June 11, 2025 • 23 mins
EPISODE 12
From building gravity-defying camera rigs for Alfonso Cuarón’s Gravity to sending drones into live nuclear reactors and, most recently, designing job-site robots that can survive the dust and chaos of construction, Asa Hammond’s career shows how precision engineering and artistic vision can coexist. In this conversation, Asa explains how the exacting pixel-perfect discipline of Hollywood VFX became the blueprint for ultra-reliable industrial robotics, why safety and repeatability always come before flashy features, and how new UX layers—voice, LLMs, demonstration learning—are finally making complex machines feel like familiar tools. Along the way, we hear tales of full-scale “stunt-double” reactors, the reality of folding-laundry robots, and what it takes to assemble a world-class, T-shaped engineering team that thrives at the edge of possibility.
CHAPTERS
00:29 – From Film Sets to Founders: Show Intro
02:12 – Asa’s Hollywood-to-Hard-Hats Origin Story
05:57 – Engineering Robots for Gravity
09:05 – Reinventing Construction with Industrial Arms
15:56 – Flying Drones Inside Nuclear Reactors
LINKS
Connect with Asa Hammond
Website • LinkedIn • X/Twitter
Stay Connected with Founder Mode
Subscribe to our newsletter: foundermode.kit.com
Connect with Kevin
Connect with Jason
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
Yeah, as, as you can imagine, they're, they're very, very
careful people, those who run nuclear reactors.
So that was a very high spec project.
It took us a long time to do it primarily because of all the
they're, they're very careful kind of ramp up to get to the
point where you can, you can go in and fly during an active
emission. Welcome back to Founder Mode.
(00:27):
Where we talked to founders doing the wildest stuff and
getting their hands. Dirty Yeah, today's topics
robotics that go from the big screen to the biggest hyper
scaler builds. You might think film and
construction have nothing in common, but when it comes to
precision tools, robotics is actually that unexpected bridge.
Jason, you've worked in like tech and media like, what do you
think of of all the AI video models like Google's update the
(00:48):
view? So I watched that that keynote
from IO and I've been seeing allthese videos come out.
I've started to play with the tool.
This stuff is nuts. Like when we're going to talk to
Ace in a minute about film sets and robots that are doing the
most complicated moves with the camera and amazing stuff.
But like when you don't even need the $1,000,000 film set to
create something that looks likea blockbuster movie.
Like when this stuff all converges and you have these
(01:11):
films that are part AI created, part, you know, incredible
talented humans doing it. Like I think it's a 1 + 1 = 3
moment. I think it's going to completely
change the game and the creativity and the artistry that
can come from it. It's not about people's jobs
being destroyed in the entertainment industry.
It's about the job changing, which is what we talk about all
the time on the show. So like as that example, which
(01:31):
I'm talking about, like bringingthis creativity that is being
ushered in by AI into these new markets like film and
entertainment. And we're talking about, you
know, using robotics and construction with ASA.
How do you think AI, robotics and all this comes together
across industries, Kevin? Yeah, it's super interesting as
founders. Like part of the goal is to find
that new tech. And I mean, honestly, what we've
(01:51):
been doing for decades is like thinking about what's the new
and exciting things and how do you bring that back to either
new businesses or old businesses?
And I mean, for this conversation, it actually goes
back to the late 90s when ASA and I met, you know, as students
at UCLA in engineering. And so I think it's pretty
interesting. And I think there's going to be
a bunch to unpack. Here.
Yeah, exactly. And that's what makes ASA
Hammond, our guest today, so fascinating.
(02:14):
His career connects 2 things that you would not expect
Hollywood and hard hats. Yeah, Asa's worked on films like
Gravity Built Drones to inspect nuclear reactors.
I'd like the sickest climbing wall in his last factory, and
maybe if we're lucky, he'll callin from the new factory where
they're now working on robotics for construction.
Amazing. I can't wait to meet him.
This episode is a deep dive intohow you turn tools for
(02:36):
storytelling into systems for saving lives, inspecting nuclear
reactors, and building the future.
Let's get into it with ASA. Hey, so welcome to Founder mode.
How are you? Yeah, from robotics to sort of
this Oscar winning film stuff. I think we've had some just
incredible walks and conversations over the years.
And now, you know, you know, taking this to construction
(02:58):
sites. I think it's it's going to be
pretty cool to talk through thiswith you.
And yeah, just excited that you're here.
This is a fun one because you know, rewinding, you said this
before we started, but like you're always looking forward,
right? But let's look back for a
second. How did you get into robotics in
the 1st place? It's.
Kind of like, like anything sortof following my interest.
My, my main tenet in life is to not be bored generally just
(03:19):
always looking for for interest.So following my interest, I
started out in in Hollywood working on visual effects,
computer graphics in films. So kind of came up in the design
world working on main titles. I worked on a bunch of, you
know, small things in commercials 1st and then kind of
moved over into a big titles house called Imaginary forces,
(03:40):
worked on Spider man. It was kind of a big, big first
hard mode thing I did. And yeah, I just kept working up
in the in the film world, had myown company and then burned out
of that, got sick of being doingthe creative thing and, and
really just wanted to work on things that had requirements and
specs and robotics is that, and,you know, all the computer
(04:00):
graphic, spatial understanding, camera matching, all this kind
of really high spec work that wehave to do for visual effects
where every single pixel, we literally just over and over
again loop every single shot, even if it's 13, you know,
frames long and everybody in thewhole room is looking at every
pixel in the whole screen. They're all perfect.
And then it's like, you know, finals and you're like, cool,
let's go. So you kind of bring in that
(04:21):
mode of, of attention into the real world.
And so now you've got to make sure that you don't hit things
and you, you're moving perfectly.
And, you know, all this, all this kind of thing that seems
very unrelated, but, but actually is quite related.
