July 19, 2025 • 46 mins
Have you ever wondered how the future arrives faster than anyone can predict it? Tech futurist Richie Etwaru joins us to explain why "futurists are out of work" in a world where technology adoption happens so rapidly that five-year predictions are tested within months.
Richie breaks down why artificial intelligence has achieved unprecedented adoption rates, outpacing even the internet itself. The perfect storm of pre-existing infrastructure (mobile devices, internet connectivity), brilliant productization through simple interfaces, and our post-COVID appetite for technological advancement has catapulted AI into our daily lives with breathtaking speed.
The most profound insight emerges when Richie explains the fundamental shift from deterministic to probabilistic computing. Since Newton's time, we've operated in a technological world where 1+1 always equals 2—precise, narrow, and certain. Now, probabilistic AI systems offer vastly expanded capabilities but without guarantees of consistent outputs. "We are trading off a high friction, perfect, narrow, deterministic experience with a low friction, wide, imperfect experience," he explains with remarkable clarity.
Those concerned about AI "agents" taking over will find Richie's perspective refreshing. What appears as autonomous intelligence is actually just a probabilistic layer working between deterministic systems—a "model context protocol" that creates the perception of agency where none truly exists. This distinction helps ground our understanding of AI in reality rather than science fiction.
Through his work with Mobius, Richie is reimagining user interfaces—transforming them from self-directed experiences to guided interactions that leverage AI capabilities. This shift creates "the perception of personhood" in digital experiences, fundamentally changing how we interact with digital content.
Ready to navigate this new technological landscape with confidence? Listen now and discover how understanding the probabilistic nature of AI can help you thrive in this rapidly evolving digital world. Follow Richie's forthcoming insights in Wired, Forbes, and The Economist as he continues to demystify our technological future.
Thanks for tuning in to this episode of Follow The Brand! We hope you enjoyed learning about the latest marketing trends and strategies in Personal Branding, Business and Career Development, Financial Empowerment, Technology Innovation, and Executive Presence. To keep up with the latest insights and updates from us, be sure to follow us at 5starbdm.com. See you next time on Follow The Brand!
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:02):
Welcome everybody to the Follow Brand Podcast.
This is your host, grantMcGaugh, and I am going to speak
to what I call one of my mostfavorite futurists when it comes
to technology that I've spokento probably in the last four
years.
And why do I say that?
I say that because I've beenpart of the HEMS organization,
(00:23):
the Health Informatics SystemSociety, and that's in
healthcare, that's in technology, and I was able to meet a
gentleman named Richard Edwaruand he was talking to me about
different kinds of technology.
Remember this was three yearsago.
Ai had not really blown up.
You know these LLMs.
It wasn't blown up then, but heframed discussions around how
(00:45):
technology and nation states andwhere the power plays are where
people will fall into in a waythat I've never heard before,
and his complete understandingof how artificial intelligence
actually operates at asimplistic level.
I thought was just an honor tohear him speak.
(01:07):
So I want to have him speakagain to me and you about how he
sees this new world and how wecan all take advantage of it.
So, richard, you'd like toRichie I should say you'd like
to introduce yourself.
Speaker 2 (01:21):
Hey, first of all, Grant, thanks for having me and
hello to all of Grant's viewers.
I had an opportunity to meetthis man about three years ago
and I got to tell you, one ofthe hardest interviews I did
with him at that time.
So I look forward to beingbattered around a little bit
here.
Just a little bit of backgroundon me I live in the US.
(01:41):
I've been living in the USprobably for a little bit over
30 years now.
I was born in Guyana, southAmerica, ex-british colony,
probably 51% tech, you know, 49%business operator.
I've done a ton of stuff withtransforming businesses, going
(02:02):
through IPOs, starting companies, working in private equity,
writing books, doing speechesand stuff like that.
I think the one thing I wouldsay to start this conversation
is that I've been referred to asa futurist over the last maybe
25, 30 years and have made quitean impact in that area.
But you know, today futuristsare out of work because the
(02:26):
future is here and even if youtry to forecast the future, it's
moving so fast that you'rebeing held to your work.
So you know, five years ago, ifyou're a futurist you could
forecast out 20 years and ifyou're wrong, nobody cares,
because you would have probablyretired by that by then.
Right now, if you'reforecasting out five years out,
(02:46):
we're going to find out whetheryou're right or not by December,
right?
So the dynamic of managingthese horizons of time right
when you think about how timeand culture and industry moves
together.
The pace has changed, which ischanging everything, something
as simple as whether you're afuturist or not.
Speaker 1 (03:06):
I agree with you.
Another colleague of mine,christopher Lafayette, brought
something to my attention whenhe said and this was COVID time,
right, this is like 2019.
We all had to participate in ahuman experiment, whether we
liked it or not, participate ina human experiment more that we
(03:27):
liked it or not, and we had totest our tech.
Meaning cloud computing, yeah.
Mobile computing, yeah.
Now we're starting to see canwe do it at scale?
When it came to communication,video transfers, we started to
see the rise of metaverse.
I mean, we can totally livealmost in a digital type society
and do trade and commercewithin that society.
That was the metaverse where wehave morphed since then.
(03:51):
We're starting to like flex ourmuscles a little bit.
And I'll coin a statement.
You told me we're releasing anew like an animal called almost
a species, an animal into ourworld Artificial intelligence.
We are releasing this out.
It's been the most adoptedtechnology in human history.
(04:11):
It's just huge Faster thancloud computing, faster than
mobile computing, faster thanelectricity From your lens.
What you're seeing now becausewhat I'm seeing coming on my
horizon is a genetic AI.
You came out with the tele.
I'm starting to see others comeout with certain tools that are
(04:31):
coming.
Hey, this is an all-in-one tool.
It can interact with you, itcan speak to you, it can do
certain human-like functions.
To take what I'd say tier one,tier two, off the table, it's
blowing up the BPO world.
How do you see this technologyoperating?
Speaker 2 (04:48):
Well, it's a broad question.
I'm going to try to answer itthe way you eat an elephant, one
piece at a time.
Just for, I guess, a littlecomedic relief.
I would argue that AI is notthe fastest adopted technology
in the world, okay, but emojiswere the fastest adopted
(05:13):
technology in the world.
As soon as you saw an emoji,you knew what to do with it
right away and you wereabsolutely fluent.
Now, yes, of course it doesn'thave a gravitas on the
implication as artificialintelligence, but again, for
some comedic relief, here on theBrett Madrona show, so to speak
(05:33):
, there's a lot that has alreadybeen said, so let me try to
focus on, maybe, what ought tobe said that might've not been
said.
Yeah, maybe what ought to besaid that might have not been
said, yeah, one of the reasonswhy AI has been adopted so
quickly, and maybe we'll coverthree reasons why the speed is
(05:55):
there.
First, you know, technology issomething that builds on itself,
and if you were involved intechnology at any point, you
must have seen some experiencewhere the thing that you were
doing at the time was then usedto create the next thing that
(06:18):
you're going to do and was usedto create the next thing that
you're going to do.
So, even as simple as like theFortran compiler right In some
ways, c++ was written in Fortranright.
Java was written using C++right.
Javascript was written usingJava right.
So, like you, use the toolsthat are commoditized to create
(06:39):
the next frontier, the nextadvancement that creates
competitive frontier.
Right Now, if you think aboutthe internet, a lot of plumbing
and infrastructure had to be inplace for someone to consume the
internet Wires on the ground,fiber optic cables, cell towers,
chips, electric capacityhardware had to be shipped,
(07:02):
multilingual things had to befigured out before you got the
internet.
So you know.
If you think about any thirdworld country I'll just pick
Guyana for a second you knowGuyana got electricity about 40
years after Europe and the USgot electricity right.
So you've kind of you kind ofare left behind for a while.
(07:25):
Artificial intelligence howeverthe platform was already there.
You had a mobile phone, you hadinternet already right, and so
you know the 5 billion peoplethat are already pre-connected
to the internet and have amobile were able to uptake AI.
(07:47):
From the time that you foundout about it to the time that it
takes to either hit the open AImobile browser application,
which was the first or downloadan app from the App Store, right
.
That was the uptake time, andthat's why you see this massive
uptake of individuals.
You see this massive uptake ofindividuals and I think what's
(08:10):
the next point as to why I thinkthe adoption was so easy is
that OpenAI did something thatis the core of the craft, which
is to productize it right.
Artificial intelligence, as itrelates to foundational models
and it relates to machinelearning and deep learning, was
there for a while, not as goodas the foundational models that
(08:34):
OpenAI's foundational model setforth, which started a
competitive race.
Not as good, but what was therewas that it was never
productized for mass consumption.
If you wanted to run a modelprior to chat GPT, you kind of
had to have Math Lab running,right.
(08:54):
You had to download a bunch ofstuff on a desktop, you needed
to understand how the fluxcapacitor of every button works,
and then the outcome was alarge text file with weights in
it.
It's just not accessible.
What OpenAI did, which is tostart a for-profit company to be
able to create funding to fundthe nonprofit company, was to
(09:17):
consume their models andproductize it as a chatbot, and
that productizing of a chatbotis, to me, the brilliance of
OpenAI, not necessarily the coreLLM underneath, although they
did start the arms race bydifferentiating significantly
when you make a product thateverybody already knows how to
(09:38):
use they know how to chat.
