August 1, 2025 • 39 mins
Nils Pihl is the CEO and founder of Auki Labs, a deep-tech company redefining how augmented reality (AR) can power meaningful social and business experiences. What started in 2019 as a passion project to create AR overlays for tabletop games like Warhammer quickly evolved into a multimillion-dollar company at the forefront of spatial computing. Nils shares his conviction that new business models and decentralized technology will fundamentally reshape software—and even capitalism itself.
Key Takeaways
- Start Small, Think Big: Auki Labs began with playful AR overlays for games but scaled quickly by focusing on community ownership and decentralized models.
- Decentralization is the Future: True value creation in software will come from platforms owned and operated by their users, not just corporations.
- Augmented Reality, Real Impact: AR is at its best when it enhances live, shared experiences—turning the virtual into a collaborative tool for everyday life and business.
- Innovation Through Openness: Embracing open-source and collective funding accelerates innovation and broadens the impact of new technology.
Beyond the technology, Nils’s philosophy challenges traditional business frameworks: he explores open-source, collective governance, and decentralized funding models as the future of innovation. He remains motivated by the “software upgrade of capitalism,” aiming to empower people—not just companies—through technology that invites participation, not just passive consumption.
About Nils Pihil
Nils Pihl is an entrepreneur, behavioral engineer, and social transhumanist with a deep focus on the intersection of modern technology and human behavior. Nils founded Auki Labs which is advancing the development of the Auki Network and posemesh protocol, a decentralized machine perception network designed to help devices, robots, and embodied and physical AI perceive and interact with the world around it through a shared understanding of space.
A thought leader in the realms of technology and memetics, Nils applies his expertise in meme theory, game theory, and behavioral psychology to craft compelling narratives that shape the path of the AI revolution, future of augmented reality, and the acceleration of robotic technologies.
About The Glow Up
A "glow up" signifies a positive transformation, reflecting the journey of becoming a better, more successful version of oneself.
At The Glow Up, we humanize the startup and innovation landscape by focusing on the essential aspects of the entrepreneurial journey. Groundbreaking ideas are often ahead of their time, making resilience and perseverance vital for founders and product leaders.
In our podcast, we engage with innovators to discuss their transformative ideas, the challenges they face, and how they create value for future success.
A "glow up" signifies a positive transformation, reflecting the journey of becoming a better, more successful version of oneself.
At The Tech Glow Up, we humanize the startup and innovation landscape by focusing on the essential aspects of the entrepreneurial journey. Groundbreaking ideas are often ahead of their time, making resilience and perseverance vital for founders and product leaders.
In our podcast, we engage with innovators to discuss their transformative ideas, the challenges they face, and how they create value for future success.
If you're a founder or product leader seeking your own glow up, or a seasoned entrepreneur with stories to share, we invite you to join our guest list via this link.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Nathan C Bowser (00:00):
Hello and welcome to The Glow Up.
I'm Nathan C and today I amtalking with Nils Pihl, CEO of
Auki Labs.
Nils, it's great to talk withyou today.
Thanks for joining!
Nils Pihl (00:11):
Happy to see you here online! We met again in real
life not long ago at AWE.
Nathan C Bowser (00:17):
It was so good seeing you at AWE and I can't
wait to dive in and hear alittle bit more about the
inspirations and work that wentinto that very dynamic and
interactive booth that you hadthere.
Before we get into maybe thecoolest spatial computing event
of the year, let's wind back alittle bit and could you
(00:38):
introduce yourself and the workthat you do in innovation at
Auki Labs?
Nils Pihl (00:43):
Sure! So I am Nils.
Auki is a company focused oncollaborative machine
perception.
The idea is that digital deviceshave to have a shared
understanding of the physicalworld.
It started out very innocentlyin 2019 with us trying to make
(01:04):
augmented reality overlays forthe tabletop game Warhammer.
That's where we got started.
And, what we ran into as aproblem was that shared
augmented reality was verydifficult.
Like we want to be able to see ashared overlay over the table.
And I didn't know the firstthing about augmented reality
(01:26):
back in 2019 when I started thecompany.
So I naively imagined, like Ithink many people imagine, that
AR was in a much better statethan it actually is.
And I had the cold shower ofrealizing that like, oh the GPS
is actually not fit for purposeat all.
You can't use the GPS forWarhammer.
We need millimeters ofprecision.
(01:48):
And the GPS doesn't even knowI'm in this house right now.
But it was COVID, so rather thangiving up, I had my Warhammer
and I could just sit and lean onmy autism for two years and
think about how do you solvethis problem?
And over those two years I cameto realize that a very big
(02:08):
problem for the future that Ihad always imagined, you know, a
future with AR glasses androbots, is that digital devices
just don't understand thephysical world very well, and
they don't have a sharedunderstanding.
By 2021, I had inventedsomething that made it possible
to do what I wanted to do withthe Warhammer thing, and that
(02:30):
helped me raise a very generousamount of seed funding to build
an open source collaborativeprotocol for all kinds of
machines to see the worldtogether.
And this year, we finallylaunched an SDK at AWE, that is
(02:50):
focused on collaborative machineperception.
And we launched our firstproduct built on that SDK
earlier this year.
A spatial AI tool for retailersand we just recently broke 2
million ARR on that product andhope to hit 10-15 million ARR by
(03:11):
the end of this year on thatproduct.
Nathan C Bowser (03:15):
I think we just discovered the title of this
episode, which is somethingalong the lines of"From
Warhammer to 15 Million in ARRwith Augmented Reality and
Robots." Actually we gotta putthe time-based in there so
"Within Five Years" is a prettyamazing journey to capture!
Nils Pihl (03:38):
Many people predicted that if I spend too much time
playing Warhammer, I will end upworking in retail.
And look I did!
Nathan C Bowser (03:45):
This is very much like the entrepreneur's
curse, you fulfill thosepredictions of other people.
You just do it in a way thatthey've never imagined, right?
Like, so I'll be working totransform the back of house
supply chain operations forretail, not just helping people
find where the lettuce is.
(04:07):
That was really fantastic Nils!I love how this journey kind of
starts innocently, includes alot of special interests.
A lot of like inspiration.
You mentioned the seed round andthis goal to produce open source
technology to connect people androbots and AI to an
(04:28):
understanding of the world.
Building a business around opensource feels like it has extra
complication.
I'm curious and might not alwaysbe the thing that like venture
investors are looking for.
Can you talk a little bit about,why open source for something,
(04:49):
so cutting edge and on thefuture of where it seems like
technology is going?
Nils Pihl (04:54):
Yeah, but now we're really going on a trip here! So
I believe that the internetitself is missing a protocol.
The internet itself is missing away for machines to shake hands
with each other and say, here'swhat I know about the physical
world.
And I really want for oursolution to become a part of the
(05:21):
internet itself and not become abig corporation.
The reason why was I had asomewhat mushroom induced moment
of clarity when working on theAugmented reality solution and
thinking about the AR glassescoming, right?
So.
(05:41):
You know, go back to 2020 or soand I'm learning about how will
we do positioning in a post GPSworld?
And by 2020 there were only twocandidates that I could see it
could possibly be.
Either we're gonna do itentirely blind using ultra
wideband trilateration.
(06:03):
You know, the AirTag was prettynew back then and ultra wideband
seemed like a pretty promisingway forward where, you know,
maybe we could even build a pairof AR glasses that didn't even
have a camera! Maybe we could.
Nathan C Bowser (06:16):
Kind of like the earlier teens idea of IoT
driven by beacons, right?
Nils Pihl (06:21):
Absolutely!
Nathan C Bowser (06:22):
But that was back on Bluetooth.
Nils Pihl (06:23):
Yeah ultra wideband was promising to maybe save the
idea of the Bluetooth beacon.
Do something that could be ahigher precision.
The other alternative, and Ithink what the industry had
already settled on, you know,people didn't believe ultra
wideband and now I don't think Ido either, was visual
(06:45):
positioning.
