11-04-25 *INTERVIEW* AI Powered Soldiers with Mike LeBlanc of Foundation Future Industries

Episode Transcript

Speaker 1 (00:00):
So listeners may recall that, oh, two weeks ago, ish
I had a guest on the show, author author Alex
de Mill, who, along with his famous father Nelson de Mill,
they co wrote a novel called The Ten Men. And
you will recall, if you heard that conversation, that Nelson

(00:20):
passed away before the book was finished. And I had
Alex on the show to talk about his dad and
talk about the book. And the subject of the book
is the military using humanoid essentially human looking robots as soldiers.
And there's a lot more to the story. That's just
the very kind of top line. And we're not to
talk about the book today, but that had me thinking
about this, and I started doing a little bit of

(00:41):
reading online about robots as soldiers, about using robots in
war and I found a fascinating company online called Foundation
Future Industries. And Matt mike LeBlanc, who is I had
Matt earlier on the show.

Speaker 2 (00:58):
Mike LeBlanc, who is co.

Speaker 1 (01:00):
Founder of Foundation Future Industries and served in the US
Marines and is joining us talk about the future of
robots in warfare.

Speaker 2 (01:11):
So thanks so much for being here, Mike.

Speaker 1 (01:13):
I really appreciate I'm really looking forward to this conversation. Yeah,
thanks for having me Ross one one nerdy question because
I'm I'm a nerd and I assume you are too.
Your company's name Foundation Future Industries. Is that a reference
to Isaac Asimov?

Speaker 3 (01:32):
It is.

Speaker 4 (01:33):
Yeah, that's that's where we got the foundation and then
wanting to be like Tony Starks what got us the industries?

Speaker 2 (01:39):
Awesome?

Speaker 1 (01:40):
All right, So tell me a little bit, just at
the highest level, what your company does, and then we're
going to jump into some of the ins and outs
of the potential of robots in war.

Speaker 2 (01:51):
Yeah.

Speaker 4 (01:51):
So you know, everybody's familiar with you know, open AI, chat,
GBT and what Generator of AI has done. The puzzle
that really started to unlock was the idea of embodied AI.
And so all of a sudden, you've seen tons of
robotics companies come out over the past couple of years
that are putting that same AI into robots, not just

(02:12):
having them know things like Alexa, but also being able
to move through the world and intelligibly then actually move
things around. And so we started last April, we already
have a fully walking, fully functioning humanoid.

Speaker 3 (02:26):
So in the shape for the human human, it's about
the same size as me.

Speaker 4 (02:29):
It's a one hundred and seventy five pounds about six feet tall,
and we have these working twenty four.

Speaker 3 (02:34):
Hours in factories right now.

Speaker 4 (02:36):
And then we also have some defense contracts with Army, Navy,
and Air Force. With that we're using to test these
for deployments into combat and non combat rules.

Speaker 1 (02:45):
Okay, before we get to the combat, give us a
sense of the functions these robots are performing for companies
that are not in the defense space right.

Speaker 4 (02:56):
Now, it's almost all auto manufacturing, and so you know, well,
first you know, when you walk into any of these plans,
they're filled with robots already.

Speaker 3 (03:04):
The robots are just not.

Speaker 4 (03:05):
In the shapes of humans, right and so every time
I see a human, I asked the plant manager, why
is a human still doing this? Why hadn't you automated it?
You'd be surprised. They have very good reasons for this.
It's usually because they don't have high dexterity and locomotion,
and so you know, it can plug right in and
pack parts, move clips around things like that.

Speaker 1 (03:23):
Okay, So one of the things I wonder about the
shape of the robots humanoid, why is it an optimal
shape functionally. Is it something that is easier to integrate
into some kind of work environment because people who are

(03:44):
managing them can quote unquote relate to them better because
they look kind of like people I like. I wonder
especially the first question, I really wonder, are we an
optimal shape for something? You know?

Speaker 4 (03:58):
I think that the real thing is to ask is
are we optimal beings?

Speaker 1 (04:03):
Right?