Who here doesn't know how tosend a text to someone and get a
text back?
Know how to send a text tosomeone and get a text back?
Then to find the right place todo it, which is these language
models, these foundationalmodels, and I want to talk about
probabilism after this.
They're imperfect.
How do you productizeimperfection?
A chatbot is a reallyinteresting way to productize
(10:02):
imperfection, because there's nogood answer, there's no perfect
answer.
That has to come back Right,and so that's another reason why
we saw that speed of adoptionof it.
And the third is something thatyou'd mentioned earlier in your
preamble here to this question,which is mankind now has a
thirst for technology and thefuture.
(10:23):
Mankind now has a thirst fortechnology and the future.
When we went through COVID,everybody, from my mom to you
know the guy who works in mylocal hardware store here in a
(10:43):
small town that I live ineverybody saw the power of
technology, whether it wasthrough Instacart deliveries or
the forms that we all filled outto get vaccines earlier or
going to.
I forgot what was the websitethat gave you all the stats of
how many deaths?
Speaker 1 (10:55):
Oh, yeah, I think it was the US government, or one of
them, whatever it was right.
Speaker 2 (11:00):
Everybody got to see the value of technology, not
just the sort of soft, flimsysort of magic of it or the
excitement of it or the dopaminehit of it or the vanity
advancements of it, but theactual utility of it.
Now, and we've all as anaggregate I would say as an
(11:28):
aggregate we've tripled ourawareness and our appreciation
for the utility of technology.
We come out of COVID.
Everyone's like okay, where'sthe future?
This is great.
You guys have been talking fortech so long.
We're ready for it.
What's up?
Where's the future, man?
So you have that thirst sort ofwaiting in society to adopt
this, and I think it's thosethree kind of pieces of that
(11:50):
storm that came together thatcreated this massive transition
of human awareness and thisreminder that we are evolving.
What AI really reminds us allat the core is that we are an
evolving species, we areconstantly evolving and we're
all living through the Internetagain, we're living through
(12:13):
electricity again, we're livingthrough the printing press again
, and so everyone, I believe, onthe planet in some way is now
aware of the fact that we'reliving in a pretty pivotal point
of history.
Speaker 1 (12:26):
There's no question about that.
I call it a leveler, let's say,like that of economies of scale
, where, like, let's say, ifyou're an oil baron, let's just
say you know you had a certainamount of control.
You know you had a certainamount of control.
(12:49):
If you're a real estate bearer,you had a certain amount of
control.
Technology has now given awhole new level of control to a
whole new set of individualsthat aren't oil bearers.
They don't own, you know,massive real estate, but they
have massive value.
For the first time in humanhistory we have an evaluation of
a company of a trillion dollars.
(13:09):
That's almost unbelievable.
When you think about it, peoplesay, well, you know, it's taken
certain people a long time toget what you would call economic
advantage.
And then I say you know what alot of these companies right now
, and then NASDAQ, when you lookat the Fortune 5, those top 10,
and we call them themagnificent seven haven't been
around for hundreds of years.
(13:30):
These are newer typetechnological advancements that
have taken these companies to awhole nother level.
When I think of Mobius this is,in my opinion, when I think of
when I first saw the AirGlass, Isaid, all right, somebody's
actually thinking out of the box, because it's almost like when
I saw the Google goggles that Icould say they came out for a
(13:52):
metaverse view and it's just Idon't know.
It reminded me of the 8-track.
I don't know how long that'sgoing to last, but we'll see the
8-track player.
But you said, hey, let's takethe glasses off and still have
what we want, which is digitalinteraction in a physical space.
It's kind of an expandedawareness of what we can do.
Right, very cool stuff.
(14:15):
Now you've come out.
I mean, this is a big word andthis is also taken.
First time I heard aboutgenetic AI was probably about a
year ago.
Gentic AI was probably about ayear ago and it's just another
explosive type tech where peopleare looking at it in many
different ways.
You've taken it and I like yourviewpoint.
You're taking like hey, and Igo back to the Microsoft and IBM
(14:38):
world.
I'm not reinventing a computer,I'm making it easier to use.
You seem to be making this kindof tech easier to use.
Speaker 2 (14:55):
What's your opinion on these things?
Well, look, my career has beenabout reinventing user
interfaces my entire career.
While I was at PrudentialSecurities, which is now
Wachovia, I worked extensivelyon connecting the CICS mainframe
green screens to the VisualBasic compiler to start to build
point-and-click apps forbrokers to be able to have
(15:16):
information, faster and betterworkflow as they serviced
investors to the brokeragebusiness of Prudential
Securities.
And that was the first patentand you know it got very
interesting between Microsoftand IBM.
But that was my first timerecognizing that user interfaces
will change.
The second time I did that wasat Lehman Brothers where
(15:42):
reporting so client reportingwas very static.
Everybody got the same PDF, thesame design of a PDF, and you
know I focused my energy thereon figuring out dynamicism of
PDFs.
How do you get a bankingstatement that has a pie chart
of your asset allocation on pageone but somebody else gets a
bar chart of their assetallocation on page one, but
(16:02):
somebody else gets a bar chartof their asset allocation on
page one because they want tosee their data differently or
their portfolio is betterdisplayed as a bar chart versus
a pie chart.
And that was the second time Ikind of reoriented user
interfaces from static reportsto sort of personalized reports
Again another patent, anotherkind of partnership with Adobe,
(16:24):
and today everybody haspersonalized PDFs.
I did it again at UBS I won'tbore everybody with that around
working with Apple to put thefirst wealth management app on
the iPad at the time of the iPadunveiling by Apple when you met
me.
The thesis there was that thescreens will become more
(16:49):
engaging, and as I was foundingthe company, I had decided to
call the company Meta M-E-T-A,and right before I set out to
kind of start, you know, takingthe company forward, I set out
to kind of start taking thecompany forward.
Facebook renames itself to Meta,and I had no interest in the
(17:10):
Metaverse.
I was pretty sure that we'renot going to strap toasters to
our head for a pretty longperiod of time.
And so in some ways, zuckerbergirritated me to get into the
metaverse by naming Facebookmeta.
Now, I'm nowhere the size andscale of Zuckerberg.
(17:32):
He doesn't know that I exist,but that's what happened.
He irritated the hell out of me, and so I thought well, what's
this stupidness anyway about themetaverse?
Why would anybody do this?
And so that was the first timeI actually thought, you know,
maybe there's something here andI looked at it and I said
(17:52):
there's no way the headsets work.
We ought to be able to get thescreens to give us more.
And the long and short of it,40 patents after I was able to
devise a way to use computervision, to use the screens that
you and I are looking at rightnow to be immersive and
interactive at the same time,and essentially, I built a
version of the metaverse thatdidn't need the headset.
I'm really not interested inthat.
(18:15):
And so as the metaverse windeddown and we were using computer
vision a lot, I started goingback to my original thesis which
is why I founded the companywhich is how can we get more
from the screen?
It has nothing to do withimmersivity and interactivity,
and so today, mobius is back toits original thesis.
I got sidetracked by Zuckerberg,who irritated me on the name,
(18:36):
and where we are today is that Iam looking again at the user
interface.
One of the net nets ofartificial intelligence is that
the user interface is going tochange, and you know, to put it
into context, google changed theuser interface by putting a
search bar.
The user interface of AI isgoing to be a microphone icon.
(19:02):
That's it Top, turn it on orturn it off, the mouse and the
keyboard will still have a place, similar to how engineers have
a place in the IDE today, rightor in black and white.
You know terminals, and so whatwe're doing at Mobius today is
(19:22):
we're really focusing on thatuser interface.
What is the intelligent userinterface?
Now, I won't talk too muchabout that solution.
I can connect with folksafterwards.
But the value for your audienceis to try to parse the
artificial intelligence markettoday to know where you want to
make your bets and generally,when these big industrial
(19:44):
revolutions come about and youwere talking about oil, it's the
same thing.
You can parse it into threeverticals.
You've got the producers of thenew utility, You've got the
appliers of the new utility andyou've got the consumers of the
new utility right.
So you've got the producers ofAI right and a whole supply
(20:06):
chain in there.
I put the silicon and thetraining in the production
bucket.
So NVIDIA and OpenAI I havethem in the same bucket, so to
speak.
Right, hyperscalers I kind ofpop them in there as well.
The appliers of artificialintelligence is where I sit.
Let's take the core utilitythat's being produced by others.
By all means.
Pay them their cogs, make thema part of supply chain, but
(20:28):
apply it to do somethingcompletely unique and different,
right?
Not dissimilar from how youknow.
Amazon applied the internetwithout spending a dollar on
building the internet, right.
And then you have the consumersof artificial intelligence.
What's also different here, andhow your audience wants to
navigate.
Whether you're a producer of AIand you know, be smart about
(20:50):
your capital investment there.
Whether you're a producer of AIand be smart about your capital
investment there and who you'recompeting against, and whether
there's a David and Goliath gameto play there, I argue that
there is a very low likelihoodthat a David and Goliath game
could be played out on theproduction side.
