November 10, 2025 • 49 mins
Dan and Chris unpack whether today’s surge in AI deployment across enterprise workflows, manufacturing, healthcare, and scientific research signals a lasting transformation or an overhyped bubble. Drawing parallels to the dot-com era, they explore how technology integration is reshaping industries, affecting jobs, and even influencing human cognition, ultimately asking: is this a bubble, or just a fizzy new phase of innovation?
Featuring:
Links:
Sponsors:
- Outshift by Cisco - The open source collective building the Internet of Agents. Backed by Outshift by Cisco, AGNTCY gives developers the tools to build and deploy multi-agent software at scale. Identity, communication protocols, and modular workflows—all in one global collaboration layer. Start building at AGNTCY.org.
- Shopify – The commerce platform trusted by millions. From idea to checkout, Shopify gives you everything you need to launch and scale your business—no matter your level of experience. Build beautiful storefronts, market with built-in AI tools, and tap into the platform powering 10% of all U.S. eCommerce. Start your one-dollar trial at shopify.com/practicalai
- Fabi.ai - The all-in-one data analysis platform for modern teams. From ad hoc queries to advanced analytics, Fabi lets you explore data wherever it lives—spreadsheets, Postgres, Snowflake, Airtable and more. Built-in Python and AI assistance help you move fast, then publish interactive dashboards or automate insights delivered straight to Slack, email, spreadsheets or wherever you need to share it. Learn more and get started for free at fabi.ai
Upcoming Events:
- Join us at the Midwest AI Summit on November 13 in Indianapolis to hear world-class speakers share how they’ve scaled AI solutions. Don’t miss the AI Engineering Lounge, where you can sit down with experts for hands-on guidance. Reserve your spot today!
- Register for upcoming webinars here!
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Jerod (00:04):
Welcome to the Practical AI podcast, where we break down
the real world applications ofartificial intelligence and how
it's shaping the way we live,work, and create. Our goal is to
help make AI technologypractical, productive, and
accessible to everyone. Whetheryou're a developer, business
leader, or just curious aboutthe tech behind the buzz, you're
(00:24):
in the right place. Be sure toconnect with us on LinkedIn, X,
or Blue Sky to stay up to datewith episode drops, behind the
scenes content, and AI insights.You can learn more at
practicalai.fm.
Now onto the show.
Sponsor (00:40):
Well, friends, when you're building and shipping AI
products at scale, there's oneconstant, complexity. Yes.
You're bringing the models, datapipelines, deployment
infrastructure, and then someonesays, let's turn this into a
business. Cue the chaos. That'swhere Shopify steps in whether
you're spinning up a storefrontfor your AI powered app or
(01:00):
launching a brand around thetools you built.
Shopify is the commerce platformtrusted by millions of
businesses and 10% of all USecommerce from names like
Mattel, Gymshark, to foundersjust like you. With literally
hundreds of ready to usetemplates, powerful built in
marketing tools, and AI thatwrites product descriptions for
(01:21):
you, headlines, even polishesyour product photography.
Shopify doesn't just get youselling, it makes you look good
doing it. And we love it. We useit here at Changelog.
Check us outmerch.changelog.com. That's our
store front, and it handles theheavy lifting too. Payments,
inventory, returns, shipping,even global logistics. It's like
(01:43):
having an ops team built intoyour stack to help you sell. So
if you're ready to sell, you areready for Shopify.
Sign up now for your $1 permonth trial and start selling
today atshopify.com/practicalai. Again,
that is shopify.com/practicalai.
Daniel (02:17):
Welcome to another episode of the Practical AI
Podcast. In these fullyconnected episodes where it's
just Chris and I without aguest, we like to explore some
topics from trending AI news oror things that are being
discussed in the community.Hopefully, we learn along the
way and and our listeners learnalong the way, and and we help
(02:39):
you all level up your AI andmachine learning game. I'm
Daniel Witenack. I am CEO atPrediction Guard, and I am
joined as always by my cohost,Chris Benson, who is a principal
AI research engineer at LockheedMartin.
How you doing, Chris?
Chris (02:54):
Hey. Doing good today, Daniel. Just cruising along,
chewing some bubble gum, doingmy thing.
Daniel (03:02):
We both needed a sugary, but not sugary caffeinated
drink. I think you've got a Pibbzero. I've got a Coke zero.
Chris (03:12):
There you go.
Daniel (03:13):
It's been a long day of work and now we get to talk
about some fun things.Hopefully, we won't pop the AI
bubble today, but, certainly
Chris (03:25):
That's okay. We'll we'll clean up the situation if we
have to.
Daniel (03:29):
Yeah. Exactly. It it's interesting, Chris. I was
looking around. Of course, we'realways seeing things or people
are forwarding us various thingsin the news related to AI.
One this week that I saw wasmaybe an kind of interesting one
with a humanoid looking robot,but I'd rather maybe discuss
(03:49):
this other stuff, which I thinka couple of these things that I
saw connected. And Chris, alittle while ago, we kind of did
a hot takes and debates typeepisode. I think today, there's
certainly on this topic, thereare various sides of this topic
and strong opinions about it. Wecan just talk through some of
(04:11):
those. Certainly, there is adebate going on.
And, this is really thequestion, are we in an AI
bubble? Which seems to be talkedabout all the time. Like people
see something in the news andthey're like, oh, we're
definitely in an AI bubble. Orpeople are like, oh, this isn't
an actual AI bubble. It'sdifferent than maybe the .com
(04:34):
bubble.
Chris, you hear similaranecdotes and and maybe what is
your maybe just the generalconcept of of bubble. We're not
talking about bubble gum, but
Chris (04:45):
Yeah. I mean, the general notion of a bubble, you know, is
it's really a financial conceptwhere the valuation of an
organization exceeds what itsactual value is. In other words,
what it's putting out in termsof product and services, and the
returns that are yielded bythose. And so, you know, and,
(05:08):
you know, if you were talkingfor instance about the .com
bubble and for those of you whomay not have depending on if
you're in the younger skew ofour audience, around 2,000 ish
in in that there was all thisinternet craze and rage and hype
not dissimilar from the hypewe've been seeing in recent
(05:30):
years over AI And a lot ofcompanies came about and new
startups and stuff, and they gotvalued very, very high, but they
had very little, in some cases,no revenue and thus no profits
available and so big valuationby the market with absolutely
(05:50):
nothing coming out of them ofvalue and so it there's a point
where the market kind ofrealizes that and corrects and
and the the giant.com, you know,facade came falling down and
kinda led into a recessionaryperiod around the globe.
And so it was kind of a bigthing over several years. And so
(06:12):
through this entire period of ofAI buildup, that's been a
concern that has come upregularly. This is not the first
time, you know, we've beenhearing about AI bubbles and
stuff. So, I don't think there'sa year that's gone by that we've
been doing this podcast where ithasn't been raised as an issue.
What do you think?
Daniel (06:30):
Maybe so. Yeah. And I think there is, you know,
genuine concern because previousbubbles that have burst, have
actually caused real harm. Asyou mentioned, whether that's
economic kind of recession,certainly it distorts the way
people invest or maybe what theyinvest into, which kind of has
(06:54):
an effect potentially onretirement or 401ks. There's a
reduction in trust in certaintypes of organizations or
financial institutions ortechnology companies, that sort
of thing.
And it could be companies, itcould be assets. Like, a lot of
(07:15):
times people talk about cryptoas an asset that Or I remember
not too long ago talking to tonsof people about NFTs, right? And
this really chaotic time withNFTs and there's a lot of people
that lost a lot of money inthat. Indeed. So it is a valid
(07:36):
concern and I think the questionthat's on people's mind is, are
we in an AI bubble?
And one of the interestingarticles that I saw this week,
Chris, was that Powell, theFederal Reserve Chair, Jerome
(07:56):
Powell in The US, for those thataren't listening from The US,
that's the Federal ReserveChair, Jerome Powell, often can
kind of anything that's saidabout the economy by whoever's
in this position is taken with alot of weight because, well,
(08:20):
Chris, you may have comments onthis, but I'm not an economist.
But generally it sign is of atleast an intelligent opinion on
things where there's a lotthat's gone on or a direction
that, you know, the FederalReserve wants people to think.
Chris (08:39):
So so the two second background without going off on
a a a a track is Jerome Powellin his capacity is responsible
for a particular committee atthe Federal Reserve, and they
are responsible for monetarypolicy, which the Federal
Reserve sets as opposed to whatthe president and congress
(09:00):
together set, which is fiscalpolicy. And one of the tools,
the largest tool that theymainly do that with is through
the setting of interest rateswhich trickles through the
entire economy in a whole bunchof different ways. I don't wanna
go into any more depth thanthat, but they they thus slow
down or speed up the economy toeither tackle an underperforming
(09:23):
economy or inflation on theopposite side, and they're
trying to balance it between thetwo. And so, yeah, chairman
Powell noted that there wasactually revenue associated with
AI expenditure. And thus, itwasn't bubble like in his view,
when he is certainly an expertin a lot of ways, think I think
(09:44):
that there are some otherconsiderations in there.
But it's definitely a powerfulstatement coming from that
particular individual.
Daniel (09:52):
Yeah, yeah. The title of the article in Fortune which
we'll link is Powell says thatunlike the .com boom, AI
spending isn't a bubble. And thequote is, I won't go into
particular names, but theyactually have earnings. Now I
guess we could speculate on thispodcast. I don't know what names
he's talking about.
