Latest / Founder Talk with Alex Sheridan / The Rise of Autonomous Lawnmowers (And What It Means for Jobs) With Ilya Sagalovich

Transcript

The Rise of Autonomous Lawnmowers (And What It Means for Jobs) With Ilya Sagalovich · Founder Talk with Alex Sheridan

0:00
Are you at all worried that AI could potentially take over
0:03
humans at some point? We're trying to compete with
0:06
China on AI and we try to compete with Russia you.
0:08
Have to, otherwise we're really screwed up.
0:10
We don't go aggressive with that, yeah.
0:11
Exactly. And at the same time, you can
0:13
almost think of it as being like, OK, what kind of security
0:16
protocols we have in place that if the AI in one of the
0:20
countries goes bad, Yeah, that we know it's the AI that's going
0:23
bad and other people. Right.
0:24
Oh. Man, that's a great point.
0:26
You're doing some really unique things.
0:27
So our company develops software that makes commercial equipment
0:32
smart. This is kind of crazy.
0:34
So 0 tone lawn mowers that drive without a human operator.
0:41
So Ilya Sagalovic. You got it, Did I?
0:43
Yes, Sagalovic. Yeah.
0:45
Cool, There we go. That was Haven Shine as the
0:48
business, right? And I think just I'm really
0:51
excited to talk to you today because I don't know, I just
0:52
nerd out about autonomous and AI and like where the future is
0:55
going with this stuff. And you're doing some really
0:58
unique things. Like when I saw what you were
0:59
doing, I got referred to you by another person that was on the
1:01
podcast, Ryan, and I looked at your profile and LinkedIn and I
1:05
went to your website. Now it's like, man, no one else
1:07
is doing this stuff. This is really cool.
1:09
So first, just to give the audience context, you know,
1:12
other business owners, entrepreneurs, people like that,
1:15
how would you sum up what your business is in a sentence or
1:18
two? So our company develops software
1:24
that makes commercial equipment smart.
1:28
It makes it very smart. And what's an example of that?
1:31
Just so the audience knows. To be more specific, we're
1:34
focusing now on the automation of commercial landscaping
1:39
equipment. Yeah.
1:41
So 0 tone lawn mowers that drive without a human operator.
1:46
It's wild. Yeah.
1:49
So is it when you say lawn like landscaping equipment?
1:53
I saw the zero turn lawn mower, which it was just like driving
1:55
by itself and I'm like this is kind of crazy.
1:58
What else is it? Is there in that product line or
2:00
is that it for now? So we actually have different
2:05
levels of autonomy, everything from products just do remote
2:09
control of commercial mouse. So you're just.
2:12
Sitting in your house, mowing the lawn, just controlling it
2:15
like that in cynical in. Some way or remote control being
2:20
added to an existing machine on site all the way to our psychic
2:25
product which allows one human to operate multiple machines at
2:28
the same time, almost like in a tethered flock like
2:32
configuration. And then we have a product which
2:36
is what we call our autonomous machine, our hero product, which
2:42
operates by itself on large open sites.
2:47
Very cool. Can you explain how autonomous
2:50
vehicles just work in general? Because like, obviously we've,
2:53
you know, I've taken a Waymo in Arizona, which was kind of
2:56
crazy. I remember getting in and I was
2:58
like, I was put there. I was in Phoenix with my
3:01
girlfriend on a trip and I was like, they got these Waymo
3:04
things where they just drive you around.
3:05
There's nobody in the car. It's like an Uber, like a
3:07
driverless Uber. And she was like, that seems
3:10
kind of crap. I don't know if I trust that.
3:11
And I was like, let's just check it out.
3:12
And so we fly into the airport and we get there and we're, you
3:16
know, getting ready to hail an Uber or whatever to our hotel
3:19
and we see one of the Waymos. And I'm like, babe, this is one
3:21
of the Waymos. Check it out and I tell you
3:23
what, like really it just the way it pull it drops someone off
3:27
fine. And then the way it pulled out
3:29
was like really just happened to be like really awkward.
3:32
It like it kind of stalled in the ministry and then kind of
3:35
went and she was like, hell no, I'm not doing that.
3:38
And I said, all right, let's just maybe we'll try it later on
3:40
in the trip. So later on in the trip, a day
3:43
later, we're going to dinner. It's like a mile away from the
3:45
hotel. I said, let's just try the way
3:47
MO. It's a mile away.
3:49
So like nothing crazy can happen.
