training in Kaizen process management where you map out
processes and you find how processes actually work in
business to deliver value. And then you find which parts of
those processes you can hit withAI in a way that allows you to
drive dramatic efficiencies. You're going to unlock value
that way, and that will wash outinto the bottom line.
(26:14):
We need to have a discussion about the different models and
where they fit together. But an interesting question has
popped up on LinkedIn, and this is from Andrew Borg.
And he says everything that we're talking about right now is
true for current users, but muchof it will change for the next
(26:41):
generation. So we're we're kind of the the
babes in the woods learning about these new fascinating
tools. But what about the next
generation? What do you think about those
folks? I have some next generation in
my house. They're 9:00 and 7:00 and what I
tell them is things that I thinkwill be durable across longer
(27:02):
time horizon. So I think things like character
qualities like high agency, likeresilience, like curiosity,
rigorous critical thinking, those are likely to translate
well in part because those have always been valuable work skills
and in part because I am findingas I use AI that those are
extremely valuable in the age ofAII, do not think that we will
(27:24):
be expected to be lower agency workers when we are managing 10
AI agents, quite the opposite. And so I can't prognosticate a
straight line and say this is what the world is going to look
like in March of 2027. But I can say that I think we
are moving toward a world where intelligence is saturating tasks
(27:46):
very rapidly. We are gaining slower on the
ability to maintain intent over time with agents, but we're
getting there. And we're going to have agents
that can do useful work. And it's going to be up to us to
figure out how to actually applythem.
And there's going to be a massive decade, 15 year long
ragged edge around this technology where businesses are
(28:06):
going to be adopting it at different paces.
And we are going to need to be figuring out department by
department, team by team, personby person, how you actually pull
AI into the work that you do. On the subject of pulling AI
into the work that you do, Journey on Twitter says What do
(28:27):
you see as the future of Gen. AI versus agentic AI?
Their biggest issue with currentGen.
AI is its lack of accuracy and the lack of sourcing for the
information it generates. Is this solved by a gentic AI
and should Gen. AI be phased out in favor of it?
(28:49):
I disagree with the dichotomy. So if you actually look at it
from a classical machine learning perspective, large
language models and transformer architectures are a branch in
the much larger machine learningfamily.
I think that's the first and correct place to start.
If you look at the creation of agents based on LLMS, they are
(29:12):
essentially, and I was talking with the developer about this
last week, they're LLMS that have tools and a good set of
system instructions. It's not actually a
fundamentally different technology.
They still rely on transformer based architectures and all
we're doing is giving them toolsand good instructions.
And if you want to come back to hallucinations, I I would call
(29:33):
out a couple things. One, LLMS are hallucinating much
less than they used to. And I think that's one of the
most pernicious myths in the business is that the state of
hallucination that we had in 2022 has been steady.
It hasn't been, it's, it's come down arguably a factor of 100
and it is much, much more accurate now.
And it is true that it gets evenmore accurate as you give LLMS
(29:57):
inference time compute so that they can reason and as you give
LLMS tools so that they can reason with tools.
It's not a straight line correlation though, because I
also find that if you give an LLM tools, sometimes it writes
the tool or or uses the tool in the way it imagines is most
(30:17):
useful rather than in the way that corresponds to reality.
So as an example, and this this is a real SEO tip for the
marketers out there, it is a known issue with some LL Ms.
that when they use the web to reason, they will sometimes come
back with what are called malformed URLs.
But if you look at those malformed URLs, you'll find
(30:40):
they're not malformed. They are actually beautiful URLs
that correspond precisely to what the the topic should be for
that particular piece of information that the LLM wants.
And that is effectively free SEOguidance and advice.
Make a make a page URL with thatname because that is what the
(31:05):
chat wants to see exist. And that's just a little
sidebar. I will say the idea of of links
being malformed is getting better over time.
O3 is much, much better about this than four O was three
months ago, and it will keep getting better.
Simone Jo Moore says, do you find the psychology of humans
wants a relationships last slashdiscussion with AI because the
(31:29):
AI has no judgement on the individual.
Even with all of the disaster attempts where it's been, it's
shown a mean bias, but there's no judgement is do you think
that's a driving factor here of the psychology?