I'm assuming it's some of theselarger names that would be
(10:17):
whatever OpenAI, Anthropic,Cohere, you know, whatever the
ones are that people would thinkof. I imagine some of those are,
quote, the names. But he saysthey sort of actually have
earnings. The other interestingpiece that I saw, Chris, which
is maybe on the other side ofthis argument, and we can talk
(10:40):
maybe after this about, youknow, the different ways that
people argue that we are oraren't in an AI bubble. But this
other one was from the New YorkTimes, which is reporting that
NVIDIA is now worth$5,000,000,000,000 and this is
the quote from the article, asit consolidates power in AI
(11:01):
boom.
So it's the subheading, the AIchipmaker has become a linchpin
in the Trump administration'strade negotiations with Asia. So
there's there's a, you know,some policy and political angle
to this article, but the generalidea with the article is that
this $5,000,000,000,000valuation maybe is is part of an
(11:25):
AI boom. So yeah, that'scertainly interesting.
Chris (11:30):
It is. And I think we're seeing that, going back to those
names that were left unsaid, younamed a few of those names. And,
you know, I think any of thelarge cloud service providers
that are offering a collectionof AI services probably, you
know, round out some of thosenames. One of the things, one of
(11:53):
the distinctions that I think isinteresting to to weigh as we
talk about this is the fact thatwhile he noted, Mr. Powell noted
that some of these organizationshave earnings and we are seeing,
you know, that reflected in AIrelated stocks that are just
(12:13):
dominating S and P 500 returnsto the to the degree of about
75% of those returns.
80% of earnings growth, and 90%of capital spending growth. I
mean, are phenomenal numberswhen you think about that in
terms of percentages of totalmarkets that are out there, you
know, or of these at leastexchanges that are representing
(12:35):
markets. So I mean, that'sreally shocking, but I think one
of the things that is that thathe did not say, speaking as
someone who has nowhere near theeconomic expertise of mister
Powell, but did did have someuniversity studies in economics
that you know, those thosenumbers are concentrated in a
(12:58):
tiny fraction of companiesoverall. And so my question that
I would ask is for for some ofthose giant companies that are
that are making huge earningsfrom many, many, many thousands
of customers spending hugeamounts of money on AI growth,
We have also talked about thefact that there is there's also
(13:23):
in the media quite a bit ofquestioning of ROI on a lot of
those AI investments at variouscompanies. And if you're the
cloud provider making bookeys ofmoney, that's great.
But if you're one of thecompanies out there that maybe
is spending on that, but maybenot seeing an ROI on that
(13:44):
expenditure, That is a differentstory right there. And so, you
know, I you know, maybe it'sboom for one and a bubble for
another. Maybe it's not just auniversal bubble, but it depends
on who you're talking to, tosome degree.
Daniel (13:59):
Yeah, and I guess that's where maybe this is murky is
the, some people might definebubble differently. I think that
it is one of the kind of keyarguments for the kind of
affirmative of this that we arein an AI bubble is this kind of
valuations and speculation. Imean, we've highlighted a few
(14:21):
stories over time the lastcouple of years of these crazy
valuations where essentiallythere is an unproven revenue
model. So I'm looking at onesource here that's saying, over
50% of VC funding in Q2 of thisyear, 2025, went into AI
(14:43):
companies. And VC is risky, ofcourse, as an investment model.
Most of those businesses willfail. That's kind of always
expected in the AI space. Butalso this year, just from
looking around at differentfolks raising, these companies
are getting much, much highermultiples or higher valuations.
(15:07):
So for people that need areminder, sometimes your company
might receive a valuation that'sa certain multiple of the
revenue that you're bringing in.And so that might be fifteen,
twenty, 25x, 30x plus for an AIcompany, where other companies,
(15:28):
kind of your normal run of themill SaaS company in tech that's
raising is definitely notraising at those multiples right
now.
So that would be, I guess, theargument or an argument for that
affirmative is these kind ofspeculation and valuation that's
(15:48):
maybe reminiscent of that .comera.
Chris (15:51):
Yeah. By the way, another term just to connect kind of
that financial world with whatwe're talking about in terms of
observations is those of youinvest, some of you may have
heard the notion of a beta,which is essentially a a
multiple of valuation againstyour earnings. And so if you
(16:14):
have a very high beta that'ssaying you're being valued very
high against what your actualreal life earnings are and a
lower beta, which would beconsidered less risky, would be
that that valuation is not soextravagant. And so I guess and
that's another way of of lookingat this is if you're looking at
a company's, you know, portfolioanalysis, you know, that some
(16:37):
analyst is doing, and theremight be a beta number attached
if you're is that and this waspointed out in some of the
articles is that the betas oftoday, while high, are not
nearly as high as the betas ofthe .com era. And so, you know,
that's another sign.
Well, you know, is there somebubble? Maybe. Is it as bad in
(17:00):
terms of the sheer speculationof the .com era? Maybe not. You
know, that's one metric by whichwe can evaluate.
Sponsor (17:23):
Well, friends, it is time to let go of the old way of
exploring your data. It'sholding you back. But what
exactly is the old way? Well,I'm here with Mark DePuy, co
founder and CEO of Fabi, acollaborative analytics platform
designed to help big explorerslike yourself. So, Mark, tell me
about this old way.
So the old way, Adam, if you're a a product manager or
(17:44):
a founder and you're trying toinsights from your data, you're
wrestling with your Postgresinstance or Snowflake or your
spreadsheets, or if you are andyou don't maybe even have the
support of a data analyst ordata scientist to help you with
that work. Or if you are, forexample, a data scientist or
engineer or analyst, you'rewrestling with a bunch of
different tools, local JupyterNotebooks, Google Colab, or even
(18:07):
your legacy BI to try to buildthese dashboards that someone
may or may not go and look at.In this new way that we're
building at ABBYY, we arecreating this all in one
environment where productmanagers and founders can very
quickly go and explore dataregardless of where it is. It
can be in a spreadsheet, can bein Airtable, it can be in
Postgres, really easy to doeverything from an ad hoc
(18:29):
analysis to much more advancedanalysis if, again, you're more
experienced. So with Pythonbuilt in right there, in our AI
assistant, you can move veryquickly through advanced
analysis.
And the really cool part is thatyou can go from ad hoc analysis
and data science to publishingthese as interactive data apps
and dashboards, or better yet,delivering insights as automated
(18:53):
workflows to meet yourstakeholders where they are in,
say, Slack or email orspreadsheets. So if this is
something that you'reexperiencing, if you're a
founder or a product managertrying to get more from your
data or from your data teamtoday and you're just underwater
and feel like you're wrestlingwith your legacy, you know, BI
tools and and notebooks, comecheck out the new way and come
try out Fabi.
There you go. Well, friends, if you're trying to get
(19:14):
more insights from your data,stop wrestling with it. Start
exploring it the new way withFabi. Learn more and get started
for free at fabi.ai. That'sfabi.ai.
Again, fabi.ai.
Daniel (19:34):
Yeah. Chris, I I think you were starting to get into
maybe something we've alludedto, which is the the other side,
the negative side of saying, no,we're not in a bubble
corresponding to this firstargument of the speculative
investing, which is that, youknow, maybe the earnings or the
(19:55):
kind of business fundamentals orthe scale is, you know, the
diversification is a bitstronger, in this particular,
time than in the .com time. So,you know, on the one side,
Jerome Powell talking about realearnings, right? Which is maybe
different from some of the .comera. On the other side, there
(20:19):
does seem to be a diversifiedset of revenue streams in the AI
space.
So it's not just AI models, forexample, there's infrastructure
related to this, there's chips,right? GPUs and even like unique
types of chips that are beingdeveloped for AI. There's
(20:40):
service offerings, on top of AI.We've talked a lot on this show
about how the service providers,the large consultancies are
doing quite well in the AIspace. And those things are
things that are already scaling,whether that's chips or the
service offerings, etcetera,cloud offerings around these
(21:03):
things.
And I guess this is something Iwasn't aware of and maybe it's
connected to what you weretalking about before, but the
magnitude of the investmentrelative to kind of the GDP is
still relatively low compared toother kind of, if you wanna
think back to actual, othertransformative revolutions like
(21:28):
railroads or electrification andthat sort of thing. So this kind
of counter argument would be,no, there's really something
more here in terms of theearnings and business
fundamentals with a lot of theseAI companies where there's
earnings, there'sdiversification and the
magnitude of the investment iskind of different.
Chris (21:51):
I think I agree with that though, once again, kind of on
the other, if you're counteringthat just a little bit, then I
think that you're gonna findthat there's a selection of
companies that absolutely fitthat profile that you just
outlined and stuff. But I thinkthat there's still also quite a
few out there to balance thatthat have no earnings and stuff
(22:13):
like, know, very little and thatthe business model is is still
quite questionable and thusleading into kind of that hype
based speculation and stuff. SoI think I definitely feel like
we're seeing both sides of that.And it kinda comes back to
something I mentioned at thebeginning of the call of like,
(22:34):
you know, the the the bubbleesque nature for if we're if
we're in that kind of .cominternet, period, as an event,
you know, like, as ourcontextual reference, that it
doesn't quite fit that. And itit seems to be very much in,
like, how you're engaged andwhat your, you know, what your
ideas, do you have customers,are you providing services, and
(22:56):
thus have earnings that thatsupport that versus the ones
that are not.