Speaker 4 (04:03):
Because the world as we imagine it is perfect. You
hear every software company come out and talk about how
you're going to open up your phone and it's going
to make you coffee and open your windows. That's not
how the world really works. The world works in workarounds.
You know. You can go to the fanciest restaurant on
the planet and if their table is wobbly, they're gonna
put a piece of cardboard over it.

Speaker 3 (04:25):
It is just we are a world of workarounds.

Speaker 4 (04:27):
And because we are like that, having a human shaped
robot makes you able to fix any.

Speaker 3 (04:33):
Problem that you have.

Speaker 4 (04:34):
So instead of trying to create the perfect optimal shape
for whatever that small task is, you can create this
one sque right, put all of your money into making
this as cheap and effishient as possible, and you will
replace any human sized task, which is a ton of tasks.

Speaker 3 (04:48):
So that's how we kind of landed on the form factor.

Speaker 2 (04:51):
Very clever, very very clever.

Speaker 1 (04:54):
I should mention well, I mentioned that, I mentioned that
Mike served in the Marines. And this is kind of
funny based on a comment I said to you before
we went on the air. But a listener wants to
know if the humanoid robots will eat crayons.

Speaker 3 (05:12):
You know it is.

Speaker 4 (05:13):
It's kind of funny you asked that because I always
think everybody's all afraid of AI. AI really has taken
over the spot of the midwit. If you have like
a one ten IQ one ten to one twenty, your
job is pretty much done for right now, and it
is moving out to the edges. So they cannot eat
crans yet, they cannot win Nobel prizes yet.

Speaker 2 (05:33):
But AI works from the middle out. Uh huh that's
did you say mid wit? Yeah?

Speaker 3 (05:41):
Yeah, the midwits I had.

Speaker 4 (05:42):
I had an old professor that used to always refer
to the midwits because he said midwits are the most
dangerous people on earth because most people can speak and
understand within about thirty IQ points of themselves, and so
that's why when you have really smart people that run
for office, they don't connect with anybody, or we have
really dumb people, but all of your politicians basically fall
into that category of midwidth.

Speaker 2 (06:02):
Wow.

Speaker 1 (06:04):
Wow, that is a fantastic and exceptionally elitist concept that
I fully approve of.

Speaker 2 (06:13):
I have a book on my bookshelf.

Speaker 1 (06:15):
Well, it's in a box right now because we're renting
a house and I didn't unpack everything, but I have
a book on my bookshelf that I've had for probably
thirty years, entitled in Defensive Elitism. And I'm not going
to get into all that right now, but boy, would
the concept of midwit fit in there perfectly. And actually
there are actually enormous societal conversations to be had about.

Speaker 2 (06:35):
That, as you just did regarding who AI is going
to replace.

Speaker 1 (06:38):
All right, so I probably need to have you back
to talk about that stuff because that's fascinating. Okay, let's
talk about now, robots as warriors, robots as soldiers, robots
in defense. First of all, and this is something you
mentioned to me just in passing kind of before we
got onto the on the air. Your company foundation, Usure Industries,

(07:01):
is one of the few, one of the maybe the
only that is looking to make robots to be used
as soldiers.

Speaker 2 (07:10):
Uh why, yeah, well, you know, I.

Speaker 4 (07:14):
Mean there are there are a lot of companies that
are doing this in China and for China, so you know,
the the nuclear bomb has already been created and it's
already going to be used by by potential adversaries.

Speaker 3 (07:27):
But in the US there.

Speaker 4 (07:28):
Are there are six humanoid companies that have been valued
at a billion dollars or more. We're one of those six.
Of those six, we're the only ones that are working
with the Department of War, and most of those six
have signed a non weaponization pledge saying that these robots
will never be armed.

Speaker 3 (07:43):
In any way.

Speaker 2 (07:43):
Wow.

Speaker 4 (07:44):
And you know, I think I think this really comes
back to it's it's just different approaches to.

Speaker 3 (07:48):
Life, right.

Speaker 4 (07:49):
There are there are other USAU people who look at
the world as everyone is really nice and we need
to try to educate them to be even nicer. That's
where the world should be. And then there are the
Hobbes people that say the reason you lock your doors
at night is because you know that people are fundamentally
greedy and want to take your stuff.

Speaker 2 (08:05):
Mm hmm.