The idea that the device tellsyou what it's looking at and you
tell the device where it is.
What visual positioning is, isthat you compare an image feed
of what the device is seeing toa model of the world that you've
previously made so that you cantry to match where was this
frame taken and saying the quietpart out loud means that you're
(07:08):
quite literally telling someonewhat it is you are looking at
for this service to be able towork.
And that's not the scariestthing in the world when it's our
phone and we only whip it outevery now and then.
But when we're wearing ARglasses all day, the thought
that Mark Zuckerberg or JohnHanke or any one of these guys
(07:29):
would be able to know what mykids are looking at at all
times.
I realized that the positioninginfrastructure needs to be
privacy preserving, otherwisethe augmented reality future
that I was so excited about andI was listening to a lot of
Terence McKenna at the time andI was imagining the cyberdelic
(07:50):
future.
I could see how it could veryeasily turn very dystopian if we
build this under the traditionalSilicon Valley surveillance
capitalist business model.
And we would have to find someother kind of business model.
Luckily for me, I had a coupleof early backers that were very
(08:13):
big into Web3 and DePIN whobasically bullied me in 2021
'cause I am, I think to thisday, very largely anti-crypto.
But they bullied me into lookingseriously at what it is that
(08:34):
Web3 could enable you to do,both from a technological point
of view, but also from abusiness model point of view.
And that you could build abusiness model that can survive
the corporation and become apart of the internet itself.
And by late 2021, I had embracedthat fully.
(08:55):
That like, yes there is a lightat the end of the tunnel.
In all of this Web3 noise, thereis a promise of a new kind of
business model, civilizationscale infrastructure funded,
owned, and operated by thepeople.
And that is radically exciting!Those investors have no qualms
(09:20):
with open source at all.
In fact, they think open sourceis fantastic, right?
The decentralization communitydon't want power to be
centralized with anyone and arevery happy for the technology to
be open source.
So I, accepted their support.
They gave us almost$20 millionin seed funding to build this
(09:42):
thing.
And that also was veryconvincing because normally
traditional VC would not give 20million to an idea that was so
early.
And it gave us an opportunity totry to do things that normally a
startup would not be able to do.
(10:03):
Part of the selling point, likeoh my god! Web3 has the
capability to crowdfundinfrastructure projects that no
VC would.
Well, not no VC.
There are certainly some VCsthat write these checks every
now and then, but it's not easyto get 20 million in seed
(10:23):
funding from traditional VCs.
But being able to have this newbusiness model and have, again,
civilization scale projects thatare funded, owned, and operated
by the people deeply red-pilledme into a potential future
software upgrade of capitalismitself.
(10:46):
This is an exit for surveillancecapitalism, we can build another
kind of company.
And now I'm all about that.
Nathan C Bowser (10:56):
Hell yeah! I love that a very sort of, almost
cute idea about augmentedreality led to such an expansive
understanding of theinfrastructure that goes into 3D
like coming economies powered byrobots, and like what that's
(11:18):
gonna mean.
This very compelling idea thatlike traditional models for
funding in startups
Nils Pihl (11:27):
are
Nathan C Bowser (11:28):
extremely risk adverse and have maybe not been
the greatest predictors of likewhere things are going.
And that crypto, you know, Web3is like one of the areas where
there's a lot of capital andinvestment that's going on
pushing some of these new modelsand, I think really the most
(11:51):
exemplary thing that I want tocall out is when you're thinking
about the problem space, whenyou're thinking about the value,
when you're thinking aboutdelivering something that will
unlock even just like the nextthings you're trying to do in
your work that by reallyupleveling where you take your
(12:12):
solution, right?
It's not just a Web3 platformthat you can use in Auki.
It's literally giving the entireecosystem a missing piece and
allowing that to likepotentially change the way that
we interact with technologies.
Beautiful artwork by KeiichiMatsuda in that hyper reality,
(12:35):
corporate augmented dystopia.
But like man, I don't personallywant a single corporate oligarch
owning all my spatial data!
Nils Pihl (12:46):
Last year Elon was on Lex Fridman's podcast and he was
asked,"What will it take to winthe AI race?" And Elon said it
will take two things, right?
You need to have the bestcompute.
That's why he's building thisdata center in Houston or
wherever, and you need the bestdata.
And what he said was,"Oh we'vealready, you know, consumed all
(13:07):
the data on the internet."Turned out there wasn't that
much data on the internet tobegin with.
But reality scales, he said.
And then he said something thatshould have scared us all, which
was Optimus, his humanoid robotis going to be the greatest
source of data in the world.
Without any shame he just goeson Lex Fridman's podcast to say
(13:28):
like, hey, I'm building a robotto spy on you in your homes.
And I actually think that theproblem with the traditional VC
model is not that they are riskaverse because I mean, clearly
they do take risk and they losea lot of money making bad bets.
I think it's a more perniciousproblem of group think around
business models.
(13:48):
To this day, we get a lot ofpressure from VCs that like,"Why
are you letting your customersown this data?
You could own this data.
Imagine if you owned all thisdata." It's like, no exactly!
Imagine if I owned all thisdata.
That would be terrible! Weshould let our customers own
their own data.
And it is possible to buildother business models.
(14:10):
You know,"Data is the new oil"is so deeply ingrained in
Silicon Valley, that they oftendon't take anything seriously
that is not an attempt tocollect data.
Like a good friend of minerecently raised a round for
building smart glasses and I'man advisor to his company and I
(14:31):
got to listen in to a couple ofthe VC pitches and there's a lot
of pressure to just be like,"Areyou going to be collecting the
data from the glasses?
That's going to be so valuable."The call is coming from inside
the house.
The VCs are behind a lot of theexcesses of surveillance
capitalism, and I think they'vejust gotten caught in a group
think loop that this is the onlybusiness model that could work.
Nathan C Bowser (14:55):
One of the things that I talk a lot about
at Awesome Future is this ideathat innovation usually fails in
the last mile connection fromthat good idea to where the
customers actually are.
And I think your conversationaround like owning people's data
is one of those really clearinstances where like in a
(15:18):
boardroom level, at a VC level,obviously owning the data is a
resource and you wanna have asmany resources available to your
org as possible.
On the flip side, right you'realso calling out that spatial
experiences, AR, VR, even likeany experiences, any digital
(15:39):
experience at this momentdoesn't have enough metadata
attached to it to actuallyinteract with the world or to
really know like where it is andwhat it's supposed to be doing
beyond little actual boxes thatwe tape on the floor for robots.
And so spatial experiences,digital experiences are lacking
(16:00):
key understanding and metadatajust period.
And you're trying to own it all,even though it's like it
incomplete.
There's just gaps in reallyunderstanding like what happens
in that last little bit and, wewant all the data, but at this
moment it is incomplete anddoesn't know what it's supposed
to be doing.
Is that actually as valuable asyou think it is?
(16:21):
And, I love this position thatAuki's taken which is, it's
actually way more valuable if weincorporate this in this kind of
spatial data, this awareness ofwhere you are, who you are.
Oh my goodness!
Nils Pihl (16:35):
A lesson from history that I think VCs should
remember, right, is in the goldrush, you should be selling
picks and shovels.
It's not about the gold itself,right?
If data is the new oil, thatdoesn't mean that the game is to
be collecting all the data.
Make picks and shovels that makeit possible for your customers
to collect data that's valuablefor them.
Nathan C Bowser (16:54):
The internet is full of basically 2D data,
right?
It's text, it's images, it'spixels, it's spreadsheets.
You know, when you actuallythink about adding a third
dimension to your data, right?
Like you start getting theseexponential growths in what's
going on.
When I started in XR about adozen years ago the conversation
(17:19):
was like, we're gonna usemachine learning to have a
semantic understanding of thephysical world.
Are we just renaming what'shappening now?