And and I think, you know, we'vespent so much time on the show
over the years really trying tocut through the hype cycle.
Sometimes quite literally, youknow, we'll get the it will
we'll get the current yearpublished hype cycle and start
talking about some of that. AndI think if you're deeply
(23:19):
engaged, as we and our listenersand the people who listen to the
show regularly are, it'sprobably easier to do that. But
I also think that there'sprobably a lot of folks out
there that are doing investingthat don't have that don't have
as much, you know, knowledge ofthat. I had a conversation
earlier today with someone whowas that person and was kind of
(23:40):
asking about some stuff and Iwas kinda doing a little bit of
mentoring maybe.
But I realized just, you know,that there are a lot of people
out there that still reallydon't know much about it and all
they hear is the hype and theythey have very little ability to
get through it. So, you know, isif they're looking to invest,
it's almost a a little bit of aflip of the of the coin on
(24:03):
whether or not they look intothe profile that you were just
describing or that other onewhere it's a little bit less
substantial.
Daniel (24:10):
Yeah, I think it's definitely good points. I think
part of the reason here thatthings are hard to parse
through, which we've also talkedabout on the show is just that
it's kind of hard to pin down atthis point what AI means and
what actually part of what wouldbe considered the bubble and
(24:34):
what is not. Part of that isthere's a lot of companies that
are trying to ride the hypecycle, right? And their product
really is not AI powered at all,but they feel the pressure to
tack on this is an AI poweredthing. Maybe they have a linear
regression model or even a rulesbased thing and say, this is AI
(24:58):
and they're writing that hypecycle.
And so on the one side, they'rekind of writing that. On the
other side, there's verysophisticated, whatever,
computer vision systems andother things that maybe are not
viewed as part of the AI hypecycle because they're not
generative AI or something likethat. And so it's kind of hard
(25:22):
to tell what fits there. And inaddition, no one really knows at
the application layer what theend highest value things that
are gonna come out of the AIworld are gonna be I I you know,
we've talked about this on theshow. I certainly don't think
it's a general chat interface.
There's much more valuablethings already, in terms of some
(25:46):
of the agentic and verticalizedthings. And so there's a lot of
just diversification, both interms of people trying to ride
the wave, but also in terms ofdefining what is AI and what is
not, because it could range froma chip producer that's making a
very unique chip that isspecialized for AI workloads all
(26:08):
the way to a very thin wrapperon top of the OpenAI API to a
proprietary computer visionmodel that's taken twenty years
to develop to a SaaS platformthat has actually no AI
component but is labeled AIbecause they can sell it for
(26:30):
more money. Maybe all of thatjust feeds into this bubble, but
also I think it creates a lot ofconfusion, which maybe is not, I
guess in thinking about it, itis similar to that .com era
because it's like everythingrelated to the web, right? And
(26:50):
you did have servers, you didhave hardware, you did have kind
of websites or platforms andthat sort
Chris (26:56):
of thing. Yeah. I mean, even then people were buying
hardware and platforms andsoftware left and right during
that period. But again, the kindof the winners out of that .com
era was a fairly small group ofcompanies that were feeding the
purchase frenzy during that. AndI think that is a similarity we
(27:19):
have to today where you have asmall group of companies that
are providing a lot of ofcapability and stuff.
I think, you know, when it comesto the others, you know, in in
that, like, you just identifiedjust, know, with the the AI
label being so marketable andyet having almost no meaning
because of the the immensediversity of possibilities that
(27:42):
you could apply AI labeling to.It really comes down to solving
business problems that are realbusiness problems as opposed to
trying to put an AI thing outthere. You know, and I think I
know that the the you and I inour in our roles hosting the
(28:03):
podcast here, we get pitched bya lot of companies, a lot. And
so we see a lot of a lot ofdifferent, you know, positions
and possibilities out there. AndI think, you know, if you look
at that, probably the ones thatreally catch our attention are
the ones that aren't the mostglammy AI things necessarily in
(28:25):
terms of how they're marketing,but where you can really see
that they're using thesetechnologies to solve business
problems in novel ways thathadn't been addressed before.
And I think, you know, that'sprobably, you know, the basis
for our own way of doing theseevaluations. I wonder, you know,
how the general population whois not just living and breathing
(28:48):
AI every minute of every day,how they're looking at some of
these different things that arecoming at them in every
advertisement and marketingeffort. You know, how do they
how do they tell the difference?You know? I I think I think that
has a lot to do with withbubbles as well.
It's just that inability tounderstand the difference
between a great investment andsomething that is really, really
(29:11):
sketchy and risky and not beingable to tell the difference
between the two.
Daniel (29:15):
Yeah. Just just anecdotally, it kind of has gone
I I remember I don't know if youhad this similar experience,
Chris, but obviously we've beendoing this show for quite some
time. We talk about AI here. Forthe most of the time of this
podcast, we didn't really talkabout AI or AI was not a kind of
(29:36):
topic of general discussion andwas not represented kind of, I
don't know, in the environmentsin which you walk through day to
day, like a airport orsomething. I remember seeing the
first, I think it was anAnthropic ad when I was in, I
forget what airport, it was anairport in Europe somewhere.
(29:57):
And I saw an Anthropic ad and Iwas like, woah, there's an AI.
And you know, I don't live inSan Francisco. San Francisco is
different. You go to SanFrancisco, it's all like on the
billboard is advanced monitoringfor your Kubernetes cluster
billboard, which you know,wouldn't work anywhere else in
The US. But outside of outsideof, San Francisco, it's like you
(30:20):
don't you don't see that sort ofthing, right?
And so I I thought, oh man, likethis is crazy. Now and that was
in like a major airport in somehub in Europe. And I'm like,
okay, well, that makes sense.But I just got back from a trip
this morning and as I waswalking through the Indianapolis
Airport, there were multiplebillboards and banners that I
(30:43):
saw that all had some AI slantrelated to a product or service.
So it's pervasive now and it iskind of the soup we live in.
Chris (30:55):
It is. So as we are trying to navigate that world,
think it's funny, you mentionedhow long we've been doing this
and a lot of our listeners thathave been with us for years,
going through this evolutionwith us are probably seeing
similar, you know, takes on thisand that they're they're
(31:15):
watching it. But I've also I'vealso come to realize that
there's still quite a massivesegment of our general
population that is only in theselast few months really becoming
aware of this stuff, and arestill trying to take it in early
on. And so they are they aresuddenly I was I mentioned that
(31:39):
conversation I had earliertoday. And that individual said,
this AI thing seemed to havecome out of nowhere.
And you know, that's about asfar from my experience as as you
could possibly be. But I made merealize that that's, that's
quite common. And more recently,I mentioned, you know, that that
my mom in her mid 80s was wasnow talking about AI and stuff.
(32:01):
So times are changing quickly.
Sponsor (32:17):
What if AI agents could work together just like
developers do? That's exactlywhat agency is making possible.
Spelled AGN, TCY, agency is nowan open source collective under
the Linux Foundation, buildingthe Internet of Agents. This is
a global collaboration layerwhere the AI agents can discover
each other, connect, and executemulti agent workflows across any
(32:42):
framework. Everything engineersneed to build and deploy multi
agent software is now availableto anyone building on agency,
including trusted identity andaccess management, open
standards for agent discovery,agent to agent communication
protocols, and modular piecesyou can remix for scalable
systems.
This is a true collaborationfrom Cisco, Dell, Google Cloud,
(33:07):
Red Hat, Oracle, and more than75 other companies all
contributing to the next gen AIstack. The code, the specs, the
services, they're dropping. Nostrings attached. Visit
agency.org. That's agntcy.org tolearn more and get involved.
Again, that's agency,agntcy.org.
Daniel (33:34):
Well, Chris, I think one, if we just kind of bring
out another argument here thatpeople are making at this time,
and I'm curious to know actuallyyour opinion on this. Might as
well give some of our ownopinions or else what else are
we doing here? That seems likefun. But one the arguments for
(33:59):
the fact that we are not in anAI bubble is that the kind of
rationale, the structural andeconomic rationale for AI
deployment is much deeper thanprevious bubbles. I think the
hyperbolic example would be likecrypto maybe, but the fact that
(34:19):
AI is already being integratedinto enterprise workflows, into
even manufacturing andhealthcare.
It's not just kind of atechnology that is looking for
an an interesting technologythat's looking for an
application, but it is actuallybeing applied across a variety
(34:42):
of industries. The second pieceof this would be kind of the tie
of AI to a very long lead up ofscientific research and deep
roots, right? That have beengoing on since the sixties,
seventies, eighties, and on, andhad all of these things leading
(35:05):
up to it, which gives it kind ofroots in kind of rigorous
science and mathematics and thatsort of thing. The technology
didn't kind of come out ofnowhere, similar to kind of the
revolution of electrification,right? People had actually been
studying these topics or atleast thinking about them for
(35:28):
centuries, right?
Thinking about, these phenomena,even though maybe they weren't
fully understood and then kindof there were breakthroughs that
created this electrification ofthe world. Right?