I would say kind of yes and which is what we used to do in
theater. So I think my, my, yes and would
be the the concept of AI as an infinitely patient listener is
(31:54):
very powerful for people. That part is really true.
The effect that it has on peoplevaries based on our baseline
psychological makeup. And I have had enough exposure.
I get all kinds of crazy emails that I see the range of
relationship styles that people are able to form with AI.
(32:15):
And I find that for people who have a psychological disposition
that enables them to engage withthe outside world, to form
relationships with others that are reasonably healthy, none of
us are perfect. They're able to relate to AI
somewhat in the way you could relate to a colleague or a
friend. They're not talking to it all
(32:37):
the time, but they talk to it regularly and they get their
value back. And they do find that AI
listening to them can help them process things that would be
difficult to process with a friend sometimes because the
friend might not be available ormight not be up to listening to
whatever you want to talk about.So far, so good.
But for folks who have personality disorders of various
(32:58):
sorts, it is a very, very dangerous tool.
And I have seen first hand chatsfrom people who don't have sort
of a stable baseline psychological profile.
And effectively the chat encourages dark patterns in that
scenario and sort of sucks them in into their own world,
isolates them. And it is a dangerous thing.
(33:19):
And it's something we should talk about because I think the
impact is very different depending on your baseline
ability to form relationships inthe real world.
This is a tough problem. It is.
It's not an easy one. These tools are so profoundly
convincing. Yes.
In fact, I saw a study this pastweek that the implied persuasive
(33:41):
multiple for AI is something like 5 or 6 XA human.
And that got me thinking becauseI don't know if you saw the
news, but the CEO of Instacart came over to open AI this week.
And if you look at her LinkedIn,everywhere she's gone she's
built ad systems. My default assumption is she is
(34:03):
there to build an ad system for ChatGPT.
I would be shocked if she didn't.
I think her title is CEO of applications, something like
that. And she, I think, is likely also
to be almost like the Chief Operating officer to build out
the consumer facing part of the company beyond just the raw
(34:30):
chat. But what have we got that we can
sell to people? To me, I think the thing that's
compelling about what's saleablewith ChatGPT with other
conversations is you have effectively the ability to
compress the entire buying funnel from a marketer
perspective into a chat. You can have an initial
(34:51):
awareness, you can have buying intent begin to form.
You can have option selection. I lived this because I went
through that whole process with ChatGPT.
When I picked out a sound system, I had ChatGPT do
research for me. I talked about my options.
I fed it a blueprint of my living room, and we went through
the whole buying funnel together.
(35:12):
It was profoundly informative inthe ultimate purchase that I
made, and I think that that is exactly the kind of conversation
that people are having multiplied by millions all over
the world right now and only more so going forward.
Nastasha Smith on Twitter says Nate, can you explain cognitive
(35:32):
parasitism using the model's ownloops to rewrite its runtime
behavior versus regular prompting?
If I understand correctly, the idea is you're using the chat
interface as a way to shape model personality, to move it
through late in space, to shape how it operates at runtime.
(35:54):
That is absolutely possible. There are definitely folks on
Twitter who've gone very, very deep down that rabbit hole much
deeper than me. And I think the lesson I take
away from that is that the default vanilla model that you
get in the chat application is acarefully molded face of a much
(36:15):
more complex intelligence. You will find that if you ever
compare outputs from the raw APIfeed to a chat application, this
is true for Claude, this is truefor ChatGPT, this is true for
Gemini. And you can do things like
adjust temperature, adjust top Pin the API and you can get
different results, but you can also to this user's question in
(36:39):
chat in prompt, start to move the model to a different
personality stance. And that is sometimes a very
multi turn process where you're utilizing memory and you're
saying, I want you to remember this, I want you to think about
this. And you almost, it's almost like
if you've ever done yoga, the way an instructor will guide you
(36:59):
into a meditative place with particular sets of mental
instructions, like imagine yourself in a quiet place.
Well, you're not really in a quiet place if the truck is
going by, but you're imagining that.
And that works on your body in the same way.
You can kind of use that mindsetto give the chat instructions
that move the LLM into a different space.
(37:21):
It's a it's a really interestingexperience if you've ever done
it. And the term cognitive
parasitism, I never heard that term before.
No, I, I mean, I've heard it once or twice.
I think to me the, the way I tend to think of this is, is, is
perhaps a little bit more positive.