3:50
You know, we're not getting on the Interstate or whatever,
3:52
which by the way, I looked up the safety stats, they're safer
3:54
than human drivers. So there's that, right?
3:57
And so we get in the way mode and it's like kind of weird at
3:59
first, you know, you're like, there's just no one in there
4:01
with you. It's just, it's a little bizarre
4:03
when you're used to your whole life driving with somebody and
4:05
there's just, it's just AI, it's technology, it's autonomous.
4:09
And then we get there and it did a great job and we're like, all
4:11
right, let's do it on the way back.
4:13
So we do it to get ice cream and then we do it on the way back to
4:16
the hotel and then we Uber it on the way from the hotel back to
4:20
the airport to leave to come back here.
4:22
And by the end of it, I was like, babe, what do you honestly
4:25
feel is more safe? Like what do you, what would you
4:27
rather drive in the Waymos or the Ubers with the humans that
4:30
are on their phones and like too distracted and stuff?
4:33
And she goes the Waymos. So like the whole thing just
4:36
like that's how fast I think our mindset can shift around this
4:38
stuff. But back to the original
4:40
question, how does the how does autonomous vehicles, how do they
4:44
work? How does that technology work?
4:47
Alex, great question about the technology and just to echo your
4:52
point, I love the way more and I love what Google and Alphabet
4:55
are doing in that space. A few other companies are trying
4:59
to get into that. This is what we call on road
5:03
autonomy, on road autonomous vehicles.
5:06
And what we do is slightly different in the sense that is
5:09
more similar to agricultural tractors, right?
5:13
The way of off road vehicles like commercial mowers and
5:17
tractors navigate is slightly different from the on the road
5:20
ones. But let's talk about let's.
5:22
Dive into all of them. I'm nerding out about this.
5:24
Yes, go ahead. So when you have something like
5:29
way more, it would usually rely on the combination of sensors.
5:34
For those who like technology, this is what they called sensor
5:36
fusion. So it takes data from lighter
5:39
sensors on the Wimo, takes data from cameras.
5:42
It also knows its approximate GPS location, and then it has a
5:48
high definition map that it navigates through.
5:50
So it either detects sidewalks, streetlights, buildings and it
5:55
localizes itself within that map.
5:57
Right. So in case one's a little bit
5:59
off it, it has a couple different things to really make
6:01
it accurate. Exactly, exactly.
6:03
And it's fascinating that the connection at the gate, a car in
6:08
the middle of the city with pedestrians and bicycles from
6:11
one place to the other safely and precisely.
6:15
And this is kind of the state-of-the-art of what
6:20
autonomous driving technology can do.
6:24
What is special about Waymo is Waymo specifically has, you can
6:29
almost think of it as an unlimited amount of capital
6:32
investment because it is backed by Alphabet, which is the parent
6:36
company of Google, right? Several other companies trying
6:41
to develop such an advanced autonomy stack, not all of them
6:45
are around today. It's incredibly fascinating to
6:48
see the ones that are actually deploying this technology on
6:51
real life applications and they're able to operate it
6:53
safely and efficiently. So Google basically owns Waymo.
6:58
Is that kind of what it? Yeah, exactly.
7:00
So Google owns Waymo and I don't think.
7:02
A lot of people know that, too, yeah.
7:04
Yeah, most people don't know that Waymo was actually started
7:08
in a sense back in 2005. Wow.
7:11
It was kind of originated from Google's self driving
7:16
competition program, which then grew into this more mature
7:20
organization. Isn't that how I think that's
7:22
how big companies should do it? Like try to find things that
7:25
might put you out of business in certain sectors and go start us,
7:28
you know, start up incubator and let it see if it can grow into a
7:31
company versus letting somebody else do it.
7:33
Exactly, Alex. It's everybody knows that this
7:36
is a multi billion dollar space. If it's successful and it's
7:40
going to be disrupted for our cities, it's going to disrupt
7:43
how we commute from one place to another.
7:45
It would change the whole idea of vehicle ownership because you
7:49
would question yourself is it makes sense for me to own the
7:51
vehicle or should I just use a way more?
7:55
But at the same time, it's incredibly complex and it takes,
8:00
I would almost say, a generation to develop this technology
8:03
correctly because of how complex it is.
8:06
What does that mean for a company like Uber?
8:08
If there's autonomous cars everywhere, is Uber still
8:10
around? Well, actually it's it's funny
8:13
that you ask about Uber because way most primary business model,
8:18
the way they make money is transporting people around.