I feel like that side isalready locked up right On the
applier side.
Man, a lot of David andGoliaths are playing out on the
(21:13):
application side right.
Large software companies thathave been underserving their
constituents for decades are nowunder attack because people are
like wait, what the hell?
Why am I paying for a SaaSlicense that I'm not consuming?
Where the interest is is on theconsumption side of artificial
intelligence, and theconsumption side has two things
(21:35):
about it.
One is you can consume it tobuild businesses really fast I'm
talking about ridiculously fast, especially if you're building
businesses that have some coredigital component to it, right?
And why is it that languagemodels make it faster to build
(21:56):
companies that have digitalcomponents?
It's really simple.
Language models were trained onthe internet, and the internet
is not only full of information,it's full of software code.
When you read a website,there's JavaScript in there,
there is cascading style sheetsin there, there's HTML and DHTML
in there, right?
And so now language models arenot just good at our language,
(22:19):
they're good at the language ofthe machine, and so building
apps, building websites,building workflows, building
marketing streams are all aseasy to kick out as the answer
to you know what color is thesky.
And so that's what makes thisconsumption side so much more
interesting, because it loweredthe barrier to entry, and the
(22:42):
other part of it is theproductivity hack, right?
Not only is it helping youbuild businesses faster, it's
helping you do what you do todayfaster, and the internet didn't
have that acute differentiationon the consumption side.
Yes, it commoditized distancesignificantly, but you can't
build a business because youwere just able to commoditize
distance.
You still had to go through themechanics of a highly
(23:04):
capital-intensive sort of thing.
I think, when you look at theAI market from that perspective
and again I'm playing in thatapplying AI layer for the
benefit of the consumer's AIlayer, because I see the user
interface transitioning fromwhere we are today to where it's
going and that's where maybe wecould talk about probabilism in
(23:27):
that user interface, but Ithink for an audience in that
user interface, but I think foran audience, I think that that
is a reasonably good way toevaluate risk and opportunity
within where we are today.
Speaker 1 (23:46):
I want to expound on what you just said, especially
toward the end, understanding.
At the end of that whole thing,you have a human and then you
have what we are calling now AI,artificial intelligence, or
what I'm calling now a bettertechnology.
It's a tool.
Now I say this, remember thisIn our human history most people
agree humans took a great leapin their environment when they
(24:11):
discovered fire.
Ai is a type of fire and if youknow how to apply the fire
because you can take the fire,you can burn the forest down or
you can cook your food If youknow how to manage the fire,
you're responsible, accountable.
(24:32):
You understand the nuances ofthe fire, you know what it could
do, what it couldn't do.
You know that, hey, if I putwater on the fire, the fire is
going to go out.
But the fire also consumesoxygen.
It consumes a lot of thingsPeople don't realize that a
human being does.
No, you've got a.
Fire is a living thing, right.
(24:52):
So I said it to say this we arenow starting out with, from the
human component or reality,operating with a, a technology
that everybody can now utilizeand, to your point, I myself can
make businesses where before Ididn't have that capability.
(25:13):
Right now we're on an interfaceyou didn't have.
You know you have to have massresources to do radio, to do
television, to do mass media.
The barrier has beensignificantly lowered.
So if you, as a human, knowshow to operate or drive, I'm
(25:35):
going to say this technology,this AI, and you realize, like
you said before, the buildingblocks are already in place.
So you don't have to be a coder, you don't have to understand
(25:56):
how the Internet actually worksto a certain degree in the back
end, but you gotta know how todraw.
And if you're making it easyfor me to get behind the wheel
of the ai engine, I seeincredible transformation that
can, that can change the worldas we know it, and I've always
said this.
I have a lot of younger peoplethat ask, asked me a lot, and
asked my opinion like well,grant, what do you think my
purpose is and I always say thisespecially to younger folks
(26:19):
your purpose is to change theworld.
I've already lived in thisworld for 60 years.
I know what it's like.
I've gone through changes inhow we do things.
Your job is to change it evenfurther.
I'm going to point this to youbecause I like how you had that
comic relief right around theemoji right.
I remember sitting therewatching the Jetsons, and they
(26:41):
would have a communicator thatlooked a whole lot like your air
glass, you know, and it wouldjust be an image out of thin air
that you would interact with.
I truly believe that man'simagination, when it could be
productized, could be one of thegreatest advancements in human
history.
So I want you to, if you don'tmind, help me unpack some of
(27:06):
this, because what I'm saying isthis Human plus AI could be the
biggest game changer on theplanet.
Speaker 2 (27:13):
Yeah, yeah.
Well, I think, based on what weknow today, you know I never
underestimate 100 years out, youknow, there could be something
even bigger and no one knowswhat it is and no one knows what
it is.
Arguably it could beinterplanetary travel, right,
(27:34):
could be even bigger.
Again, thinking about thismight be an organics, a second
organic species, not a firstdigital species, right, as we
refer to it here.
Let's talk about the trade-offthat we're making, and I'm going
to focus on the trade-offspecifically from an enterprise
(27:59):
and a technology perspective.
I think your audience sits inthat center of gravity a little
bit.
So there is a fundamentaldifference between what we did
with the Internet and what we'regoing to do with artificial
intelligence, as it's specificto generative artificial
intelligence.
There's many other types of AI.
(28:20):
Computer vision, for example,is a different story, but when
you want to come to terms withwhat is different, I would
suggest that you start here.
In the internet world, we useddeterministic mathematics and in
deterministic math, 1 plus 1 isequal to 2.
(28:43):
And you can guarantee that 1plus 1 will always be equal to 2
.
You can guarantee it Withgenerative AI, which is
probabilistic math notdeterministic math, but
probabilistic math.
You cannot guarantee that oneplus one will be equal to two
(29:08):
every time.
You can't, because it's such anobvious thing.
One plus one will be equal totwo every time.
You can't, it'll be becauseit's such an obvious thing.
One plus one will be equal totwo a lot but you can't
guarantee that it'll be everytime because it's probabilistic.
There's no such thing as a rightor wrong answer.
The only thing that's trueabout probabilistic math is that
(29:31):
it must have an answer everytime, and so that's why, when
you ask a probabilistic modelfor something and language
models are sort of like thecodfish example everybody knows
what I'm going to use here.
But let's take an image model.
If I ask an image model tocreate me an image of a blue and
(29:52):
red dog, it'll give me anawesome blue and red dog.
Try to get that same dog again.
Speaker 1 (29:59):
Yeah.
Speaker 2 (30:02):
Zero chance to get that same dog again.
Now we have the same thing inlanguage models.
Ask it a question.
If it's not too obvious of aquestion, try to get the same
response every time.
Zero chance.
And so, living in thisprobabilistic world right, where
you're treating off adeterministic relationship with
(30:23):
technology for a probabilisticrelationship with technology,
now you start to see thedifference of the mechanics of
it.
For example, if we just takeuser interface for a second, in
a deterministic world you havean application that could do
about 10 things, okay.
And you must move your mouseprecisely to here to click on
(30:46):
this one button to do those 10things.
And you must wear your readingglasses to read the little text
that's on the side to know whichof those 10 things to do those
10 things.
And you must wear your readingglasses to read the little text
that's on the side to know whichof those 10 things to do, okay.
So there's a high degree offriction to a deterministic
relationship with technology.
Okay.
However, it's guaranteed thatit's going to do 10 things and
(31:08):
it's going to do those 10 thingsexactly the same way every
single time, bar none.
In a probabilistic relationshipwith technology, the number is
not 10.
They can do a billion differentthings.
However, the perfection is notthere.
Okay, so we are trading off ahigh friction, perfect, narrow,
(31:45):
deterministic experience with alow friction, wide, imperfect
experience, and that's thetrade-off.
It's because you've gotprobabilistic math on one side
and deterministic math on theother side.
Now the artistry and the craftis to figure out the balance of
(32:08):
those two to be able to makesomething of utility for mankind
.
The challenge is that as soonas you add one ounce of
determinism into a probabilisticsystem, you lose a significant
amount of the probabilisticvalue, like, if you say, to a
(32:28):
generative model.
In this scenario, you mustgenerate the same thing every
single time.
You essentially stifle theintelligence to the ground Right
, and that's the core that wehave to understand.
Speaker 1 (32:46):
I'm glad you brought me back to one of my most
favorite things I ever studiedin high school, which was
physics.
I love physics.
I wasn't great at it I gotactually a C in the class but it
was my favorite class of allbecause of those things you just
talked about.
(33:06):
My teacher would tell us allthe time there is no right and
wrong answer.
I'm more interested how you gotto your answer than what the
answer actually is.
I want your critical thinkingand and and come up with
something that's interestingbecause you can get outside the
box, just like you were sayingso sometimes and I always liked.
(33:30):
I think Einstein brought it upwhere, as you pointed, one plus
one equals two, but in his worldit doesn't.
It equals what we call three.
What do you mean by that?
Because you by yourself, you'reone entity.
Me by myself, I'm one entity,and if we both operate in those
separate worlds, that're oneentity.
Me by myself, I'm one entity,and if we both operate in those
separate worlds, that is oneplus one.
You could kind of like two, butwe come together.
(33:51):
There's a third reality thatcomes about.