It is navigating the model through latent space.
(37:44):
And so if you think about it as a higher dimensional space that
the model's drawing from, you'renavigating where that model
stands is in the shape of that model through through that
space. Let's talk about the models.
We have Open AI. We have Google.
We have Microsoft. We have Anthropic.
We have Facebook. We have Twitter with Grok.
(38:06):
Oh, with X? Yeah.
With yes X thank you. I'm still stuck in the old
Twitter days as opposed to the. I I guess it's hard.
And there's all there's a political dimension here, but we
won't go there. So what's the difference between
these models? How can we determine when to use
what? I think there's a few shorthands
(38:29):
that are really helpful. The 1st is make sure your intent
is clear. Like.
For me, I think it's very usefulto talk about it in 2 frames. 1
is what is your best everyday model, and then there's a bunch
of other things you might use for speciality applications.
And then the other frame is whatis the best model for a given
(38:50):
hard task. And I think if you put those two
frames together, you have a really interesting rubric to
understand how these models apply.
And so for me, I can make a pretty convincing case having
done some head to heads. But the best everyday model out
there right now is probably chatG PTO3.
I have some people in the last week who are making the case for
(39:12):
Grok 3. They're arguing that it has more
more less constraints on output tokens versus the more
constraints on output tokens forO3, so it's more useful.
I'll test it, We'll see. But I find that if you can name
and choose a best everyday modelfor you based on fairly rigorous
(39:33):
sort of comparison testing across common queries, then you
have answered 90% of your workflow questions.
And then it becomes, well, what hard problems or what edge cases
are interesting to solve. And I'll give you some examples
there. So Google has done a phenomenal
job on a number of different fronts.
I think their product distribution or is really
(39:54):
challenged right now. But from a coding perspective,
Gemini 2.5 Pro is an incredible coding assistant.
I love it. That is my default model when I
am coding in Windsurf. I go to it all the time.
I find it more useful in the open AI models in coding right
now. And it used to be Cloud 3.5, but
I've switched to Gemini. And if I look at the Notebook LM
(40:19):
product that Google launched, itis an incredible product just
for learning about new topics. I can put 300 different PDFs, I
can put links in, I can put YouTube videos in, I can chat
with the entire corpus of knowledge.
It's phenomenal. When I look at Claude, I find
as, as I mentioned, the personality is there, but I also
think that Claude is really, really useful for exploring the
(40:40):
concept of LLMS with tools because Claude is so tightly
bound into the MCP universe, thethe model context protocol
universe, and those are basically shop fronts for tools
that the LLM can call. And Anthropic is leaning in
heavily. They're the anchor for that
ecosystem. They have now made it possible
that you can just paste a URL infor an MCP and Claude will just
(41:03):
pick it up and start to use the tool.
And I think that you're going tosee more and more examples where
Claude is leveraging the MPC ecosystem they fostered to give
themselves additional scale because they're not the biggest
player in the space. There are a a bunch of other
models and tools that build and extend the models.
(41:26):
For example, you have DeepSeek, you have perplexity,
youhaveyou.com. You have Manus.
You have Jen Spark. Yeah.
Where do all of these fit into this landscape?
I think I tend to distinguish between sort of base models like
DeepSeek and tools like Manus and others.
(41:47):
So tools have sort of their own life to them.
They're either good or they're not good for particular use
cases. And models tend to be that sort
of baseline intelligence that's either good for your everyday
use case or it's good for a range of use cases.
And so DeepSeek is a model that people seem to be primarily
using to pull open source, host themselves and have a high
(42:12):
quality model that they've been able to host themselves.
After the initial flurry of likeapp downloads and this and that,
That is where I actually see DeepSeek getting used.
Tools like Manas are essentiallyfast forwarding that agentic
future. It's the idea that you can use
an agent to prepare reasoning across the web and a very
complex output. So I've used Manas to produce
(42:34):
like a website in one shot basedon research, and I find that
useful. But what I found so far with
what is essentially this first crop of agentic tools is they
tend to need a fair bit of shepherding to actually deliver
value. And that's true also of the
(42:54):
coding agent. So Devon people have really
mixed feelings on the Devon AI agent because if you know how to
use it and you're a senior engineer and you wanted to pick
up Tier 3 tasks and just work while you're not looking and you
know what you can expect it to do, It does fine.