And had the opportunity to work with a, with a small group
called auto Fuss. And they, they were starting to
bring these industrial robot arms into into doing motion
(04:43):
control and doing commercials. I started, I, I ended up being
the visual effects supervisor ontheir first big mush control
shoot. And then later on when we
started ramping up for gravity, I, I ended up being the, the
person who took over and yeah, built, built the whole thing
(05:03):
for, for that show. So it was quite the learning
curve kind of went home with thethe stack of manuals for the
Kooka robots and red dot cover to cover and came back on Monday
and started start building stuff.
So that was my first real entrance into robotics was
really just we need to get this done here.
(05:24):
The AP is to that robot. So they figured out and then
ended up in a, in a kind of a more full longer term dev cycle
to, to produce the tools required to service a big movie
like like Gravity, you know, really the strongest visual
effects company, the strongest director and DP and actors out
there. Definitely it's a high bar.
(05:44):
And and so you're really workingat at a set of tooling that is
going to try and be as capable and flexible as possible and as
safe as possible to match that effort.
Yeah, so from those early days, like what sticks with you about
like building for film makers? Like to me it just seems like
this mystical thing, like I was an extra when I was in LA after
UCLA. So Ace and I went to UCLA
together. That's where actually we met,
(06:04):
you know, 1228 years ago, I would say 27 years ago at this
point, and worked on electric car club.
So building, you know, an early kind of electric battery powered
little racer thing that was justso dope.
Yeah, So it's a it's a very filmsets are are are kind of the
alternate between the most beingthe most boring place and the
most exciting place on Earth. And at the time, this is quite a
(06:26):
while ago, 2007 and something like that.
It was very, very challenging. You could not do real anything
sort of real time in terms of real time adjustment of the
camera, which also would have been really unsafe.
And so so it was specifically built to to be kind of on rails,
like virtual rails. It's going to do the same thing
(06:46):
every single time no matter what.
You got four people on on E stops watching it.
We run it at 15 percent, 25 percent, 4050%.
You know, you get it faster and faster.
Your first running with stunt double for a while proving this
move. And so you sort of ramp up from
a safety perspective and it's always exactly the same move.
And so any deviation and the people released their their
(07:09):
hands off the of the stops, you know, or they press the buttons
like a dead man switch in order to kick it back.
You know, we're working in a software called Maya and Maya is
a 3D animation software and it'skind of the big heavy lifter for
the other industry has been still is it's got easy to use
animation tools. So you know, we're sitting there
on set or me or Tarek are sitting there adjusting if few
(07:30):
keyframes to just layer in this adjustment and you push it back
out. You're trying to make that loop
as fast as possible. You have to pull her out, right,
because you haven't actually runthis move.
So to be ultra safe, you're you're pulling her out.
You're then going to run throughthe whole move half, half speed,
full speed, doing all the same things.
Yes, cool, OK, let's do it. Because she was strapped and
basically bolted into kind of anIron Maiden style thing that she
(07:53):
could rotate and it would tilt her forward and it would adjust
her like this. And then this giant camera arm
is like bolted to a massive track racing at her high speed.
You know, it's all industrial equipment.
So it's all very reliable from acompany that's been doing it
forever. Kuka, the big industrial arms
that we've basically are are jamming real time Linux be here
(08:14):
at this time kind of information.
So it's a it's a curve that we are really only adjusting the
time parameter on the curve. So whenever we lock in the
curve, that's curve is safe. We're just making parts of it go
slower or faster essentially. And even we even wired it up to
like a knob. There's like a, you know, like
kind of a knob and it'd be a person, you know, me be like
slowly slowing down a section ofit and then kind of speeding it
(08:36):
up. There's that.
Human. There's that human piece to the
robotics element too. It's not just entirely running
automated. So ASA, there's things about
this that I can imagine are similar as you crossover from
industries like film to construction, but like just take
us through the difference between designing robots for
films like Gravity and and and others to like designing
(08:59):
construction robots, which whichis what you're building now,
right? Yeah.
So it's all the same concerns, right?
Safety, reliability, ruggedness.You always have this user.
That's whatever you're going to introduce something in the end
of these industries and you havepeople who are already getting
work done. They're they rely on their tools
and those tools have to be, haveto always work every time, you
know, you reach for a hammer hammer, it's a simple tool.
(09:19):
It's like it's always got to work.
You can break a hammer, but you know, it takes some, some
malfeasance. And so, so it's really just
trying to make tools. So I feel like I'm a, I'm a tool
tool maker. I am making the simplest,
simplest way for people to get the most work done.
It's really kind of kind of whatit what it sort of always has
been. And the leverage is, is a whole
world of robotics, math, algorithmic evolving and
(09:43):
improving all the time. And then the other side is, is,
is whatever mechanical bases that we can rely on that are off
the shelf as much as possible. And that just helps you with
having it be other problems. Many of the, of the, of the, the
ruggedness. I mean, these arms are good for
15 years of life, You know, they're just there to Weld and
bolt stuff, you know, on, on a factory line forever.
(10:06):
So they've been, they've been well, well, well engineered and,
and other they're a reliable piece of the puzzle.
But then you're kind of going inand inserting yourself nowadays
at, at at this fully dynamic level.
So we've evolved past this very fixed thing that we were doing
in the past and now that 250 Hertz were telling these big
arms exactly where to be. So it's more and more on us to
(10:30):
design, you know, the, the safety and, and, you know,
smoothness basically to make sure that they're not not
jerking, not using too much energy, that they're not hitting
each other, hitting things. So there's a lot of planning
that goes into making this be aseasy to use as possible.
Also, of course, on the UX side,in general, it's got to be safe,
(10:52):
you know #1 it's got to be reliable, you know, probably #2
and then, you know, ease of use is, is, is really the third
thing. And if you don't have those
things that it's, it's just veryhard to get any kind of traction
with, with users. You know, some people make do a
lot of robots and those are all hidden away in a, in a factory,
in a manufacturing line. And that's fine.