8:21
So it's essentially A robotaxi business model, which in a sense
8:25
competes with Uber. It now competes with Tesla.
8:29
Yeah, I know Elon's big into that.
8:30
And it competes with anybody else who is trying to move
8:33
people from one location to the next, right?
8:38
What we do for our company at Havenshine, we don't actually
8:41
transport people between different locations with
8:43
transport equipment. But who knows, maybe we should
8:46
be offering free rides to some people on our lawn mowers.
8:50
Well, I'm just thinking in general, not even we're going to
8:52
get into your stuff specifically because I'm really excited to
8:54
talk about that. But even just like is Uber then
8:58
are they going to just merge their technology and their
9:00
audience and their database? Because I imagine there's a lot
9:02
of value in the amount of customers Uber has and to just
9:06
quickly integrate into them. But are they partnering with
9:09
Waymo? And I'm just curious, I know
9:10
you're not in that like smaller niche of cars, but certainly
9:15
with the autonomous, you're on top of it.
9:16
And I think you've came from that world and that kind of
9:18
stuff too. What's your read on that?
9:20
Yeah. So Uber is one of the companies
9:23
that try to develop autonomous equipment, autonomous on road
9:27
vehicles. Yeah.
9:29
And unfortunately they had one accident that has shut the
9:32
program down. And that's actually very common
9:36
for the on road vehicle industry.
9:37
One small accident can shut down an entire program because there
9:41
is a lot of regulation involved, a lot of human life concern, and
9:45
so forth. Even though humans make
9:46
mistakes. Even though way more way MO.
9:51
Oh, you like that? Well, exactly, Alex.
9:53
Like this is the thing. It's like if I got into an
9:56
accident on the way to the studio, right?
9:58
And even though it, well, let's say it was my fault for making
10:02
this accident, nobody would know about it.
10:03
Nobody would even care, right? No.
10:05
And it was like, of course there's an accident today.
10:07
There's one every day, you know. Exactly.
10:08
No. Let's say I was taking an
10:10
autonomous Waymo on my way to the studio and the Waymo got
10:13
into an accident. This would make the national
10:16
news for sure, right? And I guess my point here a
10:19
little bit is that there is definitely an unfair
10:22
discrimination for the autonomous vehicles.
10:25
They have to be perfect, even though people are not perfect.
10:28
We expect the autonomous vehicles to be perfect because
10:30
of how novel the technology is and because of, I guess,
10:35
mistrust in the technology may be partially because of all the
10:38
Terminator movies we've seen. Right Are they say?
10:41
Are autonomous vehicles safer than human drivers?
10:45
Absolutely. It depends, I guess on the
10:48
situation itself. So for example, if you've had a
10:52
pack of 12 beers, you are not a very safe driver.
10:56
At that point, most autonomous vehicles in the market would be
10:58
safer than you are. If you've been driving a heavy
11:02
duty commercial truck for the past 18 hours without taking a
11:06
break, you're extremely exhausted.
11:08
At this point, even though you could be driving, you are not
11:11
paying the attention you should be paying.
11:13
And at this point, any autonomous vehicle is safer than
11:15
you are. But if you are a professionally
11:21
skilled driver who is fresh, who has been on the road for less
11:25
than half an hour, I mean. Like to think I'm pretty good
11:27
driver, you know? Chances are you'll be safer than
11:30
some of the vehicles out there because you can anticipate and
11:34
handle certain scenarios based on your experience and based on
11:39
your level of freshness better than some autonomous vehicles
11:42
can. Like what would be an example of
11:44
a scenario that I might handle better than AI?
11:47
So let's say you're driving through a busy city and
11:52
approximately 50 feet ahead of you, you see kids who are riding
11:59
bicycles or playing in the park, right?
12:01
They are in a nearby park. They are not on the street.
12:04
They're not trying to cross the road yet, but as you see them,
12:09
there is something that clicks in your head.
12:11
They're like, what if this guy's decides suddenly to cross the
12:15
street? Ball rolls on the street or
12:17
something, right? And you start slowing down or
12:19
you start paying more attention subconsciously even as you
12:23
approach it in the section, AI unfortunately still is learning
12:28
to anticipate those scenarios. Another perfect example for
12:32
autonomy that would fail is let's say you're going through a
12:38
street and you see a puddle. As you see that puddle, you're
12:43
going to ask yourself as to why is that puddle there?
12:46
You might think, OK, you know what?