I'm going to call that thispodcast discussion right Now.
We both could never reallydetermine exactly what we were
going to say and do and theanswers, but it becomes a third
reality.
That's that plus one.
I believe there's a plus one ineverything in life.
There is, it's just are youbold enough to engage with that
(34:13):
reality?
As I said earlier, the humanplus the fire.
If you know how to manage thefire, it's a great tool.
But if you don't, you know, youdon't know exactly how it's
going go Ask Elon Musk in hislast billion dollar explosion
that happened.
There's a lot that couldpotentially go on.
So, as men, we're starting toget out of our deterministic
(34:37):
world to a certain degree,because everybody likes to stand
on solid ground, right.
But now you're saying I'm goingto stand on solid ground, right
, but now you're saying I'mgoing to stand on something I'm
not sure if it's solid, could beor couldn't be, but how do we
get there?
It's going to be a veryinteresting feature, I believe,
and a very near feature.
Speaker 2 (34:55):
I want to just follow on to that very quickly Go
ahead.
I was doing some work with theboard and to engage the board, I
started by asking you know,deterministic math was arguably
discovered by Newton in the1650s.
(35:18):
1650s or 1850s, I don'tremember a 50, it won't matter
to the story.
1850s, I don't remember a 50,it won't matter to the story.
Um.
And then I asked when do youthink probabilistic math was
invented?
And everybody was wrong,everybody oriented to much later
in history.
You know like within the last50 years, right, people make
(35:40):
bets there.
It turns out probabilistic mathwas arguably discovered by
pascal within 10 to 20 years ofdeterministic math from newton,
really around the same time inthis thing that you said that
that we need, we need, um,determinism as as a species.
(36:02):
Um, what happened at the time?
Is deterministic math picked upbecause the scientific
revolution needed determinismand probabilistic math never
picked up because of the simpleargument that God has to exist.
There can't be any questionaround whether God exists.
And so probabilistic thinkingand probabilistic math, because
(36:25):
there are no guarantees in thesystems, ended up to be seen as
like a little bit of a dark arts, if I may right, a little bit
of like a spooky kind ofsituation, a little bit too
gamblish, and society at thetime was not ready for
probabilism.
You know, are we ready forprobabilism today?
(36:46):
And it's going to come down toa couple of things.
It's going to come down to onewhat are the areas where
probabilism works?
And let's touch on how did weget to agents?
How did the concept of agentscome about, which led to the
(37:09):
largest word washing of mycareer.
Talk about cloud washing andmetaverse washing and mobile
washing and internet and onlinewashing.
I've never seen washing likethis, an online washing.
I've never seen washing likethis.
But when you think about thedeterministic systems and the
probabilistic systems, again,right.
(37:30):
And you say, okay, so a lot ofour utility to build things that
people will use sit in adeterministic world, right now
exposed by API world, right nowexposed by API, right, yeah.
And then you've got thisdeterministic UI which is, if I
(37:53):
click buy, it's incrediblydetermined that you must go call
the buy API, right.
Or if I click sort, you shouldgo call the sort API.
So this very, very rigidrelationship between actions in
the user interface and executionof that intent in what we could
argue to be called a system ofrecords, now a concept evolved
(38:19):
in between the deterministicuser interface and the
deterministic APIs and in thatspace in between is called model
context, which is if the UI isgoing to be driven by a language
model or a conversational model, right, how does that
probabilistic UI calldeterministic APIs?
(38:42):
Yeah, how do you know thatsomeone meant to buy?
Because the intention to buy isno longer determined by
clicking the buy button.
There could be a million waysthat you indicate that you want
to buy.
You could say, all right, let'sgo, and that means that you
want to buy.
You could go, I'll take it, andthat means that you want to buy
(39:05):
.
Right.
And so model context protocol isthat thing that says figure out
what that person said and then,based on all APIs that you have
available, that's where themodel context, the MCP servers,
come in, which is hey, what areall the APIs that are there and
(39:26):
how are they interconnected?
Describe them in an MCP serverso that when the language model
hears I'll take it, it does atranslation of that.
Look at the API and go.
The only API that makes sensehere is the buy API.
So the MCP, this model contextprotocol, is this membrane
that's sitting between APIs andbetween a UI that's starting to
(39:50):
expand, that uses probabilism inthe middle of two deterministic
systems, and that's where theperception of agency is coming
from.
There's nothing sentient oragentic about a system.
The concept of an agent ishorseshit.
(40:13):
That probabilistic layer inbetween is creating the
perception of autonomy.
How could it have known that?
When I said I'll take it, Imeant buy.
It's really simple.
There's nothing agentic aboutit.
I haven't met an agent as yet,other than, like you know, my
agent at the airport when I'mtrying to change my flight.
(40:33):
I have not met an agent as yet,and I'm deep and wide in the
space.
It's complete horseshit.
Speaker 1 (40:41):
I love how you said that.
I think what and how I look atit is like this and we're going
to conclude here, but I had thesame, similar, somewhat, not as
technical as you just describedit, but what I was explaining to
people was like, when we'retalking about in an AI-driven
world, we're talking aboutsomething that's just mirroring
human intelligence or behaviorto a certain degree, as we
(41:05):
understand ourselves.
So you're asking something.
Here's the three things.
I look at Something that isunconscious, that is, the AI
itself.
It's unconscious to someonewho's conscious, which is
yourself.
But then also you are operatingin a subconscious world a lot
too.
But then also you are operatingin a subconscious world a lot
(41:26):
too.
When you start talking abouthow are you going to understand
behavior or what drives you, itgets into deeper layer.
So then you start seeing, yeah,in our world, in the deeper
context of what you call thenatural world and how we operate
in everything around us, it'snot so much deterministic.
There's a lot going on, a lotmore intelligence than just
(41:50):
human intelligence that we'reoperating in.
But we're starting to discoverthis and then we're going to see
how we're going to interoperatewith it, let's say, collaborate
and understand how all this isgoing to work together.
But you brought up some very,very good points and that at
some point in time, hopefully inthe near future we're going to
have another conversationbecause we're ever evolving.
(42:11):
You're ever evolving and beforeI conclude, I want you to just
kind of sum up what you feelabout this particular podcast.
How do you feel about theFollow the Brand?
Speaker 2 (42:23):
how I feel about.
Speaker 1 (42:23):
Sorry, the last word I missed yeah, follow the brand,
the follow brand podcast.
How do you feel about thisinterview?
Speaker 2 (42:29):
look, I've been tracking you for a while and
what I like is your consistencyand your perseverance.
Um, a lot of people build thesetypes of um in some ways
philanthropic platforms to beable to help others along the
way because of privilege oraccess that you may have.
(42:51):
Um, and most people fall offbecause it's hard.
It's not easy.
Getting guests is hard, doingit every week is hard.
Getting the courage courage tobe able to engage in it and put
your own brand out there is hard, and so I think your
(43:12):
consistency and yourperseverance marries what I do
in the early stage venture world.
In the early stage venture world, it's really only two
ingredients Perseverance, yeah,and craziness.
You have to be crazy A littlebit out there, yeah.
In some ways and I say thisself-deprecatingly in some ways
(43:34):
you have to be a special kind ofstupid to do something like
this, what you're doing and whatI'm doing.
You can't be born with thistype of stupidity.
You have to cultivate it right.
Nature can't make you thiscrazy to do things like this.
So I really commend your, yourperseverance and your commitment
.
That's what makes you know BDMwhat it is.
(43:55):
Thank you.
Speaker 1 (43:58):
I truly appreciate that Before we let you go, you
got to let us know how tocontact you, and you also said
you're going to be upcoming.
You're writing on Forbes,you're doing some things for
Wire.
Tell us what you've got goingon, richie.
Speaker 2 (44:11):
Well, I think a lot of what I'm going to be doing is
as we get ready to come out ofprivate beta.
We're working with about a halfa dozen or so Fortune 500
companies in a private beta atMogus Very, very exciting
approach.
Again, we're transitioning theentire user interface that you
have, any website that you'veever had before, any document
(44:32):
that you've ever had.
We're changing it from a guidedhigh-friction experience, I'm
sorry, from a self-directed highfriction experience.
So if you think about marketing, you spend a lot of money to
get people to your website andthen you leave them alone.
Self-directed, right.
We're moving that to guide itby putting this perception of
(44:55):
personhood in it, so that youfeel like you're guided when you
arrive at the website.
What would you like?
How can I help you?
What do you want?
Right, using that sameprobabilism in the middle to
create the perception of agency.
To understand something, likeyou know I'm ready to take, it
means that you want to buy.
So the work that I'm doing rightnow is to start to bring
forward the insights thatinformed the Mobius Telly, the
(45:21):
instincts that informed theMobius Deli, the deep
understanding that informed theMobius Deli because I think part
of it is I am dying to sharewhat I've learned and kind of
picked up over the last four orfive years with the rest of the
industry.
And two is in a world whereyou're launching an artificial
intelligence product, there's alot of founder gravity that
(45:43):
comes with it.
So I'm getting into wires.
A wired Forbes has finallynudged me enough to start
writing again.
I got a piece in the Economistcoming.