You can get a bunch of nerds and, you know, we can all get
(11:13):
around and say, what's the problem?
Oh, OK, got to make sure that these little, little things are
flowing into that thing at the highest rate possible.
And so, you know, you get a teamand build that thing and
eventually they can like make one thing and then they, they
let it go. And there's all kinds of
incredible versions of that thatare, that are out there and
made. But it's just a kind of a
different world when you're, when you're out there directly
having to work with users. And I, I am blessed that my, my
(11:34):
father was a general contractor.I grew up summer jobs, you know,
I learned how to build things inthe shop attached to my,
attached to my house. So construction has been with me
from the beginning. Pictures of my dad basically
built our entire house in phases.
And there are times when I didn't have a door.
It started as like a small shackand a big piece of land and a,
and a, like a lean to carport thing.
And then, you know, now it's this beautiful huge kind of
(11:57):
multi building complex. And, and so anyway, I've, I've
been in the, in the constructionworld and kind of adjacent
watching him, he's finally retired a few years ago.
But watching him trying to trying to get technology into
his world and just seeing that fail.
It's very hard to push technology into the construction
world. The only thing that does it is,
(12:19):
is just, you know, reliability. So you sort of like walk through
like this precision, reliability, you know, and then
this sort of Arc between like, you know, kind of growing up in
a construction site, going into this cinema sort of movie
creative industry and now into back to industrial robots back
in the construction site. And so like, which industry
would you say like pushes those limits more like the limits of
(12:41):
like that precision, repeatability, sort of
ruggedness. Where are you seeing like the,
the most challenging sort of like, you know, wow, we're
really hitting the limits of that.
And obviously the technology's probably changed over those 40
years, right, from what's possible.
But like from from where you sittoday, like what's kind of like
the the limits that are that arethat you're like working within
or really tricky to sort of navigate today?
(13:02):
So it's really interesting time in in the world.
Generally you have LMS have really made talking to the sand
like possible, which is which isamazing and what a great UX
unlock. I'm confused on how to do the
thing and move this move Sandra 10° forward.
(13:24):
Well, here, let me tell you how to do it or let me do that for
you. You can choose your own
adventure and and how deeply youwant to let the LLM's, you know,
control your thing. We're very.
Kevin and I have chosen for complete control.
The LLM's actually are telling us to breathe in, breathe out
right now, like everything we'redoing is completely.
We're basically I'm not there yet.
I'm not there. Yet but master by the AI I'll.
(13:44):
Have to get some pointers. No, don't do it.
It's it may not be right. Maybe wrong.
There's this big new thing, right we can we can talk to talk
to it. So there's a lot of UX
flexibility there so so that side of it but of course we have
to work we're never going to have a thing talking to the
Internet at runtime. So on sites in all these
environments, we have to we haveto have a local local model and
and of course it has to be incredibly reliable.
(14:06):
So we think of it as as unlocking the sort of
flexibility and ease of use at the sort of high level when when
you're sitting there with an interface that so many people
can be complicated because it's got it, You know, you can tumble
around a 3D object and you're seeing all this work happening
or whatever it is. So being able to talk to it is,
is a big new thing. And then the other thing that's
happening in sort of modern robotics is sort of
demonstration learning basically.
(14:27):
And some level you can, you can show, show the sand what you
want to happen. And then you get the sand can
at, you know, 10 Hertz or whatever online while you're
doing an action actually, you know, from from a camera stream
from a or fork sensor or whatever can directly modify the
action. So that gets you, you know, the
ability to fold laundry. Some people are going after
(14:49):
laundry folding or cleaning up acluttered environment.
You know, there are a variety ofthese tasks that are, that are
kind of, or we think of it as a task primitive that are that
require that kind of in in the loop modification.
And that's now tractable. You can do that in 50 examples,
100 examples. And so you can train a thing
that's that's highly reliable, that gets to our standard of
(15:09):
reliability and safety in many, many fewer, fewer examples.
So that levered into the to the rest of all the different task
primitives that we've been able to do before moving things
around and drilling whatever. I like the language task
primitives and then your exampleof folding laundry because my
next questions about how I thinkthe company was called pre NAV,
(15:29):
you flew drones inside of critical infrastructure.
Like I'm just trying to draw some contrast here.
Like some people right now are building robots that fold
laundry. ASA historically you have done
things like teach robots how to fly inside of nuclear reactors,
right? So can you tell us a little bit
about that and like what those constraints were like?
(15:50):
Yeah, as, as you can imagine, they're, they're very, very
careful people, those who run nuclear reactors.
Makes sense? And so that was a very high spec
project. It took us a long time to do it,
primarily because of all the they're they're very careful
kind of ramp up to get to the point where you can you can go
in, fly with during an active mission.
(16:11):
So to the point where where you have to go and weigh every
single part of your drone, like every single component of the
drone, you have to get a full inall up weight so that it gets
weighed before it goes in and itgets weighed when it comes out.
There had better be the exact same mass, the amount of checks.
So it's it's other people. We're also handing it off to
operators who are fully suited up when they're going in.
So it has to be able to be used with gloves.
(16:32):
It has to be inside an environment where where or
essentially sort of hardening itto the point where you're
sending it into outer space almost.
And you, you watch, we just watch the mission through
monitors. And I just had a really great
team of people. And we're we're working with,
we've been doing a lot of development for a long period of
time. And then we got this
opportunity. And so kind of a core group of
(16:53):
of people kind of just hammered on this problem for maybe a year
and a half like that to go through all of the different
checks, proving that we could doit inside at our at our place.
And then we go to their trainingreactor or they have a full
training mock up, which is the full size thing sitting what you
know. Yeah.