At AWE, I would say one of thebiggest themes that I saw there
was that every computingplatform, whether it's Niantic,
whether it's Qualcomm, whetherit's Auki Labs was talking about
(17:44):
this idea of"geospatialintelligence" or a map of the
world that humans and robots canboth understand.
Has this idea, you know,actually changed from the terms
we were using a dozen years ago?
Or is there something likereally unique about what's going
(18:07):
on in this moment?
Nils Pihl (18:08):
Oh, it's a little bit of column A and a little bit of
column B.
I think that, the spatialcomputing industry has run out
of money.
Which is a big problem.
I think that one of the thingsthat I noticed at AWE this year
and last year is that, there'sbeen a while since there were
(18:30):
big cash infusions into the XRecosystem.
A lot of companies are lookingfor the exit.
A lot of companies are on theirlast bit of gas and as such,
they are in fundraising mode andare going to speak the language
(18:51):
that the VCs want to hear.
And this year the VCs want tohear about robots.
They didn't want to hear aboutrobots in 2021.
Like I told people about robotsin 2021 and they were like,
"Please just talk about themetaverse." But like in my
original pitch in 2021, Ipointed out that look— an iPhone
is just a robot without arms andlegs.
(19:11):
We can get started building thespatial awareness right on the
iPhone.
We don't have to wait for theglasses.
We don't have to wait for therobots.
We can build this for the iPhonenow.
And when the robots are readyand they're like,"Robots?!
Metaverse!" And now it's like,"Don't talk about the metaverse!
Tell us about robots." Butthere's also I think, a couple
of interesting new developmentsin tech that are making these
(19:36):
things a lot more realistic.
I think people have had thevision of this spatial semantic
understanding for a long time,but there's been so much
homework to do, and in just thelast two years, there's been an
explosion of things that make itpossible, and a lot of it is
open source.
But I also wanna give creditwhere credit is due.
(19:56):
They've put out a lot of amazingopen source research.
And China has been putting outjust enormous amounts of open
source research when it comes tomachine perception.
Now, things that were completesci-fi, like people literally
telling me in 2021, it's inprinciple impossible.
Things like monocular depthestimation.
(20:18):
I had people tell me were inprinciple impossible has gotten
surprisingly good.
Tencent and ByteDance bothreleased open source, monocular
depth estimators that are crazyand can run in almost 60 frames
per second.
It's remarkable! A little bit ofcolumn A, a little bit of column
B.
I think the language has changedbecause there's a little bit of
(20:41):
an era of desperation in the XRcommunity because, we had our
hype cycle, we got a lot offunding during the hype cycle,
and there hasn't been follow upinvestments since.
There are very few activeinvestors in the XR space, and
they are all in the pre-seed andseed stage.
Who writes a Series A check foran XR company, right?
(21:05):
So that's a problem.
Nathan C Bowser (21:06):
They have to have"XR AI" in their name.
Nils Pihl (21:10):
Yes! "Something-something Robots."
Nathan C Bowser (21:12):
Maybe that's the title of the show! The
market changing its interest,investors like shifting where
they're funding is probably oneof the most dramatic inflection
points, learning moments,decision points that a founder
is going to have to encounter.
(21:33):
So, through your experience inspatial computing and the amount
of like shifts of attention,even in the last five years.
I'm curious like how you'veapproached, maintaining your
core vision while adjusting tothe market, or if you have
advice for folks who are on thisjourney and experiencing so much
(21:58):
change and disruption in thecurrent models.
Nils Pihl (22:01):
There are two sides to this answer.
I have to be honest about havinghad to make a lot of pivots on
the way to where we are today.
It wasn't just from Warhammer toretail, it was from Warhammer to
AR pets, to retail.
One thing that has been constantsince 2021 at least, is this
(22:22):
vision of a collaborativemachine perception network.
How to get there has changedsomewhat over time.
Frustratingly it has changedback a lot to what it originally
was in 2021, which is also apainful lesson of who you should
(22:43):
listen to and not listen to.
When you take on investor money,one thing that you open yourself
up to is that they are gonnastart giving you advice.
And one piece of advice thatI've been given is to not say
what I'm about to say, but I'mgonna say it anyhow.
A lot of VCs have never builtanything themselves.
They have backgrounds inbanking.
(23:05):
They've never shipped a product.
They've never had to fire 20people.
They don't know anything reallyabout what it means to be an
entrepreneur, and they don'thave to be fair, you know, their
job is not to be anentrepreneur.
But they will look at the worlddifferently'cause their job is
(23:27):
not to make your ideas like cometrue.
Their job is to get a positivereturn, and it doesn't matter to
them if you end up pivoting orwhatever, as long as they get
their money.
So we've constantly been gettingpushes to pursue the latest hot
(23:48):
buzzword from our investors.
In early 2021, one of the firstthings that we thought, you
know, what are you gonna do withcollaborative machine perception
was like, well we could starthelping people navigate in
warehouses and retailenvironments and we could make
it easier to deploy robots forthese things.
(24:11):
And investors came and saidlike, no, no, no.
You need to make a B2C product.
And we had a first demo, just atech demo, which was an AR pet
that you could see in sharedaugmented reality.
Like it really just, it was justan art asset, Nathan.
Like there was no game there.
It was just like, here's ananimated animal and we can see
it together.
And they're like,"You shouldtotally build that! You need to
(24:32):
build it, build it!" And we did.
And it turned out, you know,that an AR pet was actually
worse than a 2D pet because wedidn't have collaborative
machine perception yet, right?
Like we had cracked, having ashared coordinate system so we
could see the pet together.
And that was great, but the petcouldn't see its environment.
(24:55):
So we would walk through chairsand things like this that would
break the immersion.
So you were literally better offnot using the AR.
In hindsight, you know, if wehad just stayed the course and
done what we wanted to do, wecould have launched this retail
thing two years earlier maybe.
So, you know, it's hard becausebeing an entrepreneur is also
(25:15):
very scary.
But now I've learned that look,your job is to be the
entrepreneur.
And their job is to find theright entrepreneur, not to shape
the entrepreneur.
And they should be okay withpotentially losing money.
You are supposed to take aventure scale swing, right?
That's your obligation as anentrepreneur.
You need to take a venture scaleswing, and you need to
(25:37):
understand that it's okay tofail.
As long as you're trying to dosomething that is venture scale.
And that's a little bitcounterintuitive because you
know, you feel like, oh maybe ifI take this less risky path, I
could have a profitablebusiness.
Yeah but a profitable businessis not what they are looking
for.
They're looking for a venturescale return.
Nathan C Bowser (25:55):
Oh my gosh! There is such a terrible tension
in innovation, in technology.
This is what we know, this iswhat feels safe, and this is the
size of the swing we need totake, are so often on such
(26:17):
different ends of the spectrum.
And this idea that taking theeasy route or taking the advice
of other people above you as anentrepreneurs, that vision, that
unique perspective on the futurethat only you see and what your
intuition, your team can build.
(26:39):
That you're kind of shootingyourself in the foot.
If you don't think that big, ifyou don't push beyond what is
comfortable as an entrepreneur,your role is to be larger than
life to have the idea that isbigger than today's
capabilities.
And so if you are always playingthe safe and short game, yes,
(27:00):
you may build a business, butare you swinging at the scale
that your customers, that yourvision needs?
As an entrepreneur, you need toswing at the venture scale.
I don't think people say thatenough!
Nils Pihl (27:13):
I had a call yesterday with an investor,
where we were talking about oneof the things that we are about
to release later this yeartogether with our partners at
Mentra that make this nextgeneration pair of smart
glasses.
And the investor asked like,look— these are two startups
working together.
You are an early stage startup.
They're an early stage startup.
Isn't this incredibly risky?
Aren't there so many things thatcould go wrong here?
(27:34):
To which I said yeah, it'sincredibly risky.
That's why there's a venturescale opportunity here.