I'm doing some board work, I'mdoing some conference circuits
and stuff like that, because youknow I have to come out first
before the product comes out.
Speaker 1 (46:02):
I like that.
See that personal brand, seethat Everybody asks me why
personal branding?
What is business branding like?
You have to have a face of yourbrand to communicate
effectively and simplistically.
I mean you have a lot of.
You simplify very hard conceptsand I think a lot of people are
like, wow, I think I understoodwhat he just said this last
(46:22):
hour right?
I think that's wonderful.
I am going to send you picturesfrom Guyana when I get down
there, because this is going tobe wonderful.
Take you home just a little bit, show you what I'm seeing down
there and then hopefully we'llcircle back before the end of
the year and have another good,intelligent conversation.
And I want to thank you againfor being on Follow the Brand.
(46:44):
I want to encourage all of yourfollowers to follow us at
5starBDM.
That's the number 5.
That's star.
Bdm is B for brand, d fordevelopment and for masterscom.
I want to thank you again forbeing on the show.
Welcome everybody to the Follow Brand Podcast.
This is your host, grantMcGaugh, and I am going to speak
to what I call one of my mostfavorite futurists when it comes
to technology that I've spokento probably in the last four
years.
And why do I say that?
I say that because I've beenpart of the HEMS organization,
(00:23):
the Health Informatics SystemSociety, and that's in
healthcare, that's in technology, and I was able to meet a
gentleman named Richard Edwaruand he was talking to me about
different kinds of technology.
Remember this was three yearsago.
Ai had not really blown up.
You know these LLMs.
It wasn't blown up then, but heframed discussions around how
(00:45):
technology and nation states andwhere the power plays are where
people will fall into in a waythat I've never heard before,
and his complete understandingof how artificial intelligence
actually operates at asimplistic level.
I thought was just an honor tohear him speak.
(01:07):
So I want to have him speakagain to me and you about how he
sees this new world and how wecan all take advantage of it.
So, richard, you'd like toRichie I should say you'd like
to introduce yourself.
Speaker 2 (01:21):
Hey, first of all, Grant, thanks for having me and
hello to all of Grant's viewers.
I had an opportunity to meetthis man about three years ago
and I got to tell you, one ofthe hardest interviews I did
with him at that time.
So I look forward to beingbattered around a little bit
here.
Just a little bit of backgroundon me I live in the US.
(01:41):
I've been living in the USprobably for a little bit over
30 years now.
I was born in Guyana, southAmerica, ex-british colony,
probably 51% tech, you know, 49%business operator.
I've done a ton of stuff withtransforming businesses, going
(02:02):
through IPOs, starting companies, working in private equity,
writing books, doing speechesand stuff like that.
I think the one thing I wouldsay to start this conversation
is that I've been referred to asa futurist over the last maybe
25, 30 years and have made quitean impact in that area.
But you know, today futuristsare out of work because the
(02:26):
future is here and even if youtry to forecast the future, it's
moving so fast that you'rebeing held to your work.
So you know, five years ago, ifyou're a futurist you could
forecast out 20 years and ifyou're wrong, nobody cares,
because you would have probablyretired by that by then.
Right now, if you'reforecasting out five years out,
(02:46):
we're going to find out whetheryou're right or not by December,
right?
So the dynamic of managingthese horizons of time right
when you think about how timeand culture and industry moves
together.
The pace has changed, which ischanging everything, something
as simple as whether you're afuturist or not.
Speaker 1 (03:06):
I agree with you.
Another colleague of mine,christopher Lafayette, brought
something to my attention whenhe said and this was COVID time,
right, this is like 2019.
We all had to participate in ahuman experiment, whether we
liked it or not, participate ina human experiment more that we
(03:27):
liked it or not, and we had totest our tech.
Meaning cloud computing, yeah.
Mobile computing, yeah.
Now we're starting to see canwe do it at scale?
When it came to communication,video transfers, we started to
see the rise of metaverse.
I mean, we can totally livealmost in a digital type society
and do trade and commercewithin that society.
That was the metaverse where wehave morphed since then.
(03:51):
We're starting to like flex ourmuscles a little bit.
And I'll coin a statement.
You told me we're releasing anew like an animal called almost
a species, an animal into ourworld Artificial intelligence.
We are releasing this out.
It's been the most adoptedtechnology in human history.
(04:11):
It's just huge Faster thancloud computing, faster than
mobile computing, faster thanelectricity From your lens.
What you're seeing now becausewhat I'm seeing coming on my
horizon is a genetic AI.
You came out with the tele.
I'm starting to see others comeout with certain tools that are
(04:31):
coming.
Hey, this is an all-in-one tool.
It can interact with you, itcan speak to you, it can do
certain human-like functions.
To take what I'd say tier one,tier two, off the table, it's
blowing up the BPO world.
How do you see this technologyoperating?
Speaker 2 (04:48):
Well, it's a broad question.
I'm going to try to answer itthe way you eat an elephant, one
piece at a time.
Just for, I guess, a littlecomedic relief.
I would argue that AI is notthe fastest adopted technology
in the world, okay, but emojiswere the fastest adopted
(05:13):
technology in the world.
As soon as you saw an emoji,you knew what to do with it
right away and you wereabsolutely fluent.
Now, yes, of course it doesn'thave a gravitas on the
implication as artificialintelligence, but again, for
some comedic relief, here on theBrett Madrona show, so to speak
(05:33):
, there's a lot that has alreadybeen said, so let me try to
focus on, maybe, what ought tobe said that might've not been
said.
Yeah, maybe what ought to besaid that might have not been
said, yeah, one of the reasonswhy AI has been adopted so
quickly, and maybe we'll coverthree reasons why the speed is
(05:55):
there.
First, you know, technology issomething that builds on itself,
and if you were involved intechnology at any point, you
must have seen some experiencewhere the thing that you were
doing at the time was then usedto create the next thing that
(06:18):
you're going to do and was usedto create the next thing that
you're going to do.
So, even as simple as like theFortran compiler right In some
ways, c++ was written in Fortranright.
Java was written using C++right.
Javascript was written usingJava right.
So, like you, use the toolsthat are commoditized to create
(06:39):
the next frontier, the nextadvancement that creates
competitive frontier.
Right Now, if you think aboutthe internet, a lot of plumbing
and infrastructure had to be inplace for someone to consume the
internet Wires on the ground,fiber optic cables, cell towers,
chips, electric capacityhardware had to be shipped,
(07:02):
multilingual things had to befigured out before you got the
internet.
So you know.
If you think about any thirdworld country I'll just pick
Guyana for a second you knowGuyana got electricity about 40
years after Europe and the USgot electricity right.
So you've kind of you kind ofare left behind for a while.
(07:25):
Artificial intelligence howeverthe platform was already there.
You had a mobile phone, you hadinternet already right, and so
you know the 5 billion peoplethat are already pre-connected
to the internet and have amobile were able to uptake AI.
(07:47):
From the time that you foundout about it to the time that it
takes to either hit the open AImobile browser application,
which was the first or downloadan app from the App Store, right
.
That was the uptake time, andthat's why you see this massive
uptake of individuals.
You see this massive uptake ofindividuals and I think what's
(08:10):
the next point as to why I thinkthe adoption was so easy is
that OpenAI did something thatis the core of the craft, which
is to productize it right.
Artificial intelligence, as itrelates to foundational models
and it relates to machinelearning and deep learning, was
there for a while, not as goodas the foundational models that
(08:34):
OpenAI's foundational model setforth, which started a
competitive race.
Not as good, but what was therewas that it was never
productized for mass consumption.
If you wanted to run a modelprior to chat GPT, you kind of
had to have Math Lab running,right.
(08:54):
You had to download a bunch ofstuff on a desktop, you needed
to understand how the fluxcapacitor of every button works,
and then the outcome was alarge text file with weights in
it.
It's just not accessible.
What OpenAI did, which is tostart a for-profit company to be
able to create funding to fundthe nonprofit company, was to
(09:17):
consume their models andproductize it as a chatbot, and
that productizing of a chatbotis, to me, the brilliance of
OpenAI, not necessarily the coreLLM underneath, although they
did start the arms race bydifferentiating significantly
when you make a product thateverybody already knows how to
(09:38):
use they know how to chat.
Who here doesn't know how tosend a text to someone and get a
text back?
Know how to send a text tosomeone and get a text back?
Then to find the right place todo it, which is these language
models, these foundationalmodels, and I want to talk about
probabilism after this.
They're imperfect.
How do you productizeimperfection?
A chatbot is a reallyinteresting way to productize
(10:02):
imperfection, because there's nogood answer, there's no perfect
answer.
That has to come back Right,and so that's another reason why
we saw that speed of adoptionof it.
And the third is something thatyou'd mentioned earlier in your
preamble here to this question,which is mankind now has a
thirst for technology and thefuture.
(10:23):
Mankind now has a thirst fortechnology and the future.
When we went through COVID,everybody, from my mom to you
know the guy who works in mylocal hardware store here in a
(10:43):
small town that I live ineverybody saw the power of
technology, whether it wasthrough Instacart deliveries or
the forms that we all filled outto get vaccines earlier or
going to.
I forgot what was the websitethat gave you all the stats of
how many deaths?