And then you know. Exact same size, just.
(17:14):
Yeah, OK. Because it's kind of a it's on
the reactor that's on its side. So we're going up and holding a
radiation meter at each of the little cores of the of the of
the wall. Yeah.
So so then you do it on the fullsize mock up, show what happens
to run the full procedure and then do that enough times.
The movie analogy. The film analogy is like the
stunt double. This is the stunt double nuclear
(17:34):
reactor. Yep, Yeah, yeah.
This is safe. This is safe thinking, right?
You go through and you do everything many times.
This is, I understand this is very different from software.
Yeah, not quite. Someday, maybe, but the industry
hasn't gotten there. Like you talk a lot about the
team and I think and, you know, you said you had a team of
people working together. And I know that, you know,
whenever we talk on our walks, you always talk about like, you
(17:56):
know, folks that you've worked with for a long time.
And my partner JJ and I've worked together now, you know,
over 2 decades and a lot of the same people.
I'm curious for you, like, what is it that makes a great
robotics team? And like what are the kind of
key roles that you see? And has that changed over the
last, you know, 10 years or 15 years or like are that, you
know, today like you would hire the same sort of key people that
you would have hired, you know, in that first sort of like some
(18:18):
of those movie runs? I mean, you learn a lot,
obviously over time. I don't know that I was a great,
great hire of people gravity. I mean, I was the only person
for the software really sovereign hardware piece.
There was there was somebody wasworking on some of the hardware
pieces. So I had to do it all and grab
and and grapple with it myself. And so I didn't have to really
lever other people so much on that on that project.
But then later as you grow and and working on pre now, for
(18:40):
instance, we are definitely havethe requirements of all of the
different disciplines in engineering.
Really, that's the thing that I love about robotics is, you
know, from UX all the way down to the electrical engineering,
mechanical, you're basically touching all these different
domains and you need experts in each of those domains or at
least sort of T shaped people who who've got a strong
expertise and then can can kind of do general things.
(19:01):
And in early companies, you, youreally people who are founder
types off right, people maybe not willing to go all the way to
the point where being founded themselves.
But several of the people that we've brought into this this
this new thing are are like that.
They're self starters, they're going out and it's somebody that
can rely on. Do they know a lot about the
domain I can rely on? If I, if I hand off thing, even
if it's not directly in their domain, they'll go and, and just
(19:22):
chase it down and, and give me good, good feedback, good,
clear, direct response. They're not just going to be
like, Oh, yes, Sir, you need a conversation.
That's yeah, you need reliable people.
You need people who are fluid thinking and, and then they have
to be, of course, you know, justreally, really smart and good at
their domain and be willing to work hard.
I do like the the Ace and I wereboth mechanical engineering
(19:44):
majors, and he's actually doing mechanical engineering where I'm
like, you know. You took you took the road most
traveled and he took the road left well, he.
He finished his degree and I andI dropped out and said I would
rather have to be doing these other things.
Where can folks follow your work?
And more importantly, when stealth mode ends, like where
can people look it up and where's the banner going to flip
from kind of quiet mode to loud mode?
(20:05):
LinkedIn X Twitter What's your where's your spot that folks can
can find you I. Don't do any socials.
I have I have. I have a LinkedIn page but who
cares about that? I love it.
Perfect. Well ASA, this conversation was
next level. Thanks for helping us bolt
together creative tech and hardcore robotics.
OK. It's good, guys.
That was wild. I don't think we've ever had a
(20:26):
guest blend creativity and engineering like that.
Yeah, the way ASA talks about designing robots with both like
perfectly designed automations and also the human touch of like
the dials and knobs, it reframesthe whole conversation.
Yeah, there's this notion of like he talks about UX and like
user experience, the way you're looking at like some beautiful
art project or some sort of likeincredibly like, you know,
(20:47):
choreographed art of, you know, an app on your phone and, you
know, so a bunch of key takeaways, right?
So I think this notion of like Hollywood to hardware, right?
How do you go from like, you know, visual effects for Spider
Man robots for gravity and then,you know, never really looked
back, right? And now it's that creativity
meets industrial where he's taking, you know, the precision
of running sort of a special effect over and over.
(21:08):
Now to like robotics where it's really about how do we do it
safe and how do we do it repeatable?
And then very, very simple, right?
Because that's, you know, non negotiable in construction,
right? You need to be able to have a
tool that you can use in 15 years.
Yeah, just being this tool makerat heart, right?
You know, robotics has never seen the better tools.
But again, how do you make them robot rugged, reliable and
(21:28):
intuitive? And I remember back when we were
building projects as engineeringstudents at UCLA, stuff would
break all the time. It felt like we were in the shop
repairing it. And so I think a lot of what
they're, what they've worked on is really how do you, you know,
complete, complete that repeatability for a place where
people aren't probably taking care of some of those tools,
drones in the danger zones. I mean, the story of flying in
the nuclear reactor and, you know, having to weigh every part
(21:49):
and really understand that whatever went into the reactor
also came out, which is a prettycritical thing.
In a full size stunt double nuclear reactor.
Right, wild, like the fact that they were practicing.
But again, you know, kind of practice makes perfect and
perfect practice is what really brings it home.
And so, yeah, but seeing the future is kind of this software,
right? I'm betting on AI.
But then also, you know, using that real time learning and
(22:12):
these software designed defined robots to actually sort of know
what you're expecting, you know,step by step, but also sort of
realizing that, you know, the humanoid walking around in our
house or with our kids or pets is probably not coming anywhere
soon. That's probably.
For the best. All right, so if you're building
in robotics or you just love where this like concept of tech
meets creativity, send this episode to your friend, your
(22:34):
founder, somebody that's building something.
Just like that and subscribe if you haven't.