Chris (35:43):
Totally. One of the things to your point right there
that actually gives me a littlebit of confidence maybe that
we're not in a classic bubble,at least in the .com context, is
the fact that the fundamentalunderlying, algorithmic
technology that we're looking athere today in 2025 as we do the
(36:06):
show is actually about 40 yearsold is is the the the basis
being neural networks and and itit while we don't use that
phrase as much as we used toanymore, you know, we we call
everything AI these days, butfundamentally the AI that is
powering everything that is ofsignificant value of today is
(36:29):
fundamentally neural networksthat have been enhanced and
embellished and designed furtherand you know next iteration that
kind of thing and so we'rebuilding on a forty year history
of this particular line ofalgorithmic technology And so,
you know, that that's a greatpoint you raised there a second
ago that this is not fly bynight, that there has been a it
(36:53):
it you know, the majority forme, you know, the majority of my
lifetime, this has been around.For listener, just as a two
second thing for listeners whomay not have followed the show
for years, because I know I'vementioned this before, but my
parents were working on neuralnetworks back in the in the
early nineties.
And so like, and that's actuallywas my first exposure to this
(37:17):
when I was in college. And sothis is not new stuff. It's
evolved. And I think the thebiggest piece of it is you've
had, you've had Nvidia's comeabout able to make GPUs that
could support the continuingevolution of the technology. We
it is thoroughly embedded in abunch of industries and some
(37:38):
implement like anything, someimplementations are better than
others.
But yeah, great point you'remaking in terms of that this is
not a solution in search of agrounding or in search of a
market to use it.
Daniel (37:51):
Yeah. Yeah. And I I also tend to agree now, of course,
I'm biased on this show andmaybe my bias is not that
surprising. I think some of thevaluations and the way people
are treating investment in thisarea is crazy. Not to say that
you can't invest in PredictionGuard, and I'll talk to you, but
(38:17):
I do think that generally therehave been many examples of
craziness there.
At the same time, I think thosecrazy investments are partially
founded, maybe not in kind ofthe earnings or like a proven
(38:37):
business model, but in a kind ofdeeper understanding that this
technology is actually shiftinghow work is done and is
fundamentally transformative formany industries. And that
actually is going to happen. Andso it's not, it's maybe
speculative in a certain way,but I think there are real kind
(38:58):
of foundations to thatspeculation that, I don't know,
I would almost use the metaphor,it's kind of like when people go
out and look for gold or oil,right? The concept of gold or
oil is known and it's known thatthere is value if you get this
(39:21):
material out of the ground,right? But there's speculation
and risk in trying to figure outwhere that is.
Here, it's not like you'retrying to discover a new
precious metal or a new thingthat's not known to anyone,
right? You know this thingexists, it's a matter of finding
it. And certainly many peoplelose a lot of money in that
(39:44):
speculation. Here it's similar.There is a kind of known, or
there is at least a feeling andan intuition that there will be
very transformative companiesthat will be long lasting and
this technology will be longlasting and it will be
impactful.
It's not that that's maybedoubted, but the crazy
(40:08):
valuations are in some case aredriven by that because, who
knows who's going to survivethis AI craziness. But the
technology it seems will be kindof pervasive and long lasting
and transformative.
Chris (40:24):
I think so. I think I'll throw, I'm gonna throw a wild
card into mix here for a moment.And one of the things that no
one really knows where thingsare going for sure, and we sure
talk about it a lot, is the factthat at this point, you have a
lot of of companies, especiallysome very large companies, such
(40:46):
as Amazon, are doing massivelayoffs, You know? And they're
doing that on the notion, inAmazon's case, 14,000 middle
managers. And that's just onethat's just one organization.
One very large organization, butone, and this is happening in a
lot of places, is that, youknow, we're definitely seeing
(41:09):
the replacement efforts bycompanies to use technology to
replace humans in that. And andthis has been something we've
talked about for years, youknow, certainly coming and that
would be happening and, youknow, where would the balance be
between a synergy between AI andhumans and a competition between
AI and humans. And so we've hadmany conversations over the
(41:31):
years about this. But, you know,we're definitely at that point
right now where a lot ofcompanies are beginning to bet
on AI technologies. But it'salso happening at a time when,
as we talked about, some ofthose ROIs on different efforts
are are not yielding meaningfulresults.
(41:53):
And so aside from what happenswith that, when you have
unemployment rising from AIinduced layoffs, that will
affect the economy too. Sothere's it's not just whether or
not the investment in theseparticular companies is is wise
(42:13):
or speculative and, you know,based on their fundamentals and
their earnings, but also, youknow, as as we have groups of
folks in the economy being laidoff and therefore their
purchasing power is reduced, howdoes that play back in? I've
seen a lot in the news about,like, all of this together, not
only the ROI and and speculativenature or lack or, you know,
(42:36):
whether it's a bubble or not,going back to our our kind of
original phraseology, but alsowhether or not this is going to
affect work phrases. So it's itis quite a complex thing and
when we look back on .com itseemed there were complexities
there but I think the the rawspeculation of that era made the
bubble. It was a little bit moreof a black and white thing as
(42:57):
you looked back on it, you know,historically.
After we had lived through itand looked back and kinda said,
well, yeah, you know, I guesswith 2020 hindsight, we can see
that coming. As we've notedhere, it kinda depends on who
you are and how you're doing andwhat you're claiming as your AI
and whether you're solving areal business problem on whether
(43:18):
or not you're in that bubblegroup or not in a bubble group.
So as we see the mixture pouringthrough the economy, it will be
definitely interesting to seehow this plays out in the weeks,
months and years ahead here.
Daniel (43:33):
Do you think that because we also talked recently
on the show about almost theboth cognitive and emotional,
you know, change and shift oreven manipulation that some of
these systems are doing, youknow, across the population if
you look at it, there's peoplehaving, you know, romantic
(43:55):
relationships with AI systemsor, you know, using these
systems maybe for therapeuticpurposes rather than their
therapist and cognitive load ofwork is changing because you're
vibe coding and all of thesethings. So do you think that
(44:16):
that sort of cognitive andemotional, almost like lifestyle
shift is more impactful withthis technology or with
something like the .com era andpeople kind of all coming
online?
Chris (44:30):
You know, .com era was and and sadly, I was well into
adulthood, you know, as we hitthat, you know, you know, for
those who don't realize that I'mI'm getting a little bit older
than I like to imagine. But,like, fundamentally, we went
through the bus because of allthe speculation, but eventually,
everything kind of you know, wewe did realize what it was going
(44:52):
at. It just wasn't on a atimeline and the and what you
could achieve in that shortertime given the valuations just
wasn't real. But it was real inthe long term. And I think here
what we're seeing is something alittle bit different in that you
weren't having relationshipswith your with your ISP at the
time.
(45:13):
You know, it wasn't that kind ofa relationship that you had with
that wave of technologicalinnovation. This one, it's it's
a little bit worrisome. Like, II don't believe in vibe coding
being a great strategypersonally. I I think that
thinking of AI as a pairprogramming partner is a much
sounder way of approaching it interms of turning out continuing
(45:35):
to turn out very good softwareproducts that you understand and
that you can maintain over timeand that you have humans that
that understand how theirbusiness is working. So I'm a
little pure vibe coding wheresomeone who doesn't really know
what they're doing is justasking the system and then they
end up with something.
Know, so it kinda depends onwhat you're doing with it. I see
(45:56):
there's also been some researchand we may have an episode
coming up on it that was donerecently about the human
dependence on AI causingbasically degradation. They were
measuring brain activity and andp and the subjects that were in
this study that were using theirAI for everything were showing
(46:17):
decreased brain capability overtime. And so it worries me that
we're giving up some of whatmakes us so wonderfully human At
the same time that maybe we'recreating a strong dependence
that we have in these ways thatdidn't exist in previous
revolutions. So I think I thinkthat there's a real risk here of
(46:39):
I think how you how you use AIcapabilities today makes a
difference on what your ownpersonal human future is going
to be.
And I very specifically choosehow I use AI capabilities in a
way that enables me versus kindof creates a crutch for me. So
(47:02):
that's that's kinda how I wouldanswer. That's a little bit
roundabout way of answeringthat. But I think the way you
use AI has a lot to do with itslong term effect on on your your
own personal life and as anindividual.
Daniel (47:15):
Makes sense. Yeah. I think that's a that's a great
way to look at it. Well, as weclose out here, Chris, what
would you say? Are we in an AIbubble or not?
What's your what's your vote?Yes or no?
Chris (47:26):
I'm gonna say no in the classic bubble analogy, I think.
And and I I came into thisconversation not really knowing.
I think it's this us talking itthrough. So I would say no in
the classic bubble context like.com bubble. But yes, and that
there might be instead of onebig bubble, there might be lots
of little bubbles.
Daniel (47:47):
Yeah, a fizzy like our like our PIP zero and Coke zero
that we that we started talkingabout.
Chris (47:54):
Yeah, there you go. The fizzy economy. You heard it here
first folks.
Daniel (47:58):
The fizzy AI soup. On my end, I would tend to go with the
no. I definitely think thatcompared to some other
technologies and cycles, this isalready kind of at a level of
utility and permeating manyenterprises and that sort of
(48:18):
thing. And I do think thatcreates a trickle down of things
that we will have to deal withand learn how to cope with, you
know, workforce wise andemotionally and otherwise. But,
yeah, I'll I'll go with no aswell.
So so you heard it here. NewYork Times, you can quote us if
you like. We are not in an AIbubble. Because Dan and Chris
(48:39):
said so. I'm glad that we putthat one to rest, Chris.
Chris (48:42):
Yeah. Well, there we go. Solving world problems one
episode Yeah. At a
Daniel (48:46):
Well, I guess until next time, we can solve the next
world crisis in the nextepisode, Chris. Thanks for
chatting.