12:47
It rained last night and it rained a lot, so there should be
12:50
puddles, it's normal. But if it didn't rain last
12:52
night, you would be asking yourself, why is that puddle
12:55
there? And you would slow down a lot as
12:58
you approach this puddle because you would be expecting the
13:01
chance of a pothole underneath that puddle.
13:04
So you would think there might be some huge kind of mess there
13:08
and you would slow down and you would either avoid it or you
13:11
would drive through it at very slow speed.
13:14
An autonomous vehicle that doesn't have this intuition
13:17
built yet, it might say, oh, puddle, that's OK, I'm used to
13:20
puddles go at full speed, right? Then it will probably destroy
13:23
your car that way. Right.
13:24
Exactly. So basically all autonomous,
13:27
does the technology kind of work the same?
13:29
And I want to move, I know we talked about kind of cars, maybe
13:31
trucks on the road, highway like semi conductors, semi trucks.
13:37
Is that the same technology that you would also see with the
13:39
tractors and agriculture or with the, you know, the mower, in
13:43
your case the zero term mower, Is it relatively the same
13:46
technology in terms of how it functions?
13:49
Yeah. So the autonomous safety aspect
13:53
is similar, meaning both use cameras, laser scanners,
13:59
Raiders, you name it. They both use similar safety
14:02
sensors for protecting, for detecting obstacles and
14:07
preventing collisions. The way they navigate is
14:11
substantially different. So when you are dealing with the
14:16
CD applications for autonomy or on road vehicles, they usually
14:19
localize against things and that's how they used to
14:23
navigate. They localize themselves against
14:26
buildings, against fire hydrants, against you name it.
14:32
An autonomous agricultural tractor on autonomous lawnmower
14:35
usually has nothing to localize against.
14:37
They're in the middle of a field.
14:38
They're just the middle of a field, so you have to give them
14:41
something to localize against. And for autonomous tractors and
14:45
for a fully autonomous hero solution that something is
14:49
usually GPS satellites which are located in space, which is kind
14:53
of crazy to think about if we are localizing in satellites
14:55
that are so far away in space. But that would give you you
14:58
centimeter level position of the of the absolute geospatial
15:05
position of your autonomous equipment.
15:08
For semi autonomous cytic solution this is different
15:12
because in a cytic solution we have a human operator that might
15:15
be operating one of the machines manually and then you have other
15:19
autonomous machines beside it. So those other autonomous
15:23
machines can localize against the human operator.
15:27
Which gives you a huge advantage in applications like deep shade
15:30
whenever you cannot see the satellites in applications like
15:35
rural environments where you just simply don't have cellular
15:40
connection and cannot get the proper signals from the
15:43
satellites. So from a safety perspective, if
15:46
one of the mowers, if there's a choice between like I'm going to
15:49
wreck the mower or like get near a child and potentially they
15:52
jumped out of nowhere or something, Does the mower know
15:55
to wreck the mower and not go for like, how does you know
15:58
those type of decisions, which I don't think are probably like
16:00
super common, but you know. Yeah.
16:04
And those are actually extremely, extremely difficult
16:07
scenarios to handle, Alex. On the implementation side, it's
16:10
more difficult than you might think.
16:14
And the reason for that is because in our application for
16:17
off road autonomy, usually there is no obstacles around the
16:20
machine. And of course, the decision is
16:23
always to go wreck the machine, especially if there is no driver
16:26
on it, because, you know, it's just.
16:27
A machine. It's just a machine, right?
16:30
So the decision is always that for on road vehicles, the
16:34
decision is way more complex because in an on road vehicle
16:38
there is a lot of obstacles nearby.
16:40
If you are, if you have a child who is crossing the street and
16:45
you decide to save the child's life, then you might be either
16:50
endangering the passenger who is taking a ride in the machine or
16:54
you might be endangering other passengers and other vehicles
16:57
beside you, Right, right. So that decision becomes much
16:59
more complex. How does it make that decision?
17:04
Then, so I am not an expert on social ethics, but some of it's,
17:11
it's crazy to think about that organizations like way MO, they
17:14
do employ people who need to make ethical decisions for the
17:20
organization and have to, I know it gets a little bit crazy,
17:24
right? But they would have to say like,
17:26
OK, what do we know about the person who is crossing the
17:32
sidewalk? What do we know about the person
17:33
who is riding the car? What do we know about the people
17:36
who are around the vehicle? How can we minimize the amount
17:40
of fatalities in the next in the next minute?
17:45
I wonder if it wouldn't look at like value a kids life more than
17:48
like a nine year old crossing the street with a cane or
17:50
something. Yeah, you know.