Speaker 1 (10:55):
Oh, yeah, I think it was the US government, or one of
them, whatever it was right.
Speaker 2 (11:00):
Everybody got to see the value of technology, not
just the sort of soft, flimsysort of magic of it or the
excitement of it or the dopaminehit of it or the vanity
advancements of it, but theactual utility of it.
Now, and we've all as anaggregate I would say as an
(11:28):
aggregate we've tripled ourawareness and our appreciation
for the utility of technology.
We come out of COVID.
Everyone's like okay, where'sthe future?
This is great.
You guys have been talking fortech so long.
We're ready for it.
What's up?
Where's the future, man?
So you have that thirst sort ofwaiting in society to adopt
this, and I think it's thosethree kind of pieces of that
(11:50):
storm that came together thatcreated this massive transition
of human awareness and thisreminder that we are evolving.
What AI really reminds us allat the core is that we are an
evolving species, we areconstantly evolving and we're
all living through the Internetagain, we're living through
(12:13):
electricity again, we're livingthrough the printing press again
, and so everyone, I believe, onthe planet in some way is now
aware of the fact that we'reliving in a pretty pivotal point
of history.
Speaker 1 (12:26):
There's no question about that.
I call it a leveler, let's say,like that of economies of scale
, where, like, let's say, ifyou're an oil baron, let's just
say you know you had a certainamount of control.
You know you had a certainamount of control.
(12:49):
If you're a real estate bearer,you had a certain amount of
control.
Technology has now given awhole new level of control to a
whole new set of individualsthat aren't oil bearers.
They don't own, you know,massive real estate, but they
have massive value.
For the first time in humanhistory we have an evaluation of
a company of a trillion dollars.
(13:09):
That's almost unbelievable.
When you think about it, peoplesay, well, you know, it's taken
certain people a long time toget what you would call economic
advantage.
And then I say you know what alot of these companies right now
, and then NASDAQ, when you lookat the Fortune 5, those top 10,
and we call them themagnificent seven haven't been
around for hundreds of years.
(13:30):
These are newer typetechnological advancements that
have taken these companies to awhole nother level.
When I think of Mobius this is,in my opinion, when I think of
when I first saw the AirGlass, Isaid, all right, somebody's
actually thinking out of the box, because it's almost like when
I saw the Google goggles that Icould say they came out for a
(13:52):
metaverse view and it's just Idon't know.
It reminded me of the 8-track.
I don't know how long that'sgoing to last, but we'll see the
8-track player.
But you said, hey, let's takethe glasses off and still have
what we want, which is digitalinteraction in a physical space.
It's kind of an expandedawareness of what we can do.
Right, very cool stuff.
(14:15):
Now you've come out.
I mean, this is a big word andthis is also taken.
First time I heard aboutgenetic AI was probably about a
year ago.
Gentic AI was probably about ayear ago and it's just another
explosive type tech where peopleare looking at it in many
different ways.
You've taken it and I like yourviewpoint.
You're taking like hey, and Igo back to the Microsoft and IBM
(14:38):
world.
I'm not reinventing a computer,I'm making it easier to use.
You seem to be making this kindof tech easier to use.
Speaker 2 (14:55):
What's your opinion on these things?
Well, look, my career has beenabout reinventing user
interfaces my entire career.
While I was at PrudentialSecurities, which is now
Wachovia, I worked extensivelyon connecting the CICS mainframe
green screens to the VisualBasic compiler to start to build
point-and-click apps forbrokers to be able to have
(15:16):
information, faster and betterworkflow as they serviced
investors to the brokeragebusiness of Prudential
Securities.
And that was the first patentand you know it got very
interesting between Microsoftand IBM.
But that was my first timerecognizing that user interfaces
will change.
The second time I did that wasat Lehman Brothers where
(15:42):
reporting so client reportingwas very static.
Everybody got the same PDF, thesame design of a PDF, and you
know I focused my energy thereon figuring out dynamicism of
PDFs.
How do you get a bankingstatement that has a pie chart
of your asset allocation on pageone but somebody else gets a
bar chart of their assetallocation on page one, but
(16:02):
somebody else gets a bar chartof their asset allocation on
page one because they want tosee their data differently or
their portfolio is betterdisplayed as a bar chart versus
a pie chart.
And that was the second time Ikind of reoriented user
interfaces from static reportsto sort of personalized reports
Again another patent, anotherkind of partnership with Adobe,
(16:24):
and today everybody haspersonalized PDFs.
I did it again at UBS I won'tbore everybody with that around
working with Apple to put thefirst wealth management app on
the iPad at the time of the iPadunveiling by Apple when you met
me.
The thesis there was that thescreens will become more
(16:49):
engaging, and as I was foundingthe company, I had decided to
call the company Meta M-E-T-A,and right before I set out to
kind of start, you know, takingthe company forward, I set out
to kind of start taking thecompany forward.
Facebook renames itself to Meta,and I had no interest in the
(17:10):
Metaverse.
I was pretty sure that we'renot going to strap toasters to
our head for a pretty longperiod of time.
And so in some ways, zuckerbergirritated me to get into the
metaverse by naming Facebookmeta.
Now, I'm nowhere the size andscale of Zuckerberg.
(17:32):
He doesn't know that I exist,but that's what happened.
He irritated the hell out of me, and so I thought well, what's
this stupidness anyway about themetaverse?
Why would anybody do this?
And so that was the first timeI actually thought, you know,
maybe there's something here andI looked at it and I said
(17:52):
there's no way the headsets work.
We ought to be able to get thescreens to give us more.
And the long and short of it,40 patents after I was able to
devise a way to use computervision, to use the screens that
you and I are looking at rightnow to be immersive and
interactive at the same time,and essentially, I built a
version of the metaverse thatdidn't need the headset.
I'm really not interested inthat.
(18:15):
And so as the metaverse windeddown and we were using computer
vision a lot, I started goingback to my original thesis which
is why I founded the companywhich is how can we get more
from the screen?
It has nothing to do withimmersivity and interactivity,
and so today, mobius is back toits original thesis.
I got sidetracked by Zuckerberg,who irritated me on the name,
(18:36):
and where we are today is that Iam looking again at the user
interface.
One of the net nets ofartificial intelligence is that
the user interface is going tochange, and you know, to put it
into context, google changed theuser interface by putting a
search bar.
The user interface of AI isgoing to be a microphone icon.
(19:02):
That's it Top, turn it on orturn it off, the mouse and the
keyboard will still have a place, similar to how engineers have
a place in the IDE today, rightor in black and white.
You know terminals, and so whatwe're doing at Mobius today is
(19:22):
we're really focusing on thatuser interface.
What is the intelligent userinterface?
Now, I won't talk too muchabout that solution.
I can connect with folksafterwards.
But the value for your audienceis to try to parse the
artificial intelligence markettoday to know where you want to
make your bets and generally,when these big industrial
(19:44):
revolutions come about and youwere talking about oil, it's the
same thing.
You can parse it into threeverticals.
You've got the producers of thenew utility, You've got the
appliers of the new utility andyou've got the consumers of the
new utility right.
So you've got the producers ofAI right and a whole supply
(20:06):
chain in there.
I put the silicon and thetraining in the production
bucket.
So NVIDIA and OpenAI I havethem in the same bucket, so to
speak.
Right, hyperscalers I kind ofpop them in there as well.
The appliers of artificialintelligence is where I sit.
Let's take the core utilitythat's being produced by others.
By all means.
Pay them their cogs, make thema part of supply chain, but
(20:28):
apply it to do somethingcompletely unique and different,
right?
Not dissimilar from how youknow.
Amazon applied the internetwithout spending a dollar on
building the internet, right.
And then you have the consumersof artificial intelligence.
What's also different here, andhow your audience wants to
navigate.
Whether you're a producer of AIand you know, be smart about
(20:50):
your capital investment there.
Whether you're a producer of AIand be smart about your capital
investment there and who you'recompeting against, and whether
there's a David and Goliath gameto play there, I argue that
there is a very low likelihoodthat a David and Goliath game
could be played out on theproduction side.
I feel like that side isalready locked up right On the
applier side.
Man, a lot of David andGoliaths are playing out on the
(21:13):
application side right.
Large software companies thathave been underserving their
constituents for decades are nowunder attack because people are
like wait, what the hell?
Why am I paying for a SaaSlicense that I'm not consuming?
Where the interest is is on theconsumption side of artificial
intelligence, and theconsumption side has two things
(21:35):
about it.
One is you can consume it tobuild businesses really fast I'm
talking about ridiculously fast, especially if you're building
businesses that have some coredigital component to it, right?
And why is it that languagemodels make it faster to build
(21:56):
companies that have digitalcomponents?
It's really simple.
Language models were trained onthe internet, and the internet
is not only full of information,it's full of software code.
When you read a website,there's JavaScript in there,
there is cascading style sheetsin there, there's HTML and DHTML
in there, right?
And so now language models arenot just good at our language,
(22:19):
they're good at the language ofthe machine, and so building
apps, building websites,building workflows, building
marketing streams are all aseasy to kick out as the answer
to you know what color is thesky.