Speaker 3 (08:06):
I think that we are.

Speaker 4 (08:08):
You know, it's it's us and Tesla are probably the
only two that live in the Hobbsy in reality, and
Tesla has shown.

Speaker 3 (08:14):
No interest yet in the Department of war.

Speaker 2 (08:16):
Wow.

Speaker 1 (08:19):
Yeah, interesting, it's interesting. Hobbes analogy is solitary, solitary, poor, nasty, brutal,
and short.

Speaker 2 (08:30):
I think that's right.

Speaker 1 (08:31):
I mean, you're the philosophy major, so you tell me solitary, poor, nasty, brutal,
and short.

Speaker 2 (08:36):
I think. I think.

Speaker 1 (08:37):
Okay, here's what I think. I think within the United States,
the truth is somewhat closer to the idealistic version than
the Hobbsyan version. But you must defend yourself against the
Hobbsyan version, even if that's a very very small percentage
of the population.

Speaker 2 (08:55):
And then when you start talking about, you know.

Speaker 1 (08:56):
Wars with foreign powers, then it's all Hobbes all the time.
I mean, you think about Vladimir Putin and wherever you
want to go. Obviously you have to assume that if
you're talking about war. So tell us some of the
capabilities of the phantom MK one.

Speaker 4 (09:15):
So right now we have the robot that has contracts
with the Army of the Navy and the Air Force.
As I mentioned, those are doing things like maintenance of aircraft.
So we'll look at a C one thirty and we'll
do all the inspections to make sure the tires are
filled with air, make sure all of the rivets haven't
been up properly, things like that before they take off.
But we are beginning to move into things like mount operations,

(09:37):
mount is military operations in urban terrain, so going in
clearing cities, things like that. You know, we had a
saying when I was in the Marine Corps, never send
a marine where you can send a bullet first. We
now say, never send a marine where you can send
a robot first. The idea is to have this robot
enter the room and be the bullet sponge instead of
a nineteen year old American that has to go into

(09:58):
that space. We're also working with a lot of breaching capabilities,
so breaking into doors, breaking in the windows, shooting off locks,
those those types of things to.

Speaker 3 (10:07):
Again take take marines out of harm's way.

Speaker 1 (10:09):
Okay, couple follow ups. Let's on the first function. Let's
say aircraft maintenance. Does the robot see the rivet and know,
all right, there's a rivet.

Speaker 4 (10:23):
I need to.

Speaker 1 (10:23):
Inspect that I'm going to use whatever my hands can
do or whatever tools I put in my robot hands,
and you know, make sure that rivet is tight enough
and then see here's the next rivet to the right.
Now I need to do that one or is it
somehow more of a mental map? And so if you
covered the robot's eyes, that would still be able to
do that same function.

Speaker 4 (10:45):
This, So we're we're doing this all with AI and
the perception capabilities are amazing. So when the robot is
reaching for something, you can move whatever that piece is
and it will follow the part and we'll pick it
up from where it is. It knows how to maneuver
its hands to be able to pick up things like
rivets those you know, that's really like a machine learning capability.
So just the same as they used to have people

(11:06):
at Google that would click on a bunch of pictures
of cats to teach the system what a count was,
we have the same thing through video data that shows
it what a good rivet looks like and what a
bad rivet looks like.

Speaker 1 (11:17):
Wow, so it really does see it, and it really
is responding to its environment more or less the same
way that a human with that task would.

Speaker 4 (11:28):
It's amazing once you actually start breaking down what a
human being does in any given situation, you realize so
much of it is just based on images to come
in and so having things like cameras and being able
to just make basic sense of what's in that picture.

Speaker 3 (11:42):
That's what we're doing all the time in blue collar work,
in defense, in defense work when we're driving cars. So
that's exactly what the robots do.

Speaker 1 (11:49):
Wow, we're talking with Mike LeBlanc. He's co founder of Foundation.
Future Industry is the website Foundation dot bot like Robot
Foundation dot bo he and they are, as far as
we know, the only American company making robots designed to
go to war.

Speaker 2 (12:10):
So, right now, what is.