There wouldn't be if it wasn'tincredibly risky.
To which he laughed and said,"Love that answer!" An older
version of me would've tried todefend like, no, no, it's not
risky.
You know, they're verycompetent.
But now I'm old enough to saylike, yeah, this is very risky.
(27:56):
So who knows?
This could be huge.
Nathan C Bowser (27:58):
Partnerships feel like the way that you can
make step growth or you can addfeatures or you can go to market
faster, better, stronger.
But like you said, thosepartnerships aren't always
coequal.
They're not always focused onrevenue or value.
They're focused on like shippinga next tech thing, but not
(28:19):
necessarily shipping a product.
How do you keep this partnershipfocused on value, on revenue on
the future of your youngstartups while you're also
figuring things out?
Nils Pihl (28:33):
With difficulty and with a couple of long hikes to
align incentives, betweenfounders.
Take a couple of long walks innature, talk about the future
you wanna see, check each otherout.
Do you see yourself being ableto pull through the tough times?
I have tremendous faith inCayden and Alex, the founders of
(28:54):
Mentra, the smart glassescompany that co-exhibited with
us at AWE, because I've seenthem work really, really hard
when they didn't have anyresources and they were just
fueled by passion.
And throwing some gasoline onthat fire is bound to do
something very interesting.
You know, it's gonna explode inone way or another.
(29:15):
So very excited to work withthem.
We've had many partnerships thatdidn't pan out, obviously, where
either no work ended up beingdone some conflict of interest
arose that made it impossible toproceed.
You end up arming thecompetition or whatever.
But I have maybe a naive faithin the size of the spatial
(29:42):
computing opportunity that youcan make a very big pie that you
can share with a lot of people.
There has probably never been abigger economic opportunity in
humanity's history than spatialcomputing.
Like I'm completely red-pilledabout this and have been
red-pilled about this sincebefore ChatGPT proved that AI
(30:05):
was right around the cornerbecause, 70% or so of the world
economy is still tied tophysical labor and physical
locations.
And if we look at that throughan AI lens, it means that the AI
revolution is still stuck infirst gear because AI cannot
perceive the physical world.
Almost definitionally, whoeverbuilds an OpenAI for the
(30:27):
physical world has a much biggeraddressable market than OpenAI
has.
And even skipping AI, I believethat augmented reality is an
inevitable part of humanity'sfuture.
I think it's the future oflanguage itself.
It's as inevitable as likeinternet pornography.
There is no timeline where weinvent this kind of technology
(30:50):
and don't end up adopting it,right?
It's just bound to happen.
Augmented reality is what humansdo, right?
I think looking at the arc ofhistory from the Stone Age to
now, it's been a constant pushto towards the AR glasses.
Nathan C Bowser (31:07):
From the proscenium stage somewhere in
Greece, to the expo hall at AWE.
Overlays and AR glasses isinevitable.
I love that! I'm curious, what'sthe glow up or notable
transformation or just audaciousgoal that you're working on?
Nils Pihl (31:26):
Hybrid robotics.
Let me explain.
A lot of people are excitedabout humanoid robots now.
And we believe, have believedfor some time that the future
software stack for humanoidrobots needs to have six
discrete layers that know how totalk to each other.
Locomotion layer, which is therobot's ability to walk and not
(31:50):
fall over.
This is most of what we've seendemoed today.
Like, look the robot is walking.
But we also need a manipulationlayer.
The ability to grab things andpour things.
Let's say that we wanted to havea robot that could empty all the
trash cans at AWE, right?
What would it take to buildsomething like that?
It needs to be able to walk, itneeds to be able to manipulate
its environment, but it alsoneeds a spatial semantic
(32:13):
perception layer.
It needs to be able to perceivethe difference between a trash
can and a toddler, right?
So that we don't throw thetoddlers out, but we throw the
trash out.
And it needs to also have amapping layer so it doesn't have
to bounce around like a Roombalooking for trash cans.
(32:34):
It has an understanding of wherethe trash cans were last time
and where it can walk and whereit can't walk, and it's gonna
need a positioning layer tounderstand where it is in
relation to the map.
And then finally an applicationlayer that ties all of these
things together and gives it theinstructions.
Now the cool thing is that youcan just skip the locomotion and
(32:57):
the manipulation, if you keep ahuman in the loop.
Put a pair of glasses on a humanbeing, and you have human
locomotion and humanmanipulation.
We call this hybrid robotics,right?
So together with the guys atMentra already last year we
built a really cool prototype ofan AI that sees the world
(33:18):
through your eyes.
And helps you figure out what todo in the retail environment.
It detects things like emptyshelves and automatically
reports the empty shelf, in 3Dspace with AR guidance.
So that even if you are busycarrying a box, I can get a
notification to go fix it andknow immediately where it is.
(33:40):
You have AI perception, AIcognition, combined with human
problem solving and humanlocomotion and manipulation.
And we think this is going to beincredibly big.
We think that this is what'sgoing to be the main use case
for the glasses over the nextthree years or so, and that
(34:00):
realistically we can movemillions, maybe even hundreds of
millions of these glasses toprovide AI co-pilots for people
in the physical world.
Like we've seen that an AIcopilot is incredibly helpful
for people on the upper end ofthe IQ distribution, doing
things like programming.
There's no reason it wouldn't beeven more transformative for
(34:23):
those of us that are on thelower end of the IQ
distribution.
Like people forget that about aquarter of the male population
has an IQ below 95.
And below 95 you startstruggling.
You know, that's a lot of us,right?
And at that IQ level, you startstruggling with multi-part
instructions and you startstruggling with imagining the
(34:46):
steps that you will need to taketo solve a multi-part problem,
and having an AI assistant thatcan help you think through those
things will help you be moreproductive, be more motivated,
not feel so alienated at work.
You know, retail for example,has this enormous staff turnover
issue.
If you take a job at Walmarttoday, there's a 70% chance that
(35:09):
you quit before the end of theyear, right?
A 70% chance.
And a lot of this is becauseit's an alienating place to
work.
Retail is a tough industry,right?
And giving people better toolsto succeed I think is going to
be super transformative.
So we're all about hybridrobotics.
(35:29):
Putting glasses on people,putting phones in their hands
and letting them have an AIcopilot that helps them do
better and be happier.
Nathan C Bowser (35:40):
Nils is there, on this journey, whether it's to
combine robotics, whether it'son open sourcing, a visual
understanding of the world, isthere anything that Auki is
looking for?
Connections, networking,research, support?
Nils Pihl (35:57):
So a lot of the work we do is fully open source, and
you can find it on our GitHuband we pay for contributions.
We have a grants program forpeople that wanna undertake to
build things on top of theprotocol.
If you have a great applicationsidea.
But we also have bounties, andjust ad hoc payments for people
(36:17):
that make contributions to theprotocol itself.
So you should check out theposemesh repo on our GitHub,
that's P-O-S-E-M-E-S-H, theposemesh, which is the name of
the underlying protocol of theAuki network.
And see if there's anything youwanna contribute.
And if you have an idea of whatto build on top of this
(36:38):
protocol, we're very happy towrite you a grant.
The goal here is civilizationscale, infrastructure funded,
owned and operated by thepeople.
We, the people.
Let's do it together!
Nathan C Bowser (36:49):
Oh my gosh.
That's gorgeous! Nils, each GlowUp episode likes to make space
for a community shout-out to anonprofit or other organization
doing good out in the world.
Is there somebody you'd like togive a spotlight to?
Nils Pihl (37:02):
I want to express my sincere gratitude to our Discord
community, discord.gg/aukiverse,for helping us stay motivated
when things have been tough,because startups are always
tough.
Even showing up, in physicalspace to help out at AWE.
(37:25):
We had Max from the communityfly out, to help out and set up
our booth and talk to customers.
In Singapore we had Roddy do thesame thing, even picking me up
at the airport.