You know, lots more stories likeASA coming your way, super
interesting builders and founders coming in the next
couple weeks, and obviously yourreviews help us get the story
out to more people. Yeah, all the links from today's
episode will be in the show notes.
And next week we're going to talk about the intersection of
AI in the enterprise. So how do big companies adopt or
(22:55):
not adopt AI and how can you apply that founder mode mindset
even if you're inside of a big slow moving organization?
That's a wrap, because great engineering, like great
filmmaking, is all about precision under pressure.
Yeah, as, as you can imagine, they're, they're very, very
careful people, those who run nuclear reactors.
So that was a very high spec project.
It took us a long time to do it primarily because of all the
they're, they're very careful kind of ramp up to get to the
point where you can, you can go in and fly during an active
emission. Welcome back to Founder Mode.
(00:27):
Where we talked to founders doing the wildest stuff and
getting their hands. Dirty Yeah, today's topics
robotics that go from the big screen to the biggest hyper
scaler builds. You might think film and
construction have nothing in common, but when it comes to
precision tools, robotics is actually that unexpected bridge.
Jason, you've worked in like tech and media like, what do you
think of of all the AI video models like Google's update the
(00:48):
view? So I watched that that keynote
from IO and I've been seeing allthese videos come out.
I've started to play with the tool.
This stuff is nuts. Like when we're going to talk to
Ace in a minute about film sets and robots that are doing the
most complicated moves with the camera and amazing stuff.
But like when you don't even need the $1,000,000 film set to
create something that looks likea blockbuster movie.
Like when this stuff all converges and you have these
(01:11):
films that are part AI created, part, you know, incredible
talented humans doing it. Like I think it's a 1 + 1 = 3
moment. I think it's going to completely
change the game and the creativity and the artistry that
can come from it. It's not about people's jobs
being destroyed in the entertainment industry.
It's about the job changing, which is what we talk about all
the time on the show. So like as that example, which
(01:31):
I'm talking about, like bringingthis creativity that is being
ushered in by AI into these new markets like film and
entertainment. And we're talking about, you
know, using robotics and construction with ASA.
How do you think AI, robotics and all this comes together
across industries, Kevin? Yeah, it's super interesting as
founders. Like part of the goal is to find
that new tech. And I mean, honestly, what we've
(01:51):
been doing for decades is like thinking about what's the new
and exciting things and how do you bring that back to either
new businesses or old businesses?
And I mean, for this conversation, it actually goes
back to the late 90s when ASA and I met, you know, as students
at UCLA in engineering. And so I think it's pretty
interesting. And I think there's going to be
a bunch to unpack. Here.
Yeah, exactly. And that's what makes ASA
Hammond, our guest today, so fascinating.
(02:14):
His career connects 2 things that you would not expect
Hollywood and hard hats. Yeah, Asa's worked on films like
Gravity Built Drones to inspect nuclear reactors.
I'd like the sickest climbing wall in his last factory, and
maybe if we're lucky, he'll callin from the new factory where
they're now working on robotics for construction.
Amazing. I can't wait to meet him.
This episode is a deep dive intohow you turn tools for
(02:36):
storytelling into systems for saving lives, inspecting nuclear
reactors, and building the future.
Let's get into it with ASA. Hey, so welcome to Founder mode.
How are you? Yeah, from robotics to sort of
this Oscar winning film stuff. I think we've had some just
incredible walks and conversations over the years.
And now, you know, you know, taking this to construction
(02:58):
sites. I think it's it's going to be
pretty cool to talk through thiswith you.
And yeah, just excited that you're here.
This is a fun one because you know, rewinding, you said this
before we started, but like you're always looking forward,
right? But let's look back for a
second. How did you get into robotics in
the 1st place? It's.
Kind of like, like anything sortof following my interest.
My, my main tenet in life is to not be bored generally just
(03:19):
always looking for for interest.So following my interest, I
started out in in Hollywood working on visual effects,
computer graphics in films. So kind of came up in the design
world working on main titles. I worked on a bunch of, you
know, small things in commercials 1st and then kind of
moved over into a big titles house called Imaginary forces,
(03:40):
worked on Spider man. It was kind of a big, big first
hard mode thing I did. And yeah, I just kept working up
in the in the film world, had myown company and then burned out
of that, got sick of being doingthe creative thing and, and
really just wanted to work on things that had requirements and
specs and robotics is that, and,you know, all the computer
(04:00):
graphic, spatial understanding, camera matching, all this kind
of really high spec work that wehave to do for visual effects
where every single pixel, we literally just over and over
again loop every single shot, even if it's 13, you know,
frames long and everybody in thewhole room is looking at every
pixel in the whole screen. They're all perfect.
And then it's like, you know, finals and you're like, cool,
let's go. So you kind of bring in that
(04:21):
mode of, of attention into the real world.
And so now you've got to make sure that you don't hit things
and you, you're moving perfectly.
And, you know, all this, all this kind of thing that seems
very unrelated, but, but actually is quite related.
And had the opportunity to work with a, with a small group
called auto Fuss. And they, they were starting to
bring these industrial robot arms into into doing motion
(04:43):
control and doing commercials. I started, I, I ended up being
the visual effects supervisor ontheir first big mush control
shoot. And then later on when we
started ramping up for gravity, I, I ended up being the, the
person who took over and yeah, built, built the whole thing
(05:03):
for, for that show. So it was quite the learning
curve kind of went home with thethe stack of manuals for the
Kooka robots and red dot cover to cover and came back on Monday
and started start building stuff.
So that was my first real entrance into robotics was
really just we need to get this done here.
(05:24):
The AP is to that robot. So they figured out and then
ended up in a, in a kind of a more full longer term dev cycle
to, to produce the tools required to service a big movie
like like Gravity, you know, really the strongest visual
effects company, the strongest director and DP and actors out
there. Definitely it's a high bar.