Speaker 1 (12:13):
The capability of a robot And if it doesn't have
the capability right now, when will it to uh pick
up an M four or or sit in the back
of a pickup truck with a finger on a on
a you know, on a mounted fifty or three oh
eight and and and take out the enemy and and

(12:34):
target and not and not engage in friendly fire.

Speaker 4 (12:40):
Yeah. Well, and and that's that's kind of an important point,
you know, to this perception. I think that a lot
of people picture a humanoid on the battlefield and they
picture Terminator, and they think this robot's going to determine
a good guy from a bad guy, just like a
good rivet from a bad rhytht Uh.

Speaker 3 (12:54):
You know.

Speaker 4 (12:55):
The fact of the matter is we have tons of
robots already on the battlefield. Drones are are airborne robots,
and there is no drone in the military that is
making decisions about who and when to kill. There are
I used to be in air strikes, right, one of
the many bureaucratic layers. But of all these people that
do intelligence work to make sure that we know when
we're going to drop a bomb humanoids, every conversation that

(13:17):
we have is exactly the same.

Speaker 3 (13:19):
There would be many approovals processes from.

Speaker 4 (13:20):
The humans, so the human not just in blubo human,
on top of every action of the robot.

Speaker 3 (13:26):
That's the basic idea.

Speaker 1 (13:27):
Yeah, but if you had a platoon size, much less
a battalion size group of these robots going out in
a land war in Europe or to defend Taiwan, how
do you manage a group of semi autonomous or potentially
fully autonomous things.

Speaker 2 (13:48):
How can you manage that and make.

Speaker 1 (13:50):
Sure that you have the control that you think you
have and that you intend to have.

Speaker 4 (13:55):
So it'll be very similar to how we're doing it
in auto manufacturing, which is the robot know how to
do all their tactical measures, so they know when to
pick something up and if that thing move.

Speaker 3 (14:04):
They know they know where to reach for it. Yeah,
but you have a fleet.

Speaker 4 (14:07):
Management system that says, here's the line you need to
be working on.

Speaker 3 (14:11):
Here are the parts that you need to be moving.
The same would be true of a platoon.

Speaker 4 (14:14):
You'll be able to give it a mission, just like
you would give a opportunity for marines, and then they'll
be able to carry out the individual actions. So if
there's a step over them, they're going to know to
take a step up to retire and things like that.

Speaker 3 (14:25):
But all the major decisions will be made by a human.

Speaker 1 (14:29):
I'm still wondering, like in the heat of the moment,
like do you tell a robot that's going to take
out an HVT and a building in Fallujah? You tell
them like just just assume everyone you see is a
bad guy, or do you say, well, there might be
women and children there. I bet you could probably teach
a robot what a child looks like and don't shoot

(14:49):
a child. But that sounds to me like a very
very difficult thing to program.

Speaker 4 (14:55):
Yeah, so far they haven't had any conversations that wouldn't
have a human actually being there for world about where
to pull the trigger and really in this right in auto
manufacturing makes no sense to have one to one human
to robot, but in the military it absolutely does, as
long as that person is taken out of harm's way,
just the same as we have phone operators.

Speaker 1 (15:12):
Interesting, my colleague Mandy Connell just came in and I
told her she really needs to listen to some of this,
and Mandy has what's your question for Michael Blanc?

Speaker 5 (15:21):
I mean, the biggest question that I have is how
are these unhackable? Because as many people say, hey, we've
got a great security system here, ultimately, when you are
remotely sending directions into your robot army, that is something
that can be exploited by the enemy. How do you
protect against that?

Speaker 4 (15:41):
Yeah, Well, one of the biggest differences in our AI
approach is that we are trying to put all the
compute at the edge, which means instead of having all
of the decision making capabilities take place in the cloud
and then be deployed to a robot take actions. We're
trying to put as much on the robot locally as possible,
so it would actually take you know, breaking into one

(16:01):
individual robot.

Speaker 3 (16:02):
You wouldn't be able to take over the entire fleet.

Speaker 4 (16:04):
For example, if a robot we're getting broken into acts
that would take some kind of physical response from a
person and that would be an easy thing to shut
down the robot.