Amazing community! I am deeplythankful to everyone on that
Discord.
So love to you guys!
Nathan C Bowser (37:45):
Amazing! Nils, I really appreciate, the
conversation that we've hadtoday and how there is this
careful blend in your founder'sapproach.
Open source collaboration.
Unwavering confidence in yourvision and your role as an
entrepreneur.
And a very open and pragmaticview on the pace and the process
(38:12):
of innovation.
I'm excited to see, compared tosome of the others working in
this geospatial intelligencespace you're kind of small.
And I love to see big visionsthrown down by ambitious
companies.
So thank you for embodying a lotof the ethos that we like to
(38:33):
talk about here on The Glow Up.
It's been such a pleasure todive in to this journey sparked
by Warhammer and ushering inmaybe an inevitable AR era of
computing.
So cool to talk with you today.
Thank you!
Nils Pihl (38:47):
If I can just go out on a Terence McKenna quote,"In
the cyberdelic future, artistswill rule because the world will
be made of art." This is whatspatial computing is all about.
We will be able to manifest ourknowledge and imagination in the
minds of others.
In ways that our language hasnot allowed before.
(39:09):
The AR journey is a journey toimprove the very stack of
language itself.
Nathan C Bowser (39:15):
That's fantastic!
Hello and welcome to The Glow Up.
I'm Nathan C and today I amtalking with Nils Pihl, CEO of
Auki Labs.
Nils, it's great to talk withyou today.
Thanks for joining!
Nils Pihl (00:11):
Happy to see you here online! We met again in real
life not long ago at AWE.
Nathan C Bowser (00:17):
It was so good seeing you at AWE and I can't
wait to dive in and hear alittle bit more about the
inspirations and work that wentinto that very dynamic and
interactive booth that you hadthere.
Before we get into maybe thecoolest spatial computing event
of the year, let's wind back alittle bit and could you
(00:38):
introduce yourself and the workthat you do in innovation at
Auki Labs?
Nils Pihl (00:43):
Sure! So I am Nils.
Auki is a company focused oncollaborative machine
perception.
The idea is that digital deviceshave to have a shared
understanding of the physicalworld.
It started out very innocentlyin 2019 with us trying to make
(01:04):
augmented reality overlays forthe tabletop game Warhammer.
That's where we got started.
And, what we ran into as aproblem was that shared
augmented reality was verydifficult.
Like we want to be able to see ashared overlay over the table.
And I didn't know the firstthing about augmented reality
(01:26):
back in 2019 when I started thecompany.
So I naively imagined, like Ithink many people imagine, that
AR was in a much better statethan it actually is.
And I had the cold shower ofrealizing that like, oh the GPS
is actually not fit for purposeat all.
You can't use the GPS forWarhammer.
We need millimeters ofprecision.
(01:48):
And the GPS doesn't even knowI'm in this house right now.
But it was COVID, so rather thangiving up, I had my Warhammer
and I could just sit and lean onmy autism for two years and
think about how do you solvethis problem?
And over those two years I cameto realize that a very big
(02:08):
problem for the future that Ihad always imagined, you know, a
future with AR glasses androbots, is that digital devices
just don't understand thephysical world very well, and
they don't have a sharedunderstanding.
By 2021, I had inventedsomething that made it possible
to do what I wanted to do withthe Warhammer thing, and that
(02:30):
helped me raise a very generousamount of seed funding to build
an open source collaborativeprotocol for all kinds of
machines to see the worldtogether.
And this year, we finallylaunched an SDK at AWE, that is
(02:50):
focused on collaborative machineperception.
And we launched our firstproduct built on that SDK
earlier this year.
A spatial AI tool for retailersand we just recently broke 2
million ARR on that product andhope to hit 10-15 million ARR by
(03:11):
the end of this year on thatproduct.
Nathan C Bowser (03:15):
I think we just discovered the title of this
episode, which is somethingalong the lines of"From
Warhammer to 15 Million in ARRwith Augmented Reality and
Robots." Actually we gotta putthe time-based in there so
"Within Five Years" is a prettyamazing journey to capture!
Nils Pihl (03:38):
Many people predicted that if I spend too much time
playing Warhammer, I will end upworking in retail.
And look I did!
Nathan C Bowser (03:45):
This is very much like the entrepreneur's
curse, you fulfill thosepredictions of other people.
You just do it in a way thatthey've never imagined, right?
Like, so I'll be working totransform the back of house
supply chain operations forretail, not just helping people
find where the lettuce is.
(04:07):
That was really fantastic Nils!I love how this journey kind of
starts innocently, includes alot of special interests.
A lot of like inspiration.
You mentioned the seed round andthis goal to produce open source
technology to connect people androbots and AI to an
(04:28):
understanding of the world.
Building a business around opensource feels like it has extra
complication.
I'm curious and might not alwaysbe the thing that like venture
investors are looking for.
Can you talk a little bit about,why open source for something,
(04:49):
so cutting edge and on thefuture of where it seems like
technology is going?
Nils Pihl (04:54):
Yeah, but now we're really going on a trip here! So
I believe that the internetitself is missing a protocol.
The internet itself is missing away for machines to shake hands
with each other and say, here'swhat I know about the physical
world.
And I really want for oursolution to become a part of the
(05:21):
internet itself and not become abig corporation.
The reason why was I had asomewhat mushroom induced moment
of clarity when working on theAugmented reality solution and
thinking about the AR glassescoming, right?
So.
(05:41):
You know, go back to 2020 or soand I'm learning about how will
we do positioning in a post GPSworld?
And by 2020 there were only twocandidates that I could see it
could possibly be.
Either we're gonna do itentirely blind using ultra
wideband trilateration.
(06:03):
You know, the AirTag was prettynew back then and ultra wideband
seemed like a pretty promisingway forward where, you know,
maybe we could even build a pairof AR glasses that didn't even
have a camera! Maybe we could.
Nathan C Bowser (06:16):
Kind of like the earlier teens idea of IoT
driven by beacons, right?
Nils Pihl (06:21):
Absolutely!
Nathan C Bowser (06:22):
But that was back on Bluetooth.
Nils Pihl (06:23):
Yeah ultra wideband was promising to maybe save the
idea of the Bluetooth beacon.
Do something that could be ahigher precision.
The other alternative, and Ithink what the industry had
already settled on, you know,people didn't believe ultra
wideband and now I don't think Ido either, was visual
(06:45):
positioning.
The idea that the device tellsyou what it's looking at and you
tell the device where it is.
What visual positioning is, isthat you compare an image feed
of what the device is seeing toa model of the world that you've
previously made so that you cantry to match where was this
frame taken and saying the quietpart out loud means that you're
(07:08):
quite literally telling someonewhat it is you are looking at
for this service to be able towork.
And that's not the scariestthing in the world when it's our
phone and we only whip it outevery now and then.
But when we're wearing ARglasses all day, the thought
that Mark Zuckerberg or JohnHanke or any one of these guys
(07:29):
would be able to know what mykids are looking at at all
times.
I realized that the positioninginfrastructure needs to be
privacy preserving, otherwisethe augmented reality future
that I was so excited about andI was listening to a lot of
Terence McKenna at the time andI was imagining the cyberdelic
(07:50):
future.
I could see how it could veryeasily turn very dystopian if we
build this under the traditionalSilicon Valley surveillance
capitalist business model.
And we would have to find someother kind of business model.
Luckily for me, I had a coupleof early backers that were very
(08:13):
big into Web3 and DePIN whobasically bullied me in 2021
'cause I am, I think to thisday, very largely anti-crypto.
But they bullied me into lookingseriously at what it is that
(08:34):
Web3 could enable you to do,both from a technological point
of view, but also from abusiness model point of view.
And that you could build abusiness model that can survive
the corporation and become apart of the internet itself.
And by late 2021, I had embracedthat fully.
(08:55):
That like, yes there is a lightat the end of the tunnel.