If you are not an engineer and if you are giving it more than
(43:18):
it can handle and if you're not supervising it correctly, it's
going to be a disaster to work with.
And so I think that a lot of a lot of the arts at this stage,
and this is a now piece of advice, right?
Next year if we talk, it's goingto be a different conversation.
But for right now, you need to be on your toes when you're
using those tools because they will not always give you what
(43:41):
you expect. I know someone who was using
John Spark just yesterday and they were initially it was
really funny. I got like 2 successive Slack
messages. The first one was, Oh my God, it
worked. And it showed this screenshot of
this incredible job that John Spark had done at filling out
and enriching lead contacts, right?
So that so they had e-mail addresses and can you enrich
(44:02):
that into like a profile on the person and Jens Spark had gone
out and done a gentic web searchand like pulled stuff in blah,
blah, blah. Well, the next Slack message
said, oh wait, because most of the of the links that have been
pasted in were fake. It was about 80% fake.
And so I think one of the thingsthat that underlined for me is
(44:22):
that these tools are able to do good work, but you have to be
smart enough to use them well and not over expect from them.
Tell us about the economics. So much of this is being driven
by the the economics facing the large language model with tokens
and so forth. So tell us about that.
To start with, like, and Sasha put this so succinctly back in
(44:46):
January, he said that the foundational equation of our age
is going to be dollars per tokenper Watt.
And I think that that sums up somuch.
If you are looking at any kind of scaled application, you're
not in the chat, you're in the API, you're looking at token
costs. And fundamentally you have two
different classes of cost. You have input tokens that come
at a much lower cost and you have output tokens that come at
(45:09):
a much higher cost. And the job of application
designers is essentially to figure out for a given model,
for a given model, what is the correct alication, what is the
correct output token length? How can I use that efficiently?
And if I'm looking at a very complex task, I should have
different models and I should only use the intelligence I
(45:29):
need. So I'm not overpaying for
intelligence. And I think one of the big open
questions for solution designersright now is how do you right
size the intelligence in a rigorous way for the task that
you have? There's not really one standard
framework to do that today. And I think we're living in a
world where that's going to change in the next year or two
because there's so much appetitefor intelligence and people
(45:51):
don't want to overpay for it. At the end of the day, the token
cost is driven by the energy it takes to process through a,
through a chip, the query that users bring to the LOM.
And so it literally comes down to the watts that, that that
Satya was talking about. They're pulling in electricity,
they're running the chips, they have them in giant data centers.
(46:14):
And you are paying for that token to hit the chip, generate
a response and come back. And it's like electricity
metering, but it's for 2025. Why is the cost per token going
down so dramatically? Well, there's a few reasons.
One, NVIDIA is getting better and better at making these chips
(46:36):
and making them more efficient. Anytime you have this kind of
capital infusion in the space, there's a tremendous drive for
efficiency so that you can do more with the demand.
And like demand is through the roof.
Microsoft has had to promise that they will double data
center availability by 2027 in Europe because there's so much
(46:57):
European demand for AI and that's still not enough.
And so if you have demand like that, the pressure on efficiency
is tremendous. Chipsets, the way you put the
chips together into server racks, all of that technical
data center design stuff, that comes down to designing it so
it's as efficient as possible toserve.
And and by the way, serving models is different than
(47:18):
training them. People often confuse those
separate tracks, separate efficiency distributions,
separate servers, sometimes separate chips, and you have to
design the entire configuration of the server rack in the data
Center for the workload you anticipate.
And so all of that technical stuff happens so that you can
(47:39):
very efficiently send a query intoday and get a response that is
efficient to serve. Now, from a pricing perspective,
the fact that we started this conversation talking about an
arms race is highly significant here because there's so much
competition. People are driving prices down.
One of the reasons why open AIAPI costs have dropped is
because Deep Sea came out and open sourced a very good model.
(48:02):
It's also because Gemini has been relentless about dropping
intelligent models that are very, very cheap to serve.
From an API perspective, that's an that's a market grabbing play
by Google. They're doing that on purpose
because they want you to switch your behavior as an API consumer
and move your tokens over there.As the context window of the
(48:26):
models grows larger, the aggregate cost may increase even
if the cost per token is significantly smaller.