Chris (48:54):
Absolutely. Talk to you next time.
Jerod (49:03):
All right. That's our show for this week. If you
haven't checked out our website,head to practicalai.fm and be
sure to connect with us onLinkedIn, X, or Blue Sky. You'll
see us posting insights relatedto the latest AI developments,
and we would love for you tojoin the conversation. Thanks to
our partner Prediction Guard forproviding operational support
for the show.
Check them out atpredictionguard.com. Also,
(49:26):
thanks to Breakmaster Cylinderfor the beats and to you for
listening. That's all for now,but you'll hear from us again
next week.
Welcome to the Practical AI podcast, where we break down
the real world applications ofartificial intelligence and how
it's shaping the way we live,work, and create. Our goal is to
help make AI technologypractical, productive, and
accessible to everyone. Whetheryou're a developer, business
leader, or just curious aboutthe tech behind the buzz, you're
(00:24):
in the right place. Be sure toconnect with us on LinkedIn, X,
or Blue Sky to stay up to datewith episode drops, behind the
scenes content, and AI insights.You can learn more at
practicalai.fm.
Now onto the show.
Sponsor (00:40):
Well, friends, when you're building and shipping AI
products at scale, there's oneconstant, complexity. Yes.
You're bringing the models, datapipelines, deployment
infrastructure, and then someonesays, let's turn this into a
business. Cue the chaos. That'swhere Shopify steps in whether
you're spinning up a storefrontfor your AI powered app or
(01:00):
launching a brand around thetools you built.
Shopify is the commerce platformtrusted by millions of
businesses and 10% of all USecommerce from names like
Mattel, Gymshark, to foundersjust like you. With literally
hundreds of ready to usetemplates, powerful built in
marketing tools, and AI thatwrites product descriptions for
(01:21):
you, headlines, even polishesyour product photography.
Shopify doesn't just get youselling, it makes you look good
doing it. And we love it. We useit here at Changelog.
Check us outmerch.changelog.com. That's our
store front, and it handles theheavy lifting too. Payments,
inventory, returns, shipping,even global logistics. It's like
(01:43):
having an ops team built intoyour stack to help you sell. So
if you're ready to sell, you areready for Shopify.
Sign up now for your $1 permonth trial and start selling
today atshopify.com/practicalai. Again,
that is shopify.com/practicalai.
Daniel (02:17):
Welcome to another episode of the Practical AI
Podcast. In these fullyconnected episodes where it's
just Chris and I without aguest, we like to explore some
topics from trending AI news oror things that are being
discussed in the community.Hopefully, we learn along the
way and and our listeners learnalong the way, and and we help
(02:39):
you all level up your AI andmachine learning game. I'm
Daniel Witenack. I am CEO atPrediction Guard, and I am
joined as always by my cohost,Chris Benson, who is a principal
AI research engineer at LockheedMartin.
How you doing, Chris?
Chris (02:54):
Hey. Doing good today, Daniel. Just cruising along,
chewing some bubble gum, doingmy thing.
Daniel (03:02):
We both needed a sugary, but not sugary caffeinated
drink. I think you've got a Pibbzero. I've got a Coke zero.
Chris (03:12):
There you go.
Daniel (03:13):
It's been a long day of work and now we get to talk
about some fun things.Hopefully, we won't pop the AI
bubble today, but, certainly
Chris (03:25):
That's okay. We'll we'll clean up the situation if we
have to.
Daniel (03:29):
Yeah. Exactly. It it's interesting, Chris. I was
looking around. Of course, we'realways seeing things or people
are forwarding us various thingsin the news related to AI.
One this week that I saw wasmaybe an kind of interesting one
with a humanoid looking robot,but I'd rather maybe discuss
(03:49):
this other stuff, which I thinka couple of these things that I
saw connected. And Chris, alittle while ago, we kind of did
a hot takes and debates typeepisode. I think today, there's
certainly on this topic, thereare various sides of this topic
and strong opinions about it. Wecan just talk through some of
(04:11):
those. Certainly, there is adebate going on.
And, this is really thequestion, are we in an AI
bubble? Which seems to be talkedabout all the time. Like people
see something in the news andthey're like, oh, we're
definitely in an AI bubble. Orpeople are like, oh, this isn't
an actual AI bubble. It'sdifferent than maybe the .com
(04:34):
bubble.
Chris, you hear similaranecdotes and and maybe what is
your maybe just the generalconcept of of bubble. We're not
talking about bubble gum, but
Chris (04:45):
Yeah. I mean, the general notion of a bubble, you know, is
it's really a financial conceptwhere the valuation of an
organization exceeds what itsactual value is. In other words,
what it's putting out in termsof product and services, and the
returns that are yielded bythose. And so, you know, and,
(05:08):
you know, if you were talkingfor instance about the .com
bubble and for those of you whomay not have depending on if
you're in the younger skew ofour audience, around 2,000 ish
in in that there was all thisinternet craze and rage and hype
not dissimilar from the hypewe've been seeing in recent
(05:30):
years over AI And a lot ofcompanies came about and new
startups and stuff, and they gotvalued very, very high, but they
had very little, in some cases,no revenue and thus no profits
available and so big valuationby the market with absolutely
(05:50):
nothing coming out of them ofvalue and so it there's a point
where the market kind ofrealizes that and corrects and
and the the giant.com, you know,facade came falling down and
kinda led into a recessionaryperiod around the globe.
And so it was kind of a bigthing over several years. And so
(06:12):
through this entire period of ofAI buildup, that's been a
concern that has come upregularly. This is not the first
time, you know, we've beenhearing about AI bubbles and
stuff. So, I don't think there'sa year that's gone by that we've
been doing this podcast where ithasn't been raised as an issue.
What do you think?
Daniel (06:30):
Maybe so. Yeah. And I think there is, you know,
genuine concern because previousbubbles that have burst, have
actually caused real harm. Asyou mentioned, whether that's
economic kind of recession,certainly it distorts the way
people invest or maybe what theyinvest into, which kind of has
(06:54):
an effect potentially onretirement or 401ks. There's a
reduction in trust in certaintypes of organizations or
financial institutions ortechnology companies, that sort
of thing.
And it could be companies, itcould be assets. Like, a lot of
(07:15):
times people talk about cryptoas an asset that Or I remember
not too long ago talking to tonsof people about NFTs, right? And
this really chaotic time withNFTs and there's a lot of people
that lost a lot of money inthat. Indeed. So it is a valid
(07:36):
concern and I think the questionthat's on people's mind is, are
we in an AI bubble?
And one of the interestingarticles that I saw this week,
Chris, was that Powell, theFederal Reserve Chair, Jerome
(07:56):
Powell in The US, for those thataren't listening from The US,
that's the Federal ReserveChair, Jerome Powell, often can
kind of anything that's saidabout the economy by whoever's
in this position is taken with alot of weight because, well,
(08:20):
Chris, you may have comments onthis, but I'm not an economist.
But generally it sign is of atleast an intelligent opinion on
things where there's a lotthat's gone on or a direction
that, you know, the FederalReserve wants people to think.
Chris (08:39):
So so the two second background without going off on
a a a a track is Jerome Powellin his capacity is responsible
for a particular committee atthe Federal Reserve, and they
are responsible for monetarypolicy, which the Federal
Reserve sets as opposed to whatthe president and congress
(09:00):
together set, which is fiscalpolicy. And one of the tools,
the largest tool that theymainly do that with is through
the setting of interest rateswhich trickles through the
entire economy in a whole bunchof different ways. I don't wanna
go into any more depth thanthat, but they they thus slow
down or speed up the economy toeither tackle an underperforming
(09:23):
economy or inflation on theopposite side, and they're
trying to balance it between thetwo. And so, yeah, chairman
Powell noted that there wasactually revenue associated with
AI expenditure. And thus, itwasn't bubble like in his view,
when he is certainly an expertin a lot of ways, think I think
(09:44):
that there are some otherconsiderations in there.
But it's definitely a powerfulstatement coming from that
particular individual.
Daniel (09:52):
Yeah, yeah. The title of the article in Fortune which
we'll link is Powell says thatunlike the .com boom, AI
spending isn't a bubble. And thequote is, I won't go into
particular names, but theyactually have earnings. Now I
guess we could speculate on thispodcast. I don't know what names
he's talking about.
I'm assuming it's some of theselarger names that would be
(10:17):
whatever OpenAI, Anthropic,Cohere, you know, whatever the
ones are that people would thinkof. I imagine some of those are,
quote, the names. But he saysthey sort of actually have
earnings. The other interestingpiece that I saw, Chris, which
is maybe on the other side ofthis argument, and we can talk
(10:40):
maybe after this about, youknow, the different ways that
people argue that we are oraren't in an AI bubble. But this
other one was from the New YorkTimes, which is reporting that
NVIDIA is now worth$5,000,000,000,000 and this is
the quote from the article, asit consolidates power in AI
(11:01):
boom.
So it's the subheading, the AIchipmaker has become a linchpin
in the Trump administration'strade negotiations with Asia. So
there's there's a, you know,some policy and political angle
to this article, but the generalidea with the article is that
this $5,000,000,000,000valuation maybe is is part of an
(11:25):
AI boom. So yeah, that'scertainly interesting.