In all of this Web3 noise, thereis a promise of a new kind of
business model, civilizationscale infrastructure funded,
owned, and operated by thepeople.
And that is radically exciting!Those investors have no qualms
(09:20):
with open source at all.
In fact, they think open sourceis fantastic, right?
The decentralization communitydon't want power to be
centralized with anyone and arevery happy for the technology to
be open source.
So I, accepted their support.
They gave us almost$20 millionin seed funding to build this
(09:42):
thing.
And that also was veryconvincing because normally
traditional VC would not give 20million to an idea that was so
early.
And it gave us an opportunity totry to do things that normally a
startup would not be able to do.
(10:03):
Part of the selling point, likeoh my god! Web3 has the
capability to crowdfundinfrastructure projects that no
VC would.
Well, not no VC.
There are certainly some VCsthat write these checks every
now and then, but it's not easyto get 20 million in seed
(10:23):
funding from traditional VCs.
But being able to have this newbusiness model and have, again,
civilization scale projects thatare funded, owned, and operated
by the people deeply red-pilledme into a potential future
software upgrade of capitalismitself.
(10:46):
This is an exit for surveillancecapitalism, we can build another
kind of company.
And now I'm all about that.
Nathan C Bowser (10:56):
Hell yeah! I love that a very sort of, almost
cute idea about augmentedreality led to such an expansive
understanding of theinfrastructure that goes into 3D
like coming economies powered byrobots, and like what that's
(11:18):
gonna mean.
This very compelling idea thatlike traditional models for
funding in startups
Nils Pihl (11:27):
are
Nathan C Bowser (11:28):
extremely risk adverse and have maybe not been
the greatest predictors of likewhere things are going.
And that crypto, you know, Web3is like one of the areas where
there's a lot of capital andinvestment that's going on
pushing some of these new modelsand, I think really the most
(11:51):
exemplary thing that I want tocall out is when you're thinking
about the problem space, whenyou're thinking about the value,
when you're thinking aboutdelivering something that will
unlock even just like the nextthings you're trying to do in
your work that by reallyupleveling where you take your
(12:12):
solution, right?
It's not just a Web3 platformthat you can use in Auki.
It's literally giving the entireecosystem a missing piece and
allowing that to likepotentially change the way that
we interact with technologies.
Beautiful artwork by KeiichiMatsuda in that hyper reality,
(12:35):
corporate augmented dystopia.
But like man, I don't personallywant a single corporate oligarch
owning all my spatial data!
Nils Pihl (12:46):
Last year Elon was on Lex Fridman's podcast and he was
asked,"What will it take to winthe AI race?" And Elon said it
will take two things, right?
You need to have the bestcompute.
That's why he's building thisdata center in Houston or
wherever, and you need the bestdata.
And what he said was,"Oh we'vealready, you know, consumed all
(13:07):
the data on the internet."Turned out there wasn't that
much data on the internet tobegin with.
But reality scales, he said.
And then he said something thatshould have scared us all, which
was Optimus, his humanoid robotis going to be the greatest
source of data in the world.
Without any shame he just goeson Lex Fridman's podcast to say
(13:28):
like, hey, I'm building a robotto spy on you in your homes.
And I actually think that theproblem with the traditional VC
model is not that they are riskaverse because I mean, clearly
they do take risk and they losea lot of money making bad bets.
I think it's a more perniciousproblem of group think around
business models.
(13:48):
To this day, we get a lot ofpressure from VCs that like,"Why
are you letting your customersown this data?
You could own this data.
Imagine if you owned all thisdata." It's like, no exactly!
Imagine if I owned all thisdata.
That would be terrible! Weshould let our customers own
their own data.
And it is possible to buildother business models.
(14:10):
You know,"Data is the new oil"is so deeply ingrained in
Silicon Valley, that they oftendon't take anything seriously
that is not an attempt tocollect data.
Like a good friend of minerecently raised a round for
building smart glasses and I'man advisor to his company and I
(14:31):
got to listen in to a couple ofthe VC pitches and there's a lot
of pressure to just be like,"Areyou going to be collecting the
data from the glasses?
That's going to be so valuable."The call is coming from inside
the house.
The VCs are behind a lot of theexcesses of surveillance
capitalism, and I think they'vejust gotten caught in a group
think loop that this is the onlybusiness model that could work.
Nathan C Bowser (14:55):
One of the things that I talk a lot about
at Awesome Future is this ideathat innovation usually fails in
the last mile connection fromthat good idea to where the
customers actually are.
And I think your conversationaround like owning people's data
is one of those really clearinstances where like in a
(15:18):
boardroom level, at a VC level,obviously owning the data is a
resource and you wanna have asmany resources available to your
org as possible.
On the flip side, right you'realso calling out that spatial
experiences, AR, VR, even likeany experiences, any digital
(15:39):
experience at this momentdoesn't have enough metadata
attached to it to actuallyinteract with the world or to
really know like where it is andwhat it's supposed to be doing
beyond little actual boxes thatwe tape on the floor for robots.
And so spatial experiences,digital experiences are lacking
(16:00):
key understanding and metadatajust period.
And you're trying to own it all,even though it's like it
incomplete.
There's just gaps in reallyunderstanding like what happens
in that last little bit and, wewant all the data, but at this
moment it is incomplete anddoesn't know what it's supposed
to be doing.
Is that actually as valuable asyou think it is?
(16:21):
And, I love this position thatAuki's taken which is, it's
actually way more valuable if weincorporate this in this kind of
spatial data, this awareness ofwhere you are, who you are.
Oh my goodness!
Nils Pihl (16:35):
A lesson from history that I think VCs should
remember, right, is in the goldrush, you should be selling
picks and shovels.
It's not about the gold itself,right?
If data is the new oil, thatdoesn't mean that the game is to
be collecting all the data.
Make picks and shovels that makeit possible for your customers
to collect data that's valuablefor them.
Nathan C Bowser (16:54):
The internet is full of basically 2D data,
right?
It's text, it's images, it'spixels, it's spreadsheets.
You know, when you actuallythink about adding a third
dimension to your data, right?
Like you start getting theseexponential growths in what's
going on.
When I started in XR about adozen years ago the conversation
(17:19):
was like, we're gonna usemachine learning to have a
semantic understanding of thephysical world.
Are we just renaming what'shappening now?
At AWE, I would say one of thebiggest themes that I saw there
was that every computingplatform, whether it's Niantic,
whether it's Qualcomm, whetherit's Auki Labs was talking about
(17:44):
this idea of"geospatialintelligence" or a map of the
world that humans and robots canboth understand.
Has this idea, you know,actually changed from the terms
we were using a dozen years ago?
Or is there something likereally unique about what's going
(18:07):
on in this moment?
Nils Pihl (18:08):
Oh, it's a little bit of column A and a little bit of
column B.
I think that, the spatialcomputing industry has run out
of money.
Which is a big problem.
I think that one of the thingsthat I noticed at AWE this year
and last year is that, there'sbeen a while since there were
(18:30):
big cash infusions into the XRecosystem.
A lot of companies are lookingfor the exit.
A lot of companies are on theirlast bit of gas and as such,
they are in fundraising mode andare going to speak the language
(18:51):
that the VCs want to hear.
And this year the VCs want tohear about robots.
They didn't want to hear aboutrobots in 2021.
Like I told people about robotsin 2021 and they were like,
"Please just talk about themetaverse." But like in my
original pitch in 2021, Ipointed out that look— an iPhone
is just a robot without arms andlegs.
(19:11):
We can get started building thespatial awareness right on the
iPhone.
We don't have to wait for theglasses.
We don't have to wait for therobots.
We can build this for the iPhonenow.
And when the robots are readyand they're like,"Robots?!
Metaverse!" And now it's like,"Don't talk about the metaverse!
Tell us about robots." Butthere's also I think, a couple
of interesting new developmentsin tech that are making these
(19:36):
things a lot more realistic.