(05:44):
And and so you're really workingat at a set of tooling that is
going to try and be as capable and flexible as possible and as
safe as possible to match that effort.
Yeah, so from those early days, like what sticks with you about
like building for film makers? Like to me it just seems like
this mystical thing, like I was an extra when I was in LA after
UCLA. So Ace and I went to UCLA
together. That's where actually we met,
(06:04):
you know, 1228 years ago, I would say 27 years ago at this
point, and worked on electric car club.
So building, you know, an early kind of electric battery powered
little racer thing that was justso dope.
Yeah, So it's a it's a very filmsets are are are kind of the
alternate between the most beingthe most boring place and the
most exciting place on Earth. And at the time, this is quite a
(06:26):
while ago, 2007 and something like that.
It was very, very challenging. You could not do real anything
sort of real time in terms of real time adjustment of the
camera, which also would have been really unsafe.
And so so it was specifically built to to be kind of on rails,
like virtual rails. It's going to do the same thing
(06:46):
every single time no matter what.
You got four people on on E stops watching it.
We run it at 15 percent, 25 percent, 4050%.
You know, you get it faster and faster.
Your first running with stunt double for a while proving this
move. And so you sort of ramp up from
a safety perspective and it's always exactly the same move.
And so any deviation and the people released their their
(07:09):
hands off the of the stops, you know, or they press the buttons
like a dead man switch in order to kick it back.
You know, we're working in a software called Maya and Maya is
a 3D animation software and it'skind of the big heavy lifter for
the other industry has been still is it's got easy to use
animation tools. So you know, we're sitting there
on set or me or Tarek are sitting there adjusting if few
(07:30):
keyframes to just layer in this adjustment and you push it back
out. You're trying to make that loop
as fast as possible. You have to pull her out, right,
because you haven't actually runthis move.
So to be ultra safe, you're you're pulling her out.
You're then going to run throughthe whole move half, half speed,
full speed, doing all the same things.
Yes, cool, OK, let's do it. Because she was strapped and
basically bolted into kind of anIron Maiden style thing that she
(07:53):
could rotate and it would tilt her forward and it would adjust
her like this. And then this giant camera arm
is like bolted to a massive track racing at her high speed.
You know, it's all industrial equipment.
So it's all very reliable from acompany that's been doing it
forever. Kuka, the big industrial arms
that we've basically are are jamming real time Linux be here
(08:14):
at this time kind of information.
So it's a it's a curve that we are really only adjusting the
time parameter on the curve. So whenever we lock in the
curve, that's curve is safe. We're just making parts of it go
slower or faster essentially. And even we even wired it up to
like a knob. There's like a, you know, like
kind of a knob and it'd be a person, you know, me be like
slowly slowing down a section ofit and then kind of speeding it
(08:36):
up. There's that.
Human. There's that human piece to the
robotics element too. It's not just entirely running
automated. So ASA, there's things about
this that I can imagine are similar as you crossover from
industries like film to construction, but like just take
us through the difference between designing robots for
films like Gravity and and and others to like designing
(08:59):
construction robots, which whichis what you're building now,
right? Yeah.
So it's all the same concerns, right?
Safety, reliability, ruggedness.You always have this user.
That's whatever you're going to introduce something in the end
of these industries and you havepeople who are already getting
work done. They're they rely on their tools
and those tools have to be, haveto always work every time, you
know, you reach for a hammer hammer, it's a simple tool.
(09:19):
It's like it's always got to work.
You can break a hammer, but you know, it takes some, some
malfeasance. And so, so it's really just
trying to make tools. So I feel like I'm a, I'm a tool
tool maker. I am making the simplest,
simplest way for people to get the most work done.
It's really kind of kind of whatit what it sort of always has
been. And the leverage is, is a whole
world of robotics, math, algorithmic evolving and
(09:43):
improving all the time. And then the other side is, is,
is whatever mechanical bases that we can rely on that are off
the shelf as much as possible. And that just helps you with
having it be other problems. Many of the, of the, of the, the
ruggedness. I mean, these arms are good for
15 years of life, You know, they're just there to Weld and
bolt stuff, you know, on, on a factory line forever.
(10:06):
So they've been, they've been well, well, well engineered and,
and other they're a reliable piece of the puzzle.
But then you're kind of going inand inserting yourself nowadays
at, at at this fully dynamic level.
So we've evolved past this very fixed thing that we were doing
in the past and now that 250 Hertz were telling these big
arms exactly where to be. So it's more and more on us to
(10:30):
design, you know, the, the safety and, and, you know,
smoothness basically to make sure that they're not not
jerking, not using too much energy, that they're not hitting
each other, hitting things. So there's a lot of planning
that goes into making this be aseasy to use as possible.
Also, of course, on the UX side,in general, it's got to be safe,
(10:52):
you know #1 it's got to be reliable, you know, probably #2
and then, you know, ease of use is, is, is really the third
thing. And if you don't have those
things that it's, it's just veryhard to get any kind of traction
with, with users. You know, some people make do a
lot of robots and those are all hidden away in a, in a factory,
in a manufacturing line. And that's fine.
You can get a bunch of nerds and, you know, we can all get
(11:13):
around and say, what's the problem?
Oh, OK, got to make sure that these little, little things are
flowing into that thing at the highest rate possible.
And so, you know, you get a teamand build that thing and
eventually they can like make one thing and then they, they
let it go. And there's all kinds of
incredible versions of that thatare, that are out there and
made. But it's just a kind of a
different world when you're, when you're out there directly
having to work with users. And I, I am blessed that my, my
(11:34):
father was a general contractor.I grew up summer jobs, you know,
I learned how to build things inthe shop attached to my,
attached to my house. So construction has been with me
from the beginning. Pictures of my dad basically
built our entire house in phases.