Chris (11:30):
It is. And I think we're seeing that, going back to those
names that were left unsaid, younamed a few of those names. And,
you know, I think any of thelarge cloud service providers
that are offering a collectionof AI services probably, you
know, round out some of thosenames. One of the things, one of
(11:53):
the distinctions that I think isinteresting to to weigh as we
talk about this is the fact thatwhile he noted, Mr. Powell noted
that some of these organizationshave earnings and we are seeing,
you know, that reflected in AIrelated stocks that are just
(12:13):
dominating S and P 500 returnsto the to the degree of about
75% of those returns.
80% of earnings growth, and 90%of capital spending growth. I
mean, are phenomenal numberswhen you think about that in
terms of percentages of totalmarkets that are out there, you
know, or of these at leastexchanges that are representing
(12:35):
markets. So I mean, that'sreally shocking, but I think one
of the things that is that thathe did not say, speaking as
someone who has nowhere near theeconomic expertise of mister
Powell, but did did have someuniversity studies in economics
that you know, those thosenumbers are concentrated in a
(12:58):
tiny fraction of companiesoverall. And so my question that
I would ask is for for some ofthose giant companies that are
that are making huge earningsfrom many, many, many thousands
of customers spending hugeamounts of money on AI growth,
We have also talked about thefact that there is there's also
(13:23):
in the media quite a bit ofquestioning of ROI on a lot of
those AI investments at variouscompanies. And if you're the
cloud provider making bookeys ofmoney, that's great.
But if you're one of thecompanies out there that maybe
is spending on that, but maybenot seeing an ROI on that
(13:44):
expenditure, That is a differentstory right there. And so, you
know, I you know, maybe it'sboom for one and a bubble for
another. Maybe it's not just auniversal bubble, but it depends
on who you're talking to, tosome degree.
Daniel (13:59):
Yeah, and I guess that's where maybe this is murky is
the, some people might definebubble differently. I think that
it is one of the kind of keyarguments for the kind of
affirmative of this that we arein an AI bubble is this kind of
valuations and speculation. Imean, we've highlighted a few
(14:21):
stories over time the lastcouple of years of these crazy
valuations where essentiallythere is an unproven revenue
model. So I'm looking at onesource here that's saying, over
50% of VC funding in Q2 of thisyear, 2025, went into AI
(14:43):
companies. And VC is risky, ofcourse, as an investment model.
Most of those businesses willfail. That's kind of always
expected in the AI space. Butalso this year, just from
looking around at differentfolks raising, these companies
are getting much, much highermultiples or higher valuations.
(15:07):
So for people that need areminder, sometimes your company
might receive a valuation that'sa certain multiple of the
revenue that you're bringing in.And so that might be fifteen,
twenty, 25x, 30x plus for an AIcompany, where other companies,
(15:28):
kind of your normal run of themill SaaS company in tech that's
raising is definitely notraising at those multiples right
now.
So that would be, I guess, theargument or an argument for that
affirmative is these kind ofspeculation and valuation that's
(15:48):
maybe reminiscent of that .comera.
Chris (15:51):
Yeah. By the way, another term just to connect kind of
that financial world with whatwe're talking about in terms of
observations is those of youinvest, some of you may have
heard the notion of a beta,which is essentially a a
multiple of valuation againstyour earnings. And so if you
(16:14):
have a very high beta that'ssaying you're being valued very
high against what your actualreal life earnings are and a
lower beta, which would beconsidered less risky, would be
that that valuation is not soextravagant. And so I guess and
that's another way of of lookingat this is if you're looking at
a company's, you know, portfolioanalysis, you know, that some
(16:37):
analyst is doing, and theremight be a beta number attached
if you're is that and this waspointed out in some of the
articles is that the betas oftoday, while high, are not
nearly as high as the betas ofthe .com era. And so, you know,
that's another sign.
Well, you know, is there somebubble? Maybe. Is it as bad in
(17:00):
terms of the sheer speculationof the .com era? Maybe not. You
know, that's one metric by whichwe can evaluate.
Sponsor (17:23):
Well, friends, it is time to let go of the old way of
exploring your data. It'sholding you back. But what
exactly is the old way? Well,I'm here with Mark DePuy, co
founder and CEO of Fabi, acollaborative analytics platform
designed to help big explorerslike yourself. So, Mark, tell me
about this old way.
So the old way, Adam, if you're a a product manager or
(17:44):
a founder and you're trying toinsights from your data, you're
wrestling with your Postgresinstance or Snowflake or your
spreadsheets, or if you are andyou don't maybe even have the
support of a data analyst ordata scientist to help you with
that work. Or if you are, forexample, a data scientist or
engineer or analyst, you'rewrestling with a bunch of
different tools, local JupyterNotebooks, Google Colab, or even
(18:07):
your legacy BI to try to buildthese dashboards that someone
may or may not go and look at.In this new way that we're
building at ABBYY, we arecreating this all in one
environment where productmanagers and founders can very
quickly go and explore dataregardless of where it is. It
can be in a spreadsheet, can bein Airtable, it can be in
Postgres, really easy to doeverything from an ad hoc
(18:29):
analysis to much more advancedanalysis if, again, you're more
experienced. So with Pythonbuilt in right there, in our AI
assistant, you can move veryquickly through advanced
analysis.
And the really cool part is thatyou can go from ad hoc analysis
and data science to publishingthese as interactive data apps
and dashboards, or better yet,delivering insights as automated
(18:53):
workflows to meet yourstakeholders where they are in,
say, Slack or email orspreadsheets. So if this is
something that you'reexperiencing, if you're a
founder or a product managertrying to get more from your
data or from your data teamtoday and you're just underwater
and feel like you're wrestlingwith your legacy, you know, BI
tools and and notebooks, comecheck out the new way and come
try out Fabi.
There you go. Well, friends, if you're trying to get
(19:14):
more insights from your data,stop wrestling with it. Start
exploring it the new way withFabi. Learn more and get started
for free at fabi.ai. That'sfabi.ai.
Again, fabi.ai.
Daniel (19:34):
Yeah. Chris, I I think you were starting to get into
maybe something we've alludedto, which is the the other side,
the negative side of saying, no,we're not in a bubble
corresponding to this firstargument of the speculative
investing, which is that, youknow, maybe the earnings or the
(19:55):
kind of business fundamentals orthe scale is, you know, the
diversification is a bitstronger, in this particular,
time than in the .com time. So,you know, on the one side,
Jerome Powell talking about realearnings, right? Which is maybe
different from some of the .comera. On the other side, there
(20:19):
does seem to be a diversifiedset of revenue streams in the AI
space.
So it's not just AI models, forexample, there's infrastructure
related to this, there's chips,right? GPUs and even like unique
types of chips that are beingdeveloped for AI. There's
(20:40):
service offerings, on top of AI.We've talked a lot on this show
about how the service providers,the large consultancies are
doing quite well in the AIspace. And those things are
things that are already scaling,whether that's chips or the
service offerings, etcetera,cloud offerings around these
(21:03):
things.
And I guess this is something Iwasn't aware of and maybe it's
connected to what you weretalking about before, but the
magnitude of the investmentrelative to kind of the GDP is
still relatively low compared toother kind of, if you wanna
think back to actual, othertransformative revolutions like
(21:28):
railroads or electrification andthat sort of thing. So this kind
of counter argument would be,no, there's really something
more here in terms of theearnings and business
fundamentals with a lot of theseAI companies where there's
earnings, there'sdiversification and the
magnitude of the investment iskind of different.
Chris (21:51):
I think I agree with that though, once again, kind of on
the other, if you're counteringthat just a little bit, then I
think that you're gonna findthat there's a selection of
companies that absolutely fitthat profile that you just
outlined and stuff. But I thinkthat there's still also quite a
few out there to balance thatthat have no earnings and stuff
(22:13):
like, know, very little and thatthe business model is is still
quite questionable and thusleading into kind of that hype
based speculation and stuff. SoI think I definitely feel like
we're seeing both sides of that.And it kinda comes back to
something I mentioned at thebeginning of the call of like,
(22:34):
you know, the the the bubbleesque nature for if we're if
we're in that kind of .cominternet, period, as an event,
you know, like, as ourcontextual reference, that it
doesn't quite fit that. And itit seems to be very much in,
like, how you're engaged andwhat your, you know, what your
ideas, do you have customers,are you providing services, and
(22:56):
thus have earnings that thatsupport that versus the ones
that are not.
And and I think, you know, we'vespent so much time on the show
over the years really trying tocut through the hype cycle.
Sometimes quite literally, youknow, we'll get the it will
we'll get the current yearpublished hype cycle and start
talking about some of that. AndI think if you're deeply
(23:19):
engaged, as we and our listenersand the people who listen to the
show regularly are, it'sprobably easier to do that. But
I also think that there'sprobably a lot of folks out
there that are doing investingthat don't have that don't have
as much, you know, knowledge ofthat. I had a conversation
earlier today with someone whowas that person and was kind of
(23:40):
asking about some stuff and Iwas kinda doing a little bit of
mentoring maybe.
But I realized just, you know,that there are a lot of people
out there that still reallydon't know much about it and all
they hear is the hype and theythey have very little ability to
get through it. So, you know, isif they're looking to invest,
it's almost a a little bit of aflip of the of the coin on
(24:03):
whether or not they look intothe profile that you were just
describing or that other onewhere it's a little bit less
substantial.