I think people have had thevision of this spatial semantic
understanding for a long time,but there's been so much
homework to do, and in just thelast two years, there's been an
explosion of things that make itpossible, and a lot of it is
open source.
But I also wanna give creditwhere credit is due.
(19:56):
They've put out a lot of amazingopen source research.
And China has been putting outjust enormous amounts of open
source research when it comes tomachine perception.
Now, things that were completesci-fi, like people literally
telling me in 2021, it's inprinciple impossible.
Things like monocular depthestimation.
(20:18):
I had people tell me were inprinciple impossible has gotten
surprisingly good.
Tencent and ByteDance bothreleased open source, monocular
depth estimators that are crazyand can run in almost 60 frames
per second.
It's remarkable! A little bit ofcolumn A, a little bit of column
B.
I think the language has changedbecause there's a little bit of
(20:41):
an era of desperation in the XRcommunity because, we had our
hype cycle, we got a lot offunding during the hype cycle,
and there hasn't been follow upinvestments since.
There are very few activeinvestors in the XR space, and
they are all in the pre-seed andseed stage.
Who writes a Series A check foran XR company, right?
(21:05):
So that's a problem.
Nathan C Bowser (21:06):
They have to have"XR AI" in their name.
Nils Pihl (21:10):
Yes! "Something-something Robots."
Nathan C Bowser (21:12):
Maybe that's the title of the show! The
market changing its interest,investors like shifting where
they're funding is probably oneof the most dramatic inflection
points, learning moments,decision points that a founder
is going to have to encounter.
(21:33):
So, through your experience inspatial computing and the amount
of like shifts of attention,even in the last five years.
I'm curious like how you'veapproached, maintaining your
core vision while adjusting tothe market, or if you have
advice for folks who are on thisjourney and experiencing so much
(21:58):
change and disruption in thecurrent models.
Nils Pihl (22:01):
There are two sides to this answer.
I have to be honest about havinghad to make a lot of pivots on
the way to where we are today.
It wasn't just from Warhammer toretail, it was from Warhammer to
AR pets, to retail.
One thing that has been constantsince 2021 at least, is this
(22:22):
vision of a collaborativemachine perception network.
How to get there has changedsomewhat over time.
Frustratingly it has changedback a lot to what it originally
was in 2021, which is also apainful lesson of who you should
(22:43):
listen to and not listen to.
When you take on investor money,one thing that you open yourself
up to is that they are gonnastart giving you advice.
And one piece of advice thatI've been given is to not say
what I'm about to say, but I'mgonna say it anyhow.
A lot of VCs have never builtanything themselves.
They have backgrounds inbanking.
(23:05):
They've never shipped a product.
They've never had to fire 20people.
They don't know anything reallyabout what it means to be an
entrepreneur, and they don'thave to be fair, you know, their
job is not to be anentrepreneur.
But they will look at the worlddifferently'cause their job is
(23:27):
not to make your ideas like cometrue.
Their job is to get a positivereturn, and it doesn't matter to
them if you end up pivoting orwhatever, as long as they get
their money.
So we've constantly been gettingpushes to pursue the latest hot
(23:48):
buzzword from our investors.
In early 2021, one of the firstthings that we thought, you
know, what are you gonna do withcollaborative machine perception
was like, well we could starthelping people navigate in
warehouses and retailenvironments and we could make
it easier to deploy robots forthese things.
(24:11):
And investors came and saidlike, no, no, no.
You need to make a B2C product.
And we had a first demo, just atech demo, which was an AR pet
that you could see in sharedaugmented reality.
Like it really just, it was justan art asset, Nathan.
Like there was no game there.
It was just like, here's ananimated animal and we can see
it together.
And they're like,"You shouldtotally build that! You need to
(24:32):
build it, build it!" And we did.
And it turned out, you know,that an AR pet was actually
worse than a 2D pet because wedidn't have collaborative
machine perception yet, right?
Like we had cracked, having ashared coordinate system so we
could see the pet together.
And that was great, but the petcouldn't see its environment.
(24:55):
So we would walk through chairsand things like this that would
break the immersion.
So you were literally better offnot using the AR.
In hindsight, you know, if wehad just stayed the course and
done what we wanted to do, wecould have launched this retail
thing two years earlier maybe.
So, you know, it's hard becausebeing an entrepreneur is also
(25:15):
very scary.
But now I've learned that look,your job is to be the
entrepreneur.
And their job is to find theright entrepreneur, not to shape
the entrepreneur.
And they should be okay withpotentially losing money.
You are supposed to take aventure scale swing, right?
That's your obligation as anentrepreneur.
You need to take a venture scaleswing, and you need to
(25:37):
understand that it's okay tofail.
As long as you're trying to dosomething that is venture scale.
And that's a little bitcounterintuitive because you
know, you feel like, oh maybe ifI take this less risky path, I
could have a profitablebusiness.
Yeah but a profitable businessis not what they are looking
for.
They're looking for a venturescale return.
Nathan C Bowser (25:55):
Oh my gosh! There is such a terrible tension
in innovation, in technology.
This is what we know, this iswhat feels safe, and this is the
size of the swing we need totake, are so often on such
(26:17):
different ends of the spectrum.
And this idea that taking theeasy route or taking the advice
of other people above you as anentrepreneurs, that vision, that
unique perspective on the futurethat only you see and what your
intuition, your team can build.
(26:39):
That you're kind of shootingyourself in the foot.
If you don't think that big, ifyou don't push beyond what is
comfortable as an entrepreneur,your role is to be larger than
life to have the idea that isbigger than today's
capabilities.
And so if you are always playingthe safe and short game, yes,
(27:00):
you may build a business, butare you swinging at the scale
that your customers, that yourvision needs?
As an entrepreneur, you need toswing at the venture scale.
I don't think people say thatenough!
Nils Pihl (27:13):
I had a call yesterday with an investor,
where we were talking about oneof the things that we are about
to release later this yeartogether with our partners at
Mentra that make this nextgeneration pair of smart
glasses.
And the investor asked like,look— these are two startups
working together.
You are an early stage startup.
They're an early stage startup.
Isn't this incredibly risky?
Aren't there so many things thatcould go wrong here?
(27:34):
To which I said yeah, it'sincredibly risky.
That's why there's a venturescale opportunity here.
There wouldn't be if it wasn'tincredibly risky.
To which he laughed and said,"Love that answer!" An older
version of me would've tried todefend like, no, no, it's not
risky.
You know, they're verycompetent.
But now I'm old enough to saylike, yeah, this is very risky.
(27:56):
So who knows?
This could be huge.
Nathan C Bowser (27:58):
Partnerships feel like the way that you can
make step growth or you can addfeatures or you can go to market
faster, better, stronger.
But like you said, thosepartnerships aren't always
coequal.
They're not always focused onrevenue or value.
They're focused on like shippinga next tech thing, but not
(28:19):
necessarily shipping a product.
How do you keep this partnershipfocused on value, on revenue on
the future of your youngstartups while you're also
figuring things out?
Nils Pihl (28:33):
With difficulty and with a couple of long hikes to
align incentives, betweenfounders.
Take a couple of long walks innature, talk about the future
you wanna see, check each otherout.
Do you see yourself being ableto pull through the tough times?
I have tremendous faith inCayden and Alex, the founders of
(28:54):
Mentra, the smart glassescompany that co-exhibited with
us at AWE, because I've seenthem work really, really hard
when they didn't have anyresources and they were just
fueled by passion.
And throwing some gasoline onthat fire is bound to do
something very interesting.
You know, it's gonna explode inone way or another.
(29:15):
So very excited to work withthem.
We've had many partnerships thatdidn't pan out, obviously, where
either no work ended up beingdone some conflict of interest
arose that made it impossible toproceed.
You end up arming thecompetition or whatever.