And so that's what makes thisconsumption side so much more
interesting, because it loweredthe barrier to entry, and the
(22:42):
other part of it is theproductivity hack, right?
Not only is it helping youbuild businesses faster, it's
helping you do what you do todayfaster, and the internet didn't
have that acute differentiationon the consumption side.
Yes, it commoditized distancesignificantly, but you can't
build a business because youwere just able to commoditize
distance.
You still had to go through themechanics of a highly
(23:04):
capital-intensive sort of thing.
I think, when you look at theAI market from that perspective
and again I'm playing in thatapplying AI layer for the
benefit of the consumer's AIlayer, because I see the user
interface transitioning fromwhere we are today to where it's
going and that's where maybe wecould talk about probabilism in
(23:27):
that user interface, but Ithink for an audience in that
user interface, but I think foran audience, I think that that
is a reasonably good way toevaluate risk and opportunity
within where we are today.
Speaker 1 (23:46):
I want to expound on what you just said, especially
toward the end, understanding.
At the end of that whole thing,you have a human and then you
have what we are calling now AI,artificial intelligence, or
what I'm calling now a bettertechnology.
It's a tool.
Now I say this, remember thisIn our human history most people
agree humans took a great leapin their environment when they
(24:11):
discovered fire.
Ai is a type of fire and if youknow how to apply the fire
because you can take the fire,you can burn the forest down or
you can cook your food If youknow how to manage the fire,
you're responsible, accountable.
(24:32):
You understand the nuances ofthe fire, you know what it could
do, what it couldn't do.
You know that, hey, if I putwater on the fire, the fire is
going to go out.
But the fire also consumesoxygen.
It consumes a lot of thingsPeople don't realize that a
human being does.
No, you've got a.
Fire is a living thing, right.
(24:52):
So I said it to say this we arenow starting out with, from the
human component or reality,operating with a, a technology
that everybody can now utilizeand, to your point, I myself can
make businesses where before Ididn't have that capability.
(25:13):
Right now we're on an interfaceyou didn't have.
You know you have to have massresources to do radio, to do
television, to do mass media.
The barrier has beensignificantly lowered.
So if you, as a human, knowshow to operate or drive, I'm
(25:35):
going to say this technology,this AI, and you realize, like
you said before, the buildingblocks are already in place.
So you don't have to be a coder, you don't have to understand
(25:56):
how the Internet actually worksto a certain degree in the back
end, but you gotta know how todraw.
And if you're making it easyfor me to get behind the wheel
of the ai engine, I seeincredible transformation that
can, that can change the worldas we know it, and I've always
said this.
I have a lot of younger peoplethat ask, asked me a lot, and
asked my opinion like well,grant, what do you think my
purpose is and I always say thisespecially to younger folks
(26:19):
your purpose is to change theworld.
I've already lived in thisworld for 60 years.
I know what it's like.
I've gone through changes inhow we do things.
Your job is to change it evenfurther.
I'm going to point this to youbecause I like how you had that
comic relief right around theemoji right.
I remember sitting therewatching the Jetsons, and they
(26:41):
would have a communicator thatlooked a whole lot like your air
glass, you know, and it wouldjust be an image out of thin air
that you would interact with.
I truly believe that man'simagination, when it could be
productized, could be one of thegreatest advancements in human
history.
So I want you to, if you don'tmind, help me unpack some of
(27:06):
this, because what I'm saying isthis Human plus AI could be the
biggest game changer on theplanet.
Speaker 2 (27:13):
Yeah, yeah.
Well, I think, based on what weknow today, you know I never
underestimate 100 years out, youknow, there could be something
even bigger and no one knowswhat it is and no one knows what
it is.
Arguably it could beinterplanetary travel, right,
(27:34):
could be even bigger.
Again, thinking about thismight be an organics, a second
organic species, not a firstdigital species, right, as we
refer to it here.
Let's talk about the trade-offthat we're making, and I'm going
to focus on the trade-offspecifically from an enterprise
(27:59):
and a technology perspective.
I think your audience sits inthat center of gravity a little
bit.
So there is a fundamentaldifference between what we did
with the Internet and what we'regoing to do with artificial
intelligence, as it's specificto generative artificial
intelligence.
There's many other types of AI.
(28:20):
Computer vision, for example,is a different story, but when
you want to come to terms withwhat is different, I would
suggest that you start here.
In the internet world, we useddeterministic mathematics and in
deterministic math, 1 plus 1 isequal to 2.
(28:43):
And you can guarantee that 1plus 1 will always be equal to 2
.
You can guarantee it Withgenerative AI, which is
probabilistic math notdeterministic math, but
probabilistic math.
You cannot guarantee that oneplus one will be equal to two
(29:08):
every time.
You can't, because it's such anobvious thing.
One plus one will be equal totwo every time.
You can't, it'll be becauseit's such an obvious thing.
One plus one will be equal totwo a lot but you can't
guarantee that it'll be everytime because it's probabilistic.
There's no such thing as a rightor wrong answer.
The only thing that's trueabout probabilistic math is that
(29:31):
it must have an answer everytime, and so that's why, when
you ask a probabilistic modelfor something and language
models are sort of like thecodfish example everybody knows
what I'm going to use here.
But let's take an image model.
If I ask an image model tocreate me an image of a blue and
(29:52):
red dog, it'll give me anawesome blue and red dog.
Try to get that same dog again.
Speaker 1 (29:59):
Yeah.
Speaker 2 (30:02):
Zero chance to get that same dog again.
Now we have the same thing inlanguage models.
Ask it a question.
If it's not too obvious of aquestion, try to get the same
response every time.
Zero chance.
And so, living in thisprobabilistic world right, where
you're treating off adeterministic relationship with
(30:23):
technology for a probabilisticrelationship with technology,
now you start to see thedifference of the mechanics of
it.
For example, if we just takeuser interface for a second, in
a deterministic world you havean application that could do
about 10 things, okay.
And you must move your mouseprecisely to here to click on
(30:46):
this one button to do those 10things.
And you must wear your readingglasses to read the little text
that's on the side to know whichof those 10 things to do those
10 things.
And you must wear your readingglasses to read the little text
that's on the side to know whichof those 10 things to do, okay.
So there's a high degree offriction to a deterministic
relationship with technology.
Okay.
However, it's guaranteed thatit's going to do 10 things and
(31:08):
it's going to do those 10 thingsexactly the same way every
single time, bar none.
In a probabilistic relationshipwith technology, the number is
not 10.
They can do a billion differentthings.
However, the perfection is notthere.
Okay, so we are trading off ahigh friction, perfect, narrow,
(31:45):
deterministic experience with alow friction, wide, imperfect
experience, and that's thetrade-off.
It's because you've gotprobabilistic math on one side
and deterministic math on theother side.
Now the artistry and the craftis to figure out the balance of
(32:08):
those two to be able to makesomething of utility for mankind
.
The challenge is that as soonas you add one ounce of
determinism into a probabilisticsystem, you lose a significant
amount of the probabilisticvalue, like, if you say, to a
(32:28):
generative model.
In this scenario, you mustgenerate the same thing every
single time.
You essentially stifle theintelligence to the ground Right
, and that's the core that wehave to understand.
Speaker 1 (32:46):
I'm glad you brought me back to one of my most
favorite things I ever studiedin high school, which was
physics.
I love physics.
I wasn't great at it I gotactually a C in the class but it
was my favorite class of allbecause of those things you just
talked about.
(33:06):
My teacher would tell us allthe time there is no right and
wrong answer.
I'm more interested how you gotto your answer than what the
answer actually is.
I want your critical thinkingand and and come up with
something that's interestingbecause you can get outside the
box, just like you were sayingso sometimes and I always liked.
(33:30):
I think Einstein brought it upwhere, as you pointed, one plus
one equals two, but in his worldit doesn't.
It equals what we call three.
What do you mean by that?
Because you by yourself, you'reone entity.
Me by myself, I'm one entity,and if we both operate in those
separate worlds, that're oneentity.
Me by myself, I'm one entity,and if we both operate in those
separate worlds, that is oneplus one.
You could kind of like two, butwe come together.
(33:51):
There's a third reality thatcomes about.
I'm going to call that thispodcast discussion right Now.
We both could never reallydetermine exactly what we were
going to say and do and theanswers, but it becomes a third
reality.
That's that plus one.
I believe there's a plus one ineverything in life.
There is, it's just are youbold enough to engage with that
(34:13):
reality?
As I said earlier, the humanplus the fire.
If you know how to manage thefire, it's a great tool.
But if you don't, you know, youdon't know exactly how it's
going go Ask Elon Musk in hislast billion dollar explosion
that happened.
There's a lot that couldpotentially go on.
So, as men, we're starting toget out of our deterministic
(34:37):
world to a certain degree,because everybody likes to stand
on solid ground, right.
But now you're saying I'm goingto stand on solid ground, right
, but now you're saying I'mgoing to stand on something I'm
not sure if it's solid, could beor couldn't be, but how do we
get there?
It's going to be a veryinteresting feature, I believe,
and a very near feature.
Speaker 2 (34:55):
I want to just follow on to that very quickly Go
ahead.
I was doing some work with theboard and to engage the board, I
started by asking you know,deterministic math was arguably
discovered by Newton in the1650s.