Daniel (24:10):
Yeah, I think it's definitely good points. I think
part of the reason here thatthings are hard to parse
through, which we've also talkedabout on the show is just that
it's kind of hard to pin down atthis point what AI means and
what actually part of what wouldbe considered the bubble and
(24:34):
what is not. Part of that isthere's a lot of companies that
are trying to ride the hypecycle, right? And their product
really is not AI powered at all,but they feel the pressure to
tack on this is an AI poweredthing. Maybe they have a linear
regression model or even a rulesbased thing and say, this is AI
(24:58):
and they're writing that hypecycle.
And so on the one side, they'rekind of writing that. On the
other side, there's verysophisticated, whatever,
computer vision systems andother things that maybe are not
viewed as part of the AI hypecycle because they're not
generative AI or something likethat. And so it's kind of hard
(25:22):
to tell what fits there. And inaddition, no one really knows at
the application layer what theend highest value things that
are gonna come out of the AIworld are gonna be I I you know,
we've talked about this on theshow. I certainly don't think
it's a general chat interface.
There's much more valuablethings already, in terms of some
(25:46):
of the agentic and verticalizedthings. And so there's a lot of
just diversification, both interms of people trying to ride
the wave, but also in terms ofdefining what is AI and what is
not, because it could range froma chip producer that's making a
very unique chip that isspecialized for AI workloads all
(26:08):
the way to a very thin wrapperon top of the OpenAI API to a
proprietary computer visionmodel that's taken twenty years
to develop to a SaaS platformthat has actually no AI
component but is labeled AIbecause they can sell it for
(26:30):
more money. Maybe all of thatjust feeds into this bubble, but
also I think it creates a lot ofconfusion, which maybe is not, I
guess in thinking about it, itis similar to that .com era
because it's like everythingrelated to the web, right? And
(26:50):
you did have servers, you didhave hardware, you did have kind
of websites or platforms andthat sort
Chris (26:56):
of thing. Yeah. I mean, even then people were buying
hardware and platforms andsoftware left and right during
that period. But again, the kindof the winners out of that .com
era was a fairly small group ofcompanies that were feeding the
purchase frenzy during that. AndI think that is a similarity we
(27:19):
have to today where you have asmall group of companies that
are providing a lot of ofcapability and stuff.
I think, you know, when it comesto the others, you know, in in
that, like, you just identifiedjust, know, with the the AI
label being so marketable andyet having almost no meaning
because of the the immensediversity of possibilities that
(27:42):
you could apply AI labeling to.It really comes down to solving
business problems that are realbusiness problems as opposed to
trying to put an AI thing outthere. You know, and I think I
know that the the you and I inour in our roles hosting the
(28:03):
podcast here, we get pitched bya lot of companies, a lot. And
so we see a lot of a lot ofdifferent, you know, positions
and possibilities out there. AndI think, you know, if you look
at that, probably the ones thatreally catch our attention are
the ones that aren't the mostglammy AI things necessarily in
(28:25):
terms of how they're marketing,but where you can really see
that they're using thesetechnologies to solve business
problems in novel ways thathadn't been addressed before.
And I think, you know, that'sprobably, you know, the basis
for our own way of doing theseevaluations. I wonder, you know,
how the general population whois not just living and breathing
(28:48):
AI every minute of every day,how they're looking at some of
these different things that arecoming at them in every
advertisement and marketingeffort. You know, how do they
how do they tell the difference?You know? I I think I think that
has a lot to do with withbubbles as well.
It's just that inability tounderstand the difference
between a great investment andsomething that is really, really
(29:11):
sketchy and risky and not beingable to tell the difference
between the two.
Daniel (29:15):
Yeah. Just just anecdotally, it kind of has gone
I I remember I don't know if youhad this similar experience,
Chris, but obviously we've beendoing this show for quite some
time. We talk about AI here. Forthe most of the time of this
podcast, we didn't really talkabout AI or AI was not a kind of
(29:36):
topic of general discussion andwas not represented kind of, I
don't know, in the environmentsin which you walk through day to
day, like a airport orsomething. I remember seeing the
first, I think it was anAnthropic ad when I was in, I
forget what airport, it was anairport in Europe somewhere.
(29:57):
And I saw an Anthropic ad and Iwas like, woah, there's an AI.
And you know, I don't live inSan Francisco. San Francisco is
different. You go to SanFrancisco, it's all like on the
billboard is advanced monitoringfor your Kubernetes cluster
billboard, which you know,wouldn't work anywhere else in
The US. But outside of outsideof, San Francisco, it's like you
(30:20):
don't you don't see that sort ofthing, right?
And so I I thought, oh man, likethis is crazy. Now and that was
in like a major airport in somehub in Europe. And I'm like,
okay, well, that makes sense.But I just got back from a trip
this morning and as I waswalking through the Indianapolis
Airport, there were multiplebillboards and banners that I
(30:43):
saw that all had some AI slantrelated to a product or service.
So it's pervasive now and it iskind of the soup we live in.
Chris (30:55):
It is. So as we are trying to navigate that world,
think it's funny, you mentionedhow long we've been doing this
and a lot of our listeners thathave been with us for years,
going through this evolutionwith us are probably seeing
similar, you know, takes on thisand that they're they're
(31:15):
watching it. But I've also I'vealso come to realize that
there's still quite a massivesegment of our general
population that is only in theselast few months really becoming
aware of this stuff, and arestill trying to take it in early
on. And so they are they aresuddenly I was I mentioned that
(31:39):
conversation I had earliertoday. And that individual said,
this AI thing seemed to havecome out of nowhere.
And you know, that's about asfar from my experience as as you
could possibly be. But I made merealize that that's, that's
quite common. And more recently,I mentioned, you know, that that
my mom in her mid 80s was wasnow talking about AI and stuff.
(32:01):
So times are changing quickly.
Sponsor (32:17):
What if AI agents could work together just like
developers do? That's exactlywhat agency is making possible.
Spelled AGN, TCY, agency is nowan open source collective under
the Linux Foundation, buildingthe Internet of Agents. This is
a global collaboration layerwhere the AI agents can discover
each other, connect, and executemulti agent workflows across any
(32:42):
framework. Everything engineersneed to build and deploy multi
agent software is now availableto anyone building on agency,
including trusted identity andaccess management, open
standards for agent discovery,agent to agent communication
protocols, and modular piecesyou can remix for scalable
systems.
This is a true collaborationfrom Cisco, Dell, Google Cloud,
(33:07):
Red Hat, Oracle, and more than75 other companies all
contributing to the next gen AIstack. The code, the specs, the
services, they're dropping. Nostrings attached. Visit
agency.org. That's agntcy.org tolearn more and get involved.
Again, that's agency,agntcy.org.
Daniel (33:34):
Well, Chris, I think one, if we just kind of bring
out another argument here thatpeople are making at this time,
and I'm curious to know actuallyyour opinion on this. Might as
well give some of our ownopinions or else what else are
we doing here? That seems likefun. But one the arguments for
(33:59):
the fact that we are not in anAI bubble is that the kind of
rationale, the structural andeconomic rationale for AI
deployment is much deeper thanprevious bubbles. I think the
hyperbolic example would be likecrypto maybe, but the fact that
(34:19):
AI is already being integratedinto enterprise workflows, into
even manufacturing andhealthcare.
It's not just kind of atechnology that is looking for
an an interesting technologythat's looking for an
application, but it is actuallybeing applied across a variety
(34:42):
of industries. The second pieceof this would be kind of the tie
of AI to a very long lead up ofscientific research and deep
roots, right? That have beengoing on since the sixties,
seventies, eighties, and on, andhad all of these things leading
(35:05):
up to it, which gives it kind ofroots in kind of rigorous
science and mathematics and thatsort of thing. The technology
didn't kind of come out ofnowhere, similar to kind of the
revolution of electrification,right? People had actually been
studying these topics or atleast thinking about them for
(35:28):
centuries, right?
Thinking about, these phenomena,even though maybe they weren't
fully understood and then kindof there were breakthroughs that
created this electrification ofthe world. Right?
Chris (35:43):
Totally. One of the things to your point right there
that actually gives me a littlebit of confidence maybe that
we're not in a classic bubble,at least in the .com context, is
the fact that the fundamentalunderlying, algorithmic
technology that we're looking athere today in 2025 as we do the
(36:06):
show is actually about 40 yearsold is is the the the basis
being neural networks and and itit while we don't use that
phrase as much as we used toanymore, you know, we we call
everything AI these days, butfundamentally the AI that is
powering everything that is ofsignificant value of today is
(36:29):
fundamentally neural networksthat have been enhanced and
embellished and designed furtherand you know next iteration that
kind of thing and so we'rebuilding on a forty year history
of this particular line ofalgorithmic technology And so,
you know, that that's a greatpoint you raised there a second
ago that this is not fly bynight, that there has been a it
(36:53):
it you know, the majority forme, you know, the majority of my
lifetime, this has been around.For listener, just as a two
second thing for listeners whomay not have followed the show
for years, because I know I'vementioned this before, but my
parents were working on neuralnetworks back in the in the
early nineties.
And so like, and that's actuallywas my first exposure to this
(37:17):
when I was in college. And sothis is not new stuff. It's
evolved. And I think the thebiggest piece of it is you've
had, you've had Nvidia's comeabout able to make GPUs that
could support the continuingevolution of the technology. We
it is thoroughly embedded in abunch of industries and some
(37:38):
implement like anything, someimplementations are better than
others.