But I have maybe a naive faithin the size of the spatial
(29:42):
computing opportunity that youcan make a very big pie that you
can share with a lot of people.
There has probably never been abigger economic opportunity in
humanity's history than spatialcomputing.
Like I'm completely red-pilledabout this and have been
red-pilled about this sincebefore ChatGPT proved that AI
(30:05):
was right around the cornerbecause, 70% or so of the world
economy is still tied tophysical labor and physical
locations.
And if we look at that throughan AI lens, it means that the AI
revolution is still stuck infirst gear because AI cannot
perceive the physical world.
Almost definitionally, whoeverbuilds an OpenAI for the
(30:27):
physical world has a much biggeraddressable market than OpenAI
has.
And even skipping AI, I believethat augmented reality is an
inevitable part of humanity'sfuture.
I think it's the future oflanguage itself.
It's as inevitable as likeinternet pornography.
There is no timeline where weinvent this kind of technology
(30:50):
and don't end up adopting it,right?
It's just bound to happen.
Augmented reality is what humansdo, right?
I think looking at the arc ofhistory from the Stone Age to
now, it's been a constant pushto towards the AR glasses.
Nathan C Bowser (31:07):
From the proscenium stage somewhere in
Greece, to the expo hall at AWE.
Overlays and AR glasses isinevitable.
I love that! I'm curious, what'sthe glow up or notable
transformation or just audaciousgoal that you're working on?
Nils Pihl (31:26):
Hybrid robotics.
Let me explain.
A lot of people are excitedabout humanoid robots now.
And we believe, have believedfor some time that the future
software stack for humanoidrobots needs to have six
discrete layers that know how totalk to each other.
Locomotion layer, which is therobot's ability to walk and not
(31:50):
fall over.
This is most of what we've seendemoed today.
Like, look the robot is walking.
But we also need a manipulationlayer.
The ability to grab things andpour things.
Let's say that we wanted to havea robot that could empty all the
trash cans at AWE, right?
What would it take to buildsomething like that?
It needs to be able to walk, itneeds to be able to manipulate
its environment, but it alsoneeds a spatial semantic
(32:13):
perception layer.
It needs to be able to perceivethe difference between a trash
can and a toddler, right?
So that we don't throw thetoddlers out, but we throw the
trash out.
And it needs to also have amapping layer so it doesn't have
to bounce around like a Roombalooking for trash cans.
(32:34):
It has an understanding of wherethe trash cans were last time
and where it can walk and whereit can't walk, and it's gonna
need a positioning layer tounderstand where it is in
relation to the map.
And then finally an applicationlayer that ties all of these
things together and gives it theinstructions.
Now the cool thing is that youcan just skip the locomotion and
(32:57):
the manipulation, if you keep ahuman in the loop.
Put a pair of glasses on a humanbeing, and you have human
locomotion and humanmanipulation.
We call this hybrid robotics,right?
So together with the guys atMentra already last year we
built a really cool prototype ofan AI that sees the world
(33:18):
through your eyes.
And helps you figure out what todo in the retail environment.
It detects things like emptyshelves and automatically
reports the empty shelf, in 3Dspace with AR guidance.
So that even if you are busycarrying a box, I can get a
notification to go fix it andknow immediately where it is.
(33:40):
You have AI perception, AIcognition, combined with human
problem solving and humanlocomotion and manipulation.
And we think this is going to beincredibly big.
We think that this is what'sgoing to be the main use case
for the glasses over the nextthree years or so, and that
(34:00):
realistically we can movemillions, maybe even hundreds of
millions of these glasses toprovide AI co-pilots for people
in the physical world.
Like we've seen that an AIcopilot is incredibly helpful
for people on the upper end ofthe IQ distribution, doing
things like programming.
There's no reason it wouldn't beeven more transformative for
(34:23):
those of us that are on thelower end of the IQ
distribution.
Like people forget that about aquarter of the male population
has an IQ below 95.
And below 95 you startstruggling.
You know, that's a lot of us,right?
And at that IQ level, you startstruggling with multi-part
instructions and you startstruggling with imagining the
(34:46):
steps that you will need to taketo solve a multi-part problem,
and having an AI assistant thatcan help you think through those
things will help you be moreproductive, be more motivated,
not feel so alienated at work.
You know, retail for example,has this enormous staff turnover
issue.
If you take a job at Walmarttoday, there's a 70% chance that
(35:09):
you quit before the end of theyear, right?
A 70% chance.
And a lot of this is becauseit's an alienating place to
work.
Retail is a tough industry,right?
And giving people better toolsto succeed I think is going to
be super transformative.
So we're all about hybridrobotics.
(35:29):
Putting glasses on people,putting phones in their hands
and letting them have an AIcopilot that helps them do
better and be happier.
Nathan C Bowser (35:40):
Nils is there, on this journey, whether it's to
combine robotics, whether it'son open sourcing, a visual
understanding of the world, isthere anything that Auki is
looking for?
Connections, networking,research, support?
Nils Pihl (35:57):
So a lot of the work we do is fully open source, and
you can find it on our GitHuband we pay for contributions.
We have a grants program forpeople that wanna undertake to
build things on top of theprotocol.
If you have a great applicationsidea.
But we also have bounties, andjust ad hoc payments for people
(36:17):
that make contributions to theprotocol itself.
So you should check out theposemesh repo on our GitHub,
that's P-O-S-E-M-E-S-H, theposemesh, which is the name of
the underlying protocol of theAuki network.
And see if there's anything youwanna contribute.
And if you have an idea of whatto build on top of this
(36:38):
protocol, we're very happy towrite you a grant.
The goal here is civilizationscale, infrastructure funded,
owned and operated by thepeople.
We, the people.
Let's do it together!
Nathan C Bowser (36:49):
Oh my gosh.
That's gorgeous! Nils, each GlowUp episode likes to make space
for a community shout-out to anonprofit or other organization
doing good out in the world.
Is there somebody you'd like togive a spotlight to?
Nils Pihl (37:02):
I want to express my sincere gratitude to our Discord
community, discord.gg/aukiverse,for helping us stay motivated
when things have been tough,because startups are always
tough.
Even showing up, in physicalspace to help out at AWE.
(37:25):
We had Max from the communityfly out, to help out and set up
our booth and talk to customers.
In Singapore we had Roddy do thesame thing, even picking me up
at the airport.
Amazing community! I am deeplythankful to everyone on that
Discord.
So love to you guys!
Nathan C Bowser (37:45):
Amazing! Nils, I really appreciate, the
conversation that we've hadtoday and how there is this
careful blend in your founder'sapproach.
Open source collaboration.
Unwavering confidence in yourvision and your role as an
entrepreneur.
And a very open and pragmaticview on the pace and the process
(38:12):
of innovation.
I'm excited to see, compared tosome of the others working in
this geospatial intelligencespace you're kind of small.
And I love to see big visionsthrown down by ambitious
companies.
So thank you for embodying a lotof the ethos that we like to
(38:33):
talk about here on The Glow Up.
It's been such a pleasure todive in to this journey sparked
by Warhammer and ushering inmaybe an inevitable AR era of
computing.
So cool to talk with you today.
Thank you!
Nils Pihl (38:47):
If I can just go out on a Terence McKenna quote,"In
the cyberdelic future, artistswill rule because the world will
be made of art." This is whatspatial computing is all about.
We will be able to manifest ourknowledge and imagination in the
minds of others.
In ways that our language hasnot allowed before.
(39:09):
The AR journey is a journey toimprove the very stack of
language itself.
Nathan C Bowser (39:15):
That's fantastic!
Popular Podcasts
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
NFL Daily with Gregg Rosenthal
Gregg Rosenthal and a rotating crew of elite NFL Media co-hosts, including Patrick Claybon, Colleen Wolfe, Steve Wyche, Nick Shook and Jourdan Rodrigue of The Athletic get you caught up daily on all the NFL news and analysis you need to be smarter and funnier than your friends.