(35:18):
1650s or 1850s, I don'tremember a 50, it won't matter
to the story.
1850s, I don't remember a 50,it won't matter to the story.
Um.
And then I asked when do youthink probabilistic math was
invented?
And everybody was wrong,everybody oriented to much later
in history.
You know like within the last50 years, right, people make
(35:40):
bets there.
It turns out probabilistic mathwas arguably discovered by
pascal within 10 to 20 years ofdeterministic math from newton,
really around the same time inthis thing that you said that
that we need, we need, um,determinism as as a species.
(36:02):
Um, what happened at the time?
Is deterministic math picked upbecause the scientific
revolution needed determinismand probabilistic math never
picked up because of the simpleargument that God has to exist.
There can't be any questionaround whether God exists.
And so probabilistic thinkingand probabilistic math, because
(36:25):
there are no guarantees in thesystems, ended up to be seen as
like a little bit of a dark arts, if I may right, a little bit
of like a spooky kind ofsituation, a little bit too
gamblish, and society at thetime was not ready for
probabilism.
You know, are we ready forprobabilism today?
(36:46):
And it's going to come down toa couple of things.
It's going to come down to onewhat are the areas where
probabilism works?
And let's touch on how did weget to agents?
How did the concept of agentscome about, which led to the
(37:09):
largest word washing of mycareer.
Talk about cloud washing andmetaverse washing and mobile
washing and internet and onlinewashing.
I've never seen washing likethis, an online washing.
I've never seen washing likethis.
But when you think about thedeterministic systems and the
probabilistic systems, again,right.
(37:30):
And you say, okay, so a lot ofour utility to build things that
people will use sit in adeterministic world, right now
exposed by API world, right nowexposed by API, right, yeah.
And then you've got thisdeterministic UI which is, if I
(37:53):
click buy, it's incrediblydetermined that you must go call
the buy API, right.
Or if I click sort, you shouldgo call the sort API.
So this very, very rigidrelationship between actions in
the user interface and executionof that intent in what we could
argue to be called a system ofrecords, now a concept evolved
(38:19):
in between the deterministicuser interface and the
deterministic APIs and in thatspace in between is called model
context, which is if the UI isgoing to be driven by a language
model or a conversational model, right, how does that
probabilistic UI calldeterministic APIs?
(38:42):
Yeah, how do you know thatsomeone meant to buy?
Because the intention to buy isno longer determined by
clicking the buy button.
There could be a million waysthat you indicate that you want
to buy.
You could say, all right, let'sgo, and that means that you
want to buy.
You could go, I'll take it, andthat means that you want to buy
(39:05):
.
Right.
And so model context protocol isthat thing that says figure out
what that person said and then,based on all APIs that you have
available, that's where themodel context, the MCP servers,
come in, which is hey, what areall the APIs that are there and
(39:26):
how are they interconnected?
Describe them in an MCP serverso that when the language model
hears I'll take it, it does atranslation of that.
Look at the API and go.
The only API that makes sensehere is the buy API.
So the MCP, this model contextprotocol, is this membrane
that's sitting between APIs andbetween a UI that's starting to
(39:50):
expand, that uses probabilism inthe middle of two deterministic
systems, and that's where theperception of agency is coming
from.
There's nothing sentient oragentic about a system.
The concept of an agent ishorseshit.
(40:13):
That probabilistic layer inbetween is creating the
perception of autonomy.
How could it have known that?
When I said I'll take it, Imeant buy.
It's really simple.
There's nothing agentic aboutit.
I haven't met an agent as yet,other than, like you know, my
agent at the airport when I'mtrying to change my flight.
(40:33):
I have not met an agent as yet,and I'm deep and wide in the
space.
It's complete horseshit.
Speaker 1 (40:41):
I love how you said that.
I think what and how I look atit is like this and we're going
to conclude here, but I had thesame, similar, somewhat, not as
technical as you just describedit, but what I was explaining to
people was like, when we'retalking about in an AI-driven
world, we're talking aboutsomething that's just mirroring
human intelligence or behaviorto a certain degree, as we
(41:05):
understand ourselves.
So you're asking something.
Here's the three things.
I look at Something that isunconscious, that is, the AI
itself.
It's unconscious to someonewho's conscious, which is
yourself.
But then also you are operatingin a subconscious world a lot
too.
But then also you are operatingin a subconscious world a lot
(41:26):
too.
When you start talking abouthow are you going to understand
behavior or what drives you, itgets into deeper layer.
So then you start seeing, yeah,in our world, in the deeper
context of what you call thenatural world and how we operate
in everything around us, it'snot so much deterministic.
There's a lot going on, a lotmore intelligence than just
(41:50):
human intelligence that we'reoperating in.
But we're starting to discoverthis and then we're going to see
how we're going to interoperatewith it, let's say, collaborate
and understand how all this isgoing to work together.
But you brought up some very,very good points and that at
some point in time, hopefully inthe near future we're going to
have another conversationbecause we're ever evolving.
(42:11):
You're ever evolving and beforeI conclude, I want you to just
kind of sum up what you feelabout this particular podcast.
How do you feel about theFollow the Brand?
Speaker 2 (42:23):
how I feel about.
Speaker 1 (42:23):
Sorry, the last word I missed yeah, follow the brand,
the follow brand podcast.
How do you feel about thisinterview?
Speaker 2 (42:29):
look, I've been tracking you for a while and
what I like is your consistencyand your perseverance.
Um, a lot of people build thesetypes of um in some ways
philanthropic platforms to beable to help others along the
way because of privilege oraccess that you may have.
(42:51):
Um, and most people fall offbecause it's hard.
It's not easy.
Getting guests is hard, doingit every week is hard.
Getting the courage courage tobe able to engage in it and put
your own brand out there is hard, and so I think your
(43:12):
consistency and yourperseverance marries what I do
in the early stage venture world.
In the early stage venture world, it's really only two
ingredients Perseverance, yeah,and craziness.
You have to be crazy A littlebit out there, yeah.
In some ways and I say thisself-deprecatingly in some ways
(43:34):
you have to be a special kind ofstupid to do something like
this, what you're doing and whatI'm doing.
You can't be born with thistype of stupidity.
You have to cultivate it right.
Nature can't make you thiscrazy to do things like this.
So I really commend your, yourperseverance and your commitment
.
That's what makes you know BDMwhat it is.
(43:55):
Thank you.
Speaker 1 (43:58):
I truly appreciate that Before we let you go, you
got to let us know how tocontact you, and you also said
you're going to be upcoming.
You're writing on Forbes,you're doing some things for
Wire.
Tell us what you've got goingon, richie.
Speaker 2 (44:11):
Well, I think a lot of what I'm going to be doing is
as we get ready to come out ofprivate beta.
We're working with about a halfa dozen or so Fortune 500
companies in a private beta atMogus Very, very exciting
approach.
Again, we're transitioning theentire user interface that you
have, any website that you'veever had before, any document
(44:32):
that you've ever had.
We're changing it from a guidedhigh-friction experience, I'm
sorry, from a self-directed highfriction experience.
So if you think about marketing, you spend a lot of money to
get people to your website andthen you leave them alone.
Self-directed, right.
We're moving that to guide itby putting this perception of
(44:55):
personhood in it, so that youfeel like you're guided when you
arrive at the website.
What would you like?
How can I help you?
What do you want?
Right, using that sameprobabilism in the middle to
create the perception of agency.
To understand something, likeyou know I'm ready to take, it
means that you want to buy.
So the work that I'm doing rightnow is to start to bring
forward the insights thatinformed the Mobius Telly, the
(45:21):
instincts that informed theMobius Deli, the deep
understanding that informed theMobius Deli because I think part
of it is I am dying to sharewhat I've learned and kind of
picked up over the last four orfive years with the rest of the
industry.
And two is in a world whereyou're launching an artificial
intelligence product, there's alot of founder gravity that
(45:43):
comes with it.
So I'm getting into wires.
A wired Forbes has finallynudged me enough to start
writing again.
I got a piece in the Economistcoming.
I'm doing some board work, I'mdoing some conference circuits
and stuff like that, because youknow I have to come out first
before the product comes out.
Speaker 1 (46:02):
I like that.
See that personal brand, seethat Everybody asks me why
personal branding?
What is business branding like?
You have to have a face of yourbrand to communicate
effectively and simplistically.
I mean you have a lot of.
You simplify very hard conceptsand I think a lot of people are
like, wow, I think I understoodwhat he just said this last
(46:22):
hour right?
I think that's wonderful.
I am going to send you picturesfrom Guyana when I get down
there, because this is going tobe wonderful.
Take you home just a little bit, show you what I'm seeing down
there and then hopefully we'llcircle back before the end of
the year and have another good,intelligent conversation.
And I want to thank you againfor being on Follow the Brand.
(46:44):
I want to encourage all of yourfollowers to follow us at
5starBDM.
That's the number 5.
That's star.
Bdm is B for brand, d fordevelopment and for masterscom.
I want to thank you again forbeing on the show.
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.
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
The Breakfast Club
The World's Most Dangerous Morning Show, The Breakfast Club, With DJ Envy, Jess Hilarious, And Charlamagne Tha God!