17:52
I mean, it kind of makes sense. Yeah, it's but and you know,
17:57
like it's, it's kind of crazy to think about this, Alex, because
17:59
we as people, when we are driving and we need to make the
18:02
decision split second, we will ourselves make the decision
18:06
based on a bunch of different factors.
18:08
Sure, based on how much we value our own life compared to the
18:10
life of the person we might be indenturing and so forth.
18:13
Yeah. And now we are outsourcing this
18:16
decision making right to an AI that would be responsible for
18:22
either saving somebody's life at the expense of indenturing
18:27
somebody else's life. It's.
18:28
Kind of like the I robot movie. Have you seen that where he
18:31
chooses that the robot chooses to save.
18:34
I think somebody else Will Smith versus like a child or something
18:37
like that. Because the risk of or the, the
18:39
calculation the AI robot did in its mind was that there was like
18:42
an 85% chance he could save and there was like a 5 or something
18:45
percent or the percentage are probably off, but it was like a
18:47
way higher percentage chance. The robot said I could probably
18:50
save you. There's a super low chance I'm
18:51
going to save this other person. But from a human standpoint, it
18:55
was like save the other person, you know?
18:57
And so because we we don't think as analytical as that, you know,
19:02
we kind of think about more of an emotional decision, right?
19:05
Exactly, Alexandra. Yeah.
19:08
And they think that this topic of ethics in autonomous driving
19:12
would continue to evolve for quite some time because as we
19:16
expand the utilization of autonomous vehicles, I know it
19:21
sounds terrifying and unfortunate to think, but we're
19:24
going to see more accidents. There is just no way to avoid
19:27
them. We will see more accidents.
19:29
And as those accidents happen, we will need to revisit the
19:33
ethics around how do we preserve human life and how do we make
19:38
sure that we minimize the damage to property and so on.
19:41
Yeah. What's the craziest thing that
19:43
you've seen happen with the mowers in terms of just like
19:47
maybe they ran into something or who knows, just kind of like
19:50
have you seen anything wild with those or?
19:52
Yeah. You know, the craziest thing
19:54
I've seen with our machines is not actually a real accident,
20:00
but coming close enough to an accident.
20:04
More specifically, we were doing testing operations one time in a
20:10
park in Naperville. And as we are testing things, we
20:14
see a child maybe age of 10 or so.
20:17
Enter the field with a bicycle and out of nowhere he starts
20:22
riding his bicycle right across the autonomous moment.
20:27
And as I saw him do that, I, I was in such a shock, I didn't
20:32
even know what to do or what to say.
20:34
The machine stopped. Really.
20:35
The machine detected the bicycle coming.
20:37
It detected the child coming and it stopped.
20:40
And at that moment, what I'm thinking to myself is like, wow,
20:44
he came within a few feet of that machine really and
20:48
intentionally so. He knew exactly what he was
20:51
doing. Oh, shit.
20:53
The society thinks that those machines that essentially
20:58
anything that's been deployed in public is extremely, extremely
21:01
safe, is extremely, extremely trustworthy, right?
21:04
And it's kind of safe to play chicken with it.
21:06
Jeez. And as you can imagine, for the
21:09
engineers and also run those companies, it's a nightmare.
21:12
Yes, it's a scenario because if the child makes one mistake and
21:18
happens to come too close to that moment, right?
21:19
The mower is a machine that moves at 10 miles an hour with
21:23
2000 lbs of weight on it, right? Surviving collision with that is
21:27
not pretty. Yeah, the same way you wouldn't
21:29
do that with the human driving either.
21:31
You just don't put yourself in harm's way on purpose.
21:33
You know exactly, Alex. You know, and I keep recalling
21:37
as I think about the company, I keep recalling the situation we
21:40
had. We were doing a demo a few years
21:43
ago. It was, we were at the very
21:44
early stages of autonomy. Yeah.
21:46
And we were doing a demo with a large landscaping company and
21:50
one of the feet managers, he asks me, well, what happens if
21:54
there is a child who suddenly goes and lays in front of the
21:57
mower? And I thought the question was a
22:01
little bit unusual. I asked him, well, how often
22:03
doesn't happen to you guys that you're mowing in the grass and
22:05
suddenly you see a child coming and laying in front of the.
22:07
Mower. Probably never.
22:08
And he says this never happens because shouldn't, wouldn't do
22:12
that because we have operators who are riding the machines, but
22:15
you should expect it for your machine because it's autonomous.
22:20
And I'm like, wow, you have a good point.