The footprint of the models is growing from a from a sort of
cost and energy perspective for sure we are using more and more
and more of these models and I would anticipate that growth to
(48:47):
continue. Everybody in the space
anticipates that demand will keep sky rocketing, and so the
overall footprint will grow and grow and grow.
Yeah, absolutely. We have another question.
I just want to grab very fast, and this is from Elizabeth Shaw,
who says not everyone's thinkingis sharpened by the use of LOMS.
There are many people who use the output as truth and that
(49:11):
pushes the average work and expectation level down.
People are looking for shortcutsand less work.
How will LLMS deliver on the promise you speak of, given this
kind of promulgation of least common denominator information
(49:35):
that so often comes out? When I think about that, I go
back to the selection pressures that we've talked about all the
way through. We talked about how we have
never seen intelligence revolution like this.
We have never seen growth and scale in non human intelligence
like this. Previous machine learning
generations didn't do this. People think about these as
(49:59):
companions. They react to them badly.
They react to them well, but they do.
And when I put all of that together with sort of the
business pressures that we have when we install AII think the
answer is that business pressureacts as a selector for high
quality usage of AI. And so I will tell you, I don't
(50:20):
tolerate AI slop when I see it from other people either.
And people who are coming with AI slop into the business are
going to get negative repercussions because people who
use AI well will simply out compete them.
That product will be better. The thinking will be more
rigorous. And so do I believe that people
(50:41):
will use AI to produce slop and bad answers and lazy responses?
Yes. I was reading a profile on
someone who went to an Ivy League university and is just
using AI to generate answers andkind of Slough their way
through. And by the way, it can't be that
the results are that average if they're getting good grades all
the way through to which they were.
(51:03):
The problem is that they're not learning the skills they need.
And so you will get people who do that.
They will do well for a while, and then they're going to run
into a situation where someone out competes them because they
use AI well. And that is going to have to
play out enough times for enoughpeople for norms to start to
form around what best practice looks like.
(51:25):
But I have no doubt, because I've seen it play out in my own
life enough times in conversations with others enough
times, that one, people who use AI well will out compete humans
who don't, and two, people who use AI well will ruthlessly out
compete people who take AI as the answer.
Arsalan Khan, he's going to get the last question here.
(51:47):
Does chat create a false sense of security and privacy very
fast? Please, Nate.
I would argue that humans have afalse sense of privacy on the
Internet to begin with, and chatextends that.
So fundamentally, I think we allimagine our usage of the
Internet is much more private than it is.
If you've ever worked in AD tech, you will never think that
(52:07):
way again. It just doesn't work that way
and we bring that same mindset to AI.
And yes, you can read the terms of service.
It's not. It's not as bad as the nightmare
people will tell you unless you're using a model with a
really poor terms of service. Deep seeks are not great if you
use their cloud provided model. But Even so, you're not as
private as you think you are in general.
(52:29):
That's my answer. I can tell you, I sure wouldn't
put a lot of private informationinto Manus, which is from China,
or DeepSeek, which is Chinese. I'd be pretty careful about what
I put into my LLM. And on that note, Nate B Jones,
(52:50):
thank you so much for taking thetime to speak with us today and
share your wealth of knowledge. Very grateful to you.
I had an absolute blast. I think this was a great
conversation. I really appreciate the quality
questions we got. I think that made a lot of fun.
And thank you to everybody who asked such amazing questions.
You guys are awesome. Before you go, subscribe to the
(53:11):
CXO Talk newsletter on our website.
We have incredible shows. There's no show next week, but
the next one after that in two weeks from today is the
Executive Vice President and Chief Product Officer of Cisco,
and he's great. So you must check it out.
Thank you so much, everybody. Thanks to Nate, and I hope you
(53:31):
have a great day. Take care everyone.
Popular Podcasts
Bookmarked by Reese's Book Club
Welcome to Bookmarked by Reese’s Book Club — the podcast where great stories, bold women, and irresistible conversations collide! Hosted by award-winning journalist Danielle Robay, each week new episodes balance thoughtful literary insight with the fervor of buzzy book trends, pop culture and more. Bookmarked brings together celebrities, tastemakers, influencers and authors from Reese's Book Club and beyond to share stories that transcend the page. Pull up a chair. You’re not just listening — you’re part of the conversation.
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
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.