And there are times when I didn't have a door.
It started as like a small shackand a big piece of land and a,
and a, like a lean to carport thing.
And then, you know, now it's this beautiful huge kind of
(11:57):
multi building complex. And, and so anyway, I've, I've
been in the, in the constructionworld and kind of adjacent
watching him, he's finally retired a few years ago.
But watching him trying to trying to get technology into
his world and just seeing that fail.
It's very hard to push technology into the construction
world. The only thing that does it is,
(12:19):
is just, you know, reliability. So you sort of like walk through
like this precision, reliability, you know, and then
this sort of Arc between like, you know, kind of growing up in
a construction site, going into this cinema sort of movie
creative industry and now into back to industrial robots back
in the construction site. And so like, which industry
would you say like pushes those limits more like the limits of
(12:41):
like that precision, repeatability, sort of
ruggedness. Where are you seeing like the,
the most challenging sort of like, you know, wow, we're
really hitting the limits of that.
And obviously the technology's probably changed over those 40
years, right, from what's possible.
But like from from where you sittoday, like what's kind of like
the the limits that are that arethat you're like working within
or really tricky to sort of navigate today?
(13:02):
So it's really interesting time in in the world.
Generally you have LMS have really made talking to the sand
like possible, which is which isamazing and what a great UX
unlock. I'm confused on how to do the
thing and move this move Sandra 10° forward.
(13:24):
Well, here, let me tell you how to do it or let me do that for
you. You can choose your own
adventure and and how deeply youwant to let the LLM's, you know,
control your thing. We're very.
Kevin and I have chosen for complete control.
The LLM's actually are telling us to breathe in, breathe out
right now, like everything we'redoing is completely.
We're basically I'm not there yet.
I'm not there. Yet but master by the AI I'll.
(13:44):
Have to get some pointers. No, don't do it.
It's it may not be right. Maybe wrong.
There's this big new thing, right we can we can talk to talk
to it. So there's a lot of UX
flexibility there so so that side of it but of course we have
to work we're never going to have a thing talking to the
Internet at runtime. So on sites in all these
environments, we have to we haveto have a local local model and
and of course it has to be incredibly reliable.
(14:06):
So we think of it as as unlocking the sort of
flexibility and ease of use at the sort of high level when when
you're sitting there with an interface that so many people
can be complicated because it's got it, You know, you can tumble
around a 3D object and you're seeing all this work happening
or whatever it is. So being able to talk to it is,
is a big new thing. And then the other thing that's
happening in sort of modern robotics is sort of
demonstration learning basically.
(14:27):
And some level you can, you can show, show the sand what you
want to happen. And then you get the sand can
at, you know, 10 Hertz or whatever online while you're
doing an action actually, you know, from from a camera stream
from a or fork sensor or whatever can directly modify the
action. So that gets you, you know, the
ability to fold laundry. Some people are going after
(14:49):
laundry folding or cleaning up acluttered environment.
You know, there are a variety ofthese tasks that are, that are
kind of, or we think of it as a task primitive that are that
require that kind of in in the loop modification.
And that's now tractable. You can do that in 50 examples,
100 examples. And so you can train a thing
that's that's highly reliable, that gets to our standard of
(15:09):
reliability and safety in many, many fewer, fewer examples.
So that levered into the to the rest of all the different task
primitives that we've been able to do before moving things
around and drilling whatever. I like the language task
primitives and then your exampleof folding laundry because my
next questions about how I thinkthe company was called pre NAV,
(15:29):
you flew drones inside of critical infrastructure.
Like I'm just trying to draw some contrast here.
Like some people right now are building robots that fold
laundry. ASA historically you have done
things like teach robots how to fly inside of nuclear reactors,
right? So can you tell us a little bit
about that and like what those constraints were like?
(15:50):
Yeah, as, as you can imagine, they're, they're very, very
careful people, those who run nuclear reactors.
Makes sense? And so that was a very high spec
project. It took us a long time to do it,
primarily because of all the they're they're very careful
kind of ramp up to get to the point where you can you can go
in, fly with during an active mission.
(16:11):
So to the point where where you have to go and weigh every
single part of your drone, like every single component of the
drone, you have to get a full inall up weight so that it gets
weighed before it goes in and itgets weighed when it comes out.
There had better be the exact same mass, the amount of checks.
So it's it's other people. We're also handing it off to
operators who are fully suited up when they're going in.
So it has to be able to be used with gloves.
(16:32):
It has to be inside an environment where where or
essentially sort of hardening itto the point where you're
sending it into outer space almost.
And you, you watch, we just watch the mission through
monitors. And I just had a really great
team of people. And we're we're working with,
we've been doing a lot of development for a long period of
time. And then we got this
opportunity. And so kind of a core group of
(16:53):
of people kind of just hammered on this problem for maybe a year
and a half like that to go through all of the different
checks, proving that we could doit inside at our at our place.
And then we go to their trainingreactor or they have a full
training mock up, which is the full size thing sitting what you
know. Yeah.
And then you know. Exact same size, just.
(17:14):
Yeah, OK. Because it's kind of a it's on
the reactor that's on its side. So we're going up and holding a
radiation meter at each of the little cores of the of the of
the wall. Yeah.
So so then you do it on the fullsize mock up, show what happens
to run the full procedure and then do that enough times.
The movie analogy. The film analogy is like the
stunt double. This is the stunt double nuclear
(17:34):
reactor. Yep, Yeah, yeah.
This is safe. This is safe thinking, right?
You go through and you do everything many times.
This is, I understand this is very different from software.
Yeah, not quite. Someday, maybe, but the industry
hasn't gotten there. Like you talk a lot about the
team and I think and, you know, you said you had a team of
people working together. And I know that, you know,
whenever we talk on our walks, you always talk about like, you
(17:56):
know, folks that you've worked with for a long time.