Speaker 5 (16:14):
So how do the robots work together as a team
if they are autonomous in that way?

Speaker 3 (16:20):
Yeah, Well, that's what we're really building toward.

Speaker 4 (16:22):
And we've been working with a lot of companies in
the defense space for this because the idea of having
all these coordinated not only with humanoids humanoid, but humanoid
to drone and other sensors. That's really the picture of
the future of warfare is to get all these actions together.
So that is not built out yet for any of ours.
We've been focused very much on individual actions, improving exactly

(16:43):
what each robot can do. But that's what we're building toward,
is one single picture that they're all all operating off of.

Speaker 2 (16:50):
Wow.

Speaker 1 (16:51):
Just remarkable. So what would you say, how close do
you think you are to a robot really being able
to perform the kind of function we've just been talking
about now, like go out in combat while holding a rifle.

Speaker 3 (17:04):
Yeah, I think we're looking at one to three years. Wow,
that saints get there. You know.

Speaker 4 (17:08):
One of the one of the biggest misconceptions about what's
going on with embodied AI is that people think this
is a ten to fifteen year problem.

Speaker 3 (17:16):
And we've been around for eighteen months.

Speaker 4 (17:18):
We already have a robot that is working twenty four
hours a day, five days a week on a manufacturing floor.
We did I just I just got the email we
got We did nineteen thousand parts and on one line
last last month.

Speaker 3 (17:30):
I mean, this thing is, this thing is in business.

Speaker 4 (17:32):
This thing is out there today commercially, and these steps
in the military.

Speaker 3 (17:36):
It's it's not a far it's not a far crowd amazing.

Speaker 1 (17:39):
Are are you allowed to answer the question of how
much it costs to buy one of these robots?

Speaker 3 (17:44):
Yeah? Right now.

Speaker 4 (17:45):
Commercially for our first first five customers, it is one
hundred thousand per robot per.

Speaker 2 (17:50):
Year that a year will go that's pretty much. Yeah,
that's so. Yeah, that's interesting. It's sort of a subscription model. Right,
really completely.

Speaker 4 (18:01):
Yeah, and that includes all field service and maintenance everything.
We've tried to make this really easy to buy, so
instead of having to go get a big capital expense,
which people have to do when they typically buy a robot,
you can just remove your payroll. So if you have
a work style that is working twenty four hours a day,
you're probably paying those people each sixty grand, say one
hundred and eighty thousand a year. That is where we

(18:23):
come in. Have huge forty plus percent savings. And the
robot is plugged in today so it can work all
all those twenty four hours and yeah, safety.

Speaker 2 (18:31):
Money day one, absolutely incredible. All right.

Speaker 1 (18:33):
I could talk to you for hours more, but we
got to go. I'll definitely get you back on the show.
Mich LeBlanc is co founder of Foundation Future Industries.

Speaker 2 (18:42):
Foundation.

Speaker 1 (18:43):
Dot Bot the only American company, as far as he knows,
and as far as I know, and I don't know anything,
but as far as he knows, the only American company
making robots with the intention of being used.

Speaker 2 (18:54):
In war, which is inevitable by the way.

Speaker 1 (18:57):
And these other American companies that aren't doing it that
are just cutting themselves out of a market. That's really important. Mandy,
did you want to add something real quick?

Speaker 5 (19:04):
I can hardly wait for war to be rock'm soccerm robots.

Speaker 2 (19:08):
I can't wait.

Speaker 3 (19:11):
Come on, it's gonna be amazing.

Speaker 1 (19:13):
I can't wait for a robot that has a Marine
Corps insignia on it eating crayons.

Speaker 2 (19:20):
Ah, yes, that.

Speaker 1 (19:23):
Will be the best. Mike, thank you so much. This
was really fantastic. We'll do it again. And congratulations on
your success so far.

Speaker 2 (19:29):
And I wish you a lot more.

Speaker 1 (19:31):
And I'm very grateful that you guys are focused on
defending the country when so many other companies are refusing
to take part.

Speaker 2 (19:37):
I'm proud of you for that.

Speaker 3 (19:39):
All Right, Ross, Mandy, thank you so much.

Speaker 2 (19:41):
All Right, good talking to you.