But yeah, great point you'remaking in terms of that this is
not a solution in search of agrounding or in search of a
market to use it.
Daniel (37:51):
Yeah. Yeah. And I I also tend to agree now, of course,
I'm biased on this show andmaybe my bias is not that
surprising. I think some of thevaluations and the way people
are treating investment in thisarea is crazy. Not to say that
you can't invest in PredictionGuard, and I'll talk to you, but
(38:17):
I do think that generally therehave been many examples of
craziness there.
At the same time, I think thosecrazy investments are partially
founded, maybe not in kind ofthe earnings or like a proven
(38:37):
business model, but in a kind ofdeeper understanding that this
technology is actually shiftinghow work is done and is
fundamentally transformative formany industries. And that
actually is going to happen. Andso it's not, it's maybe
speculative in a certain way,but I think there are real kind
(38:58):
of foundations to thatspeculation that, I don't know,
I would almost use the metaphor,it's kind of like when people go
out and look for gold or oil,right? The concept of gold or
oil is known and it's known thatthere is value if you get this
(39:21):
material out of the ground,right? But there's speculation
and risk in trying to figure outwhere that is.
Here, it's not like you'retrying to discover a new
precious metal or a new thingthat's not known to anyone,
right? You know this thingexists, it's a matter of finding
it. And certainly many peoplelose a lot of money in that
(39:44):
speculation. Here it's similar.There is a kind of known, or
there is at least a feeling andan intuition that there will be
very transformative companiesthat will be long lasting and
this technology will be longlasting and it will be
impactful.
It's not that that's maybedoubted, but the crazy
(40:08):
valuations are in some case aredriven by that because, who
knows who's going to survivethis AI craziness. But the
technology it seems will be kindof pervasive and long lasting
and transformative.
Chris (40:24):
I think so. I think I'll throw, I'm gonna throw a wild
card into mix here for a moment.And one of the things that no
one really knows where thingsare going for sure, and we sure
talk about it a lot, is the factthat at this point, you have a
lot of of companies, especiallysome very large companies, such
(40:46):
as Amazon, are doing massivelayoffs, You know? And they're
doing that on the notion, inAmazon's case, 14,000 middle
managers. And that's just onethat's just one organization.
One very large organization, butone, and this is happening in a
lot of places, is that, youknow, we're definitely seeing
(41:09):
the replacement efforts bycompanies to use technology to
replace humans in that. And andthis has been something we've
talked about for years, youknow, certainly coming and that
would be happening and, youknow, where would the balance be
between a synergy between AI andhumans and a competition between
AI and humans. And so we've hadmany conversations over the
(41:31):
years about this. But, you know,we're definitely at that point
right now where a lot ofcompanies are beginning to bet
on AI technologies. But it'salso happening at a time when,
as we talked about, some ofthose ROIs on different efforts
are are not yielding meaningfulresults.
(41:53):
And so aside from what happenswith that, when you have
unemployment rising from AIinduced layoffs, that will
affect the economy too. Sothere's it's not just whether or
not the investment in theseparticular companies is is wise
(42:13):
or speculative and, you know,based on their fundamentals and
their earnings, but also, youknow, as as we have groups of
folks in the economy being laidoff and therefore their
purchasing power is reduced, howdoes that play back in? I've
seen a lot in the news about,like, all of this together, not
only the ROI and and speculativenature or lack or, you know,
(42:36):
whether it's a bubble or not,going back to our our kind of
original phraseology, but alsowhether or not this is going to
affect work phrases. So it's itis quite a complex thing and
when we look back on .com itseemed there were complexities
there but I think the the rawspeculation of that era made the
bubble. It was a little bit moreof a black and white thing as
(42:57):
you looked back on it, you know,historically.
After we had lived through itand looked back and kinda said,
well, yeah, you know, I guesswith 2020 hindsight, we can see
that coming. As we've notedhere, it kinda depends on who
you are and how you're doing andwhat you're claiming as your AI
and whether you're solving areal business problem on whether
(43:18):
or not you're in that bubblegroup or not in a bubble group.
So as we see the mixture pouringthrough the economy, it will be
definitely interesting to seehow this plays out in the weeks,
months and years ahead here.
Daniel (43:33):
Do you think that because we also talked recently
on the show about almost theboth cognitive and emotional,
you know, change and shift oreven manipulation that some of
these systems are doing, youknow, across the population if
you look at it, there's peoplehaving, you know, romantic
(43:55):
relationships with AI systemsor, you know, using these
systems maybe for therapeuticpurposes rather than their
therapist and cognitive load ofwork is changing because you're
vibe coding and all of thesethings. So do you think that
(44:16):
that sort of cognitive andemotional, almost like lifestyle
shift is more impactful withthis technology or with
something like the .com era andpeople kind of all coming
online?
Chris (44:30):
You know, .com era was and and sadly, I was well into
adulthood, you know, as we hitthat, you know, you know, for
those who don't realize that I'mI'm getting a little bit older
than I like to imagine. But,like, fundamentally, we went
through the bus because of allthe speculation, but eventually,
everything kind of you know, wewe did realize what it was going
(44:52):
at. It just wasn't on a atimeline and the and what you
could achieve in that shortertime given the valuations just
wasn't real. But it was real inthe long term. And I think here
what we're seeing is something alittle bit different in that you
weren't having relationshipswith your with your ISP at the
time.
(45:13):
You know, it wasn't that kind ofa relationship that you had with
that wave of technologicalinnovation. This one, it's it's
a little bit worrisome. Like, II don't believe in vibe coding
being a great strategypersonally. I I think that
thinking of AI as a pairprogramming partner is a much
sounder way of approaching it interms of turning out continuing
(45:35):
to turn out very good softwareproducts that you understand and
that you can maintain over timeand that you have humans that
that understand how theirbusiness is working. So I'm a
little pure vibe coding wheresomeone who doesn't really know
what they're doing is justasking the system and then they
end up with something.
Know, so it kinda depends onwhat you're doing with it. I see
(45:56):
there's also been some researchand we may have an episode
coming up on it that was donerecently about the human
dependence on AI causingbasically degradation. They were
measuring brain activity and andp and the subjects that were in
this study that were using theirAI for everything were showing
(46:17):
decreased brain capability overtime. And so it worries me that
we're giving up some of whatmakes us so wonderfully human At
the same time that maybe we'recreating a strong dependence
that we have in these ways thatdidn't exist in previous
revolutions. So I think I thinkthat there's a real risk here of
(46:39):
I think how you how you use AIcapabilities today makes a
difference on what your ownpersonal human future is going
to be.
And I very specifically choosehow I use AI capabilities in a
way that enables me versus kindof creates a crutch for me. So
(47:02):
that's that's kinda how I wouldanswer. That's a little bit
roundabout way of answeringthat. But I think the way you
use AI has a lot to do with itslong term effect on on your your
own personal life and as anindividual.
Daniel (47:15):
Makes sense. Yeah. I think that's a that's a great
way to look at it. Well, as weclose out here, Chris, what
would you say? Are we in an AIbubble or not?
What's your what's your vote?Yes or no?
Chris (47:26):
I'm gonna say no in the classic bubble analogy, I think.
And and I I came into thisconversation not really knowing.
I think it's this us talking itthrough. So I would say no in
the classic bubble context like.com bubble. But yes, and that
there might be instead of onebig bubble, there might be lots
of little bubbles.
Daniel (47:47):
Yeah, a fizzy like our like our PIP zero and Coke zero
that we that we started talkingabout.
Chris (47:54):
Yeah, there you go. The fizzy economy. You heard it here
first folks.
Daniel (47:58):
The fizzy AI soup. On my end, I would tend to go with the
no. I definitely think thatcompared to some other
technologies and cycles, this isalready kind of at a level of
utility and permeating manyenterprises and that sort of
(48:18):
thing. And I do think thatcreates a trickle down of things
that we will have to deal withand learn how to cope with, you
know, workforce wise andemotionally and otherwise. But,
yeah, I'll I'll go with no aswell.
So so you heard it here. NewYork Times, you can quote us if
you like. We are not in an AIbubble. Because Dan and Chris
(48:39):
said so. I'm glad that we putthat one to rest, Chris.
Chris (48:42):
Yeah. Well, there we go. Solving world problems one
episode Yeah. At a
Daniel (48:46):
Well, I guess until next time, we can solve the next
world crisis in the nextepisode, Chris. Thanks for
chatting.
Chris (48:54):
Absolutely. Talk to you next time.
Jerod (49:03):
All right. That's our show for this week. If you
haven't checked out our website,head to practicalai.fm and be
sure to connect with us onLinkedIn, X, or Blue Sky. You'll
see us posting insights relatedto the latest AI developments,
and we would love for you tojoin the conversation. Thanks to
our partner Prediction Guard forproviding operational support
for the show.
Check them out atpredictionguard.com. Also,
(49:26):
thanks to Breakmaster Cylinderfor the beats and to you for
listening. That's all for now,but you'll hear from us again
next week.
Popular Podcasts
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.
Las Culturistas with Matt Rogers and Bowen Yang
Ding dong! Join your culture consultants, Matt Rogers and Bowen Yang, on an unforgettable journey into the beating heart of CULTURE. Alongside sizzling special guests, they GET INTO the hottest pop-culture moments of the day and the formative cultural experiences that turned them into Culturistas. Produced by the Big Money Players Network and iHeartRadio.
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