And my partner JJ and I've worked together now, you know,
over 2 decades and a lot of the same people.
I'm curious for you, like, what is it that makes a great
robotics team? And like what are the kind of
key roles that you see? And has that changed over the
last, you know, 10 years or 15 years or like are that, you
know, today like you would hire the same sort of key people that
you would have hired, you know, in that first sort of like some
(18:18):
of those movie runs? I mean, you learn a lot,
obviously over time. I don't know that I was a great,
great hire of people gravity. I mean, I was the only person
for the software really sovereign hardware piece.
There was there was somebody wasworking on some of the hardware
pieces. So I had to do it all and grab
and and grapple with it myself. And so I didn't have to really
lever other people so much on that on that project.
But then later as you grow and and working on pre now, for
(18:40):
instance, we are definitely havethe requirements of all of the
different disciplines in engineering.
Really, that's the thing that I love about robotics is, you
know, from UX all the way down to the electrical engineering,
mechanical, you're basically touching all these different
domains and you need experts in each of those domains or at
least sort of T shaped people who who've got a strong
expertise and then can can kind of do general things.
(19:01):
And in early companies, you, youreally people who are founder
types off right, people maybe not willing to go all the way to
the point where being founded themselves.
But several of the people that we've brought into this this
this new thing are are like that.
They're self starters, they're going out and it's somebody that
can rely on. Do they know a lot about the
domain I can rely on? If I, if I hand off thing, even
if it's not directly in their domain, they'll go and, and just
(19:22):
chase it down and, and give me good, good feedback, good,
clear, direct response. They're not just going to be
like, Oh, yes, Sir, you need a conversation.
That's yeah, you need reliable people.
You need people who are fluid thinking and, and then they have
to be, of course, you know, justreally, really smart and good at
their domain and be willing to work hard.
I do like the the Ace and I wereboth mechanical engineering
(19:44):
majors, and he's actually doing mechanical engineering where I'm
like, you know. You took you took the road most
traveled and he took the road left well, he.
He finished his degree and I andI dropped out and said I would
rather have to be doing these other things.
Where can folks follow your work?
And more importantly, when stealth mode ends, like where
can people look it up and where's the banner going to flip
from kind of quiet mode to loud mode?
(20:05):
LinkedIn X Twitter What's your where's your spot that folks can
can find you I. Don't do any socials.
I have I have. I have a LinkedIn page but who
cares about that? I love it.
Perfect. Well ASA, this conversation was
next level. Thanks for helping us bolt
together creative tech and hardcore robotics.
OK. It's good, guys.
That was wild. I don't think we've ever had a
(20:26):
guest blend creativity and engineering like that.
Yeah, the way ASA talks about designing robots with both like
perfectly designed automations and also the human touch of like
the dials and knobs, it reframesthe whole conversation.
Yeah, there's this notion of like he talks about UX and like
user experience, the way you're looking at like some beautiful
art project or some sort of likeincredibly like, you know,
(20:47):
choreographed art of, you know, an app on your phone and, you
know, so a bunch of key takeaways, right?
So I think this notion of like Hollywood to hardware, right?
How do you go from like, you know, visual effects for Spider
Man robots for gravity and then,you know, never really looked
back, right? And now it's that creativity
meets industrial where he's taking, you know, the precision
of running sort of a special effect over and over.
(21:08):
Now to like robotics where it's really about how do we do it
safe and how do we do it repeatable?
And then very, very simple, right?
Because that's, you know, non negotiable in construction,
right? You need to be able to have a
tool that you can use in 15 years.
Yeah, just being this tool makerat heart, right?
You know, robotics has never seen the better tools.
But again, how do you make them robot rugged, reliable and
(21:28):
intuitive? And I remember back when we were
building projects as engineeringstudents at UCLA, stuff would
break all the time. It felt like we were in the shop
repairing it. And so I think a lot of what
they're, what they've worked on is really how do you, you know,
complete, complete that repeatability for a place where
people aren't probably taking care of some of those tools,
drones in the danger zones. I mean, the story of flying in
the nuclear reactor and, you know, having to weigh every part
(21:49):
and really understand that whatever went into the reactor
also came out, which is a prettycritical thing.
In a full size stunt double nuclear reactor.
Right, wild, like the fact that they were practicing.
But again, you know, kind of practice makes perfect and
perfect practice is what really brings it home.
And so, yeah, but seeing the future is kind of this software,
right? I'm betting on AI.
But then also, you know, using that real time learning and
(22:12):
these software designed defined robots to actually sort of know
what you're expecting, you know,step by step, but also sort of
realizing that, you know, the humanoid walking around in our
house or with our kids or pets is probably not coming anywhere
soon. That's probably.
For the best. All right, so if you're building
in robotics or you just love where this like concept of tech
meets creativity, send this episode to your friend, your
(22:34):
founder, somebody that's building something.
Just like that and subscribe if you haven't.
You know, lots more stories likeASA coming your way, super
interesting builders and founders coming in the next
couple weeks, and obviously yourreviews help us get the story
out to more people. Yeah, all the links from today's
episode will be in the show notes.
And next week we're going to talk about the intersection of
AI in the enterprise. So how do big companies adopt or
(22:55):
not adopt AI and how can you apply that founder mode mindset
even if you're inside of a big slow moving organization?
That's a wrap, because great engineering, like great
filmmaking, is all about precision under pressure.
Popular Podcasts
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Special Summer Offer: Exclusively on Apple Podcasts, try our Dateline Premium subscription completely free for one month! With Dateline Premium, you get every episode ad-free plus exclusive bonus content.
The Breakfast Club
The World's Most Dangerous Morning Show, The Breakfast Club, With DJ Envy, Jess Hilarious, And Charlamagne Tha God!
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.