22:22
And it was a couple years later when we seen those kids riding
22:26
buns. You saw it.
22:26
You're like, OK. Yeah, like, wow, he was not
22:28
wrong, right? People will try to interfere
22:31
with autonomous equipment simply because it's autonomous, simply
22:35
because it's no just. To test it right.
22:37
Just to test it. Out see what's going to happen.
22:39
Let me just push the boundaries a little bit.
22:41
Exactly. And this is humans and I mean,
22:43
and this is kind of almost like a nightmare scenario for the
22:46
engineers to handle. Yeah, for the operators
22:49
operating those machines and of course for the company
22:52
management. And this is partially why we
22:56
decided to go down the path of introducing our semi autonomous
23:00
psychic product, because when people see an operator as part
23:04
of the configuration, yeah, they immediately know, oh, OK,
23:07
somebody's watching it. There's no point in the fearing
23:09
because he's just going to stop everything and come down.
23:11
With us, that's so crazy, man. So how long have you guys been,
23:15
I assume at this point there's you've got the product out there
23:18
in the marketplace, right? People like landscapers and
23:20
things that companies are using it, right?
23:21
You've been open for years. What's the what's been the
23:24
reaction to customer from a customer standpoint and from the
23:27
public and just getting the product out there now, what are
23:30
people saying about it? How are they reacting?
23:32
How is it maybe changing their businesses?
23:34
Great question, Alex. And just to kind of emphasize
23:36
here, we are at what you call the early adopter stage, I would
23:41
even say the ultra early adopter stage where the people who are
23:46
going and buying these products and deploying them, they are
23:50
almost, they are probably as excited about this technology as
23:53
you are. Yeah, yeah.
23:55
So, so they'll they love to see the progress, they love to see
23:57
it happen. They'll have to take advantage
24:00
of it and so forth. With that being said, we as a
24:04
company are showing real profitability improvements to
24:07
those who adopted. They're measurable in terms of
24:12
how much more efficient they are operating.
24:15
It's also measurable of how much quality is being improved in
24:18
their organization for the straightest of the lines that
24:22
produce and the quality of the cut.
24:26
So it can cut better than a human.
24:28
Exactly. So for the fully autonomous
24:30
models, they can cut more precisely than any human can,
24:33
right? I tried to cut this precisely.
24:35
Yeah, it's just impossible. It's just tough because I mean,
24:37
they just, they use GPS and they make something around 50
24:44
corrections per second, which is very, very hard for a human to
24:47
do. Even if you're a fast human,
24:48
it's tough for you to do. It plus we also like readjust
24:50
and like, we're just human. We just do different, you know,
24:53
like, yeah. We get destructed.
24:55
Right, we look at something else.
24:56
Yeah. We have flies trying to fly in
24:59
our face, right, if you name it, right.
25:01
So for the autonomy it, it does make sense, right.
25:05
One of the difficulties with fully autonomous deployments in
25:08
our space is path planning. I should say that's the most
25:12
difficult aspect of deploying autonomous equipment for that
25:17
operates over the road is the need to generate path plans
25:21
because path plans could adjust over time.
25:25
So for example, AS3 branches grow over time.
25:28
Or if kids visit the site and they leave the scooters and toys
25:32
on the grass, right? It requires detection of
25:35
obstacles, adjustment of the path plans and so forth.
25:38
And the cool thing about the use of our psychic solution is you
25:43
do not need to create path plans so you can come to a site, You
25:46
put them always beside you and they just go without the need to
25:50
think about how are you going to plan the path, because you just
25:53
plan it on the fly as you go. So how does it work though?
25:56
Because if you're going, I like, I'd be curious like the fully
25:58
autonomous and then what going with somebody because like, if
26:01
you're both going, like, how do you know that they're not going
26:03
to go left and you're not they're not going to go right,
26:05
you know, like it? Do you have to plan out that?
26:08
Is that what that is like? Plan out the yard and like give
26:10
it the you know, the map and stuff like that or how does that
26:12
work? So your question, Alex, is about
26:14
the cutting pattern of a psychic clock, right?
26:18
So many people ask us that question just because the
26:21
product is so new and so unique. We do have an old website, a
26:24
very good explanation of how the cutting patterns actually looks
26:27
for psychic. But the key to emphasize is
26:31
think of it as the driverless models generating A runway for
26:35
the manually operated models. So the human operator almost
26:40
sees like a runway in front of them that they can just cut
26:43
through. They don't have to think about
26:45
anything else. It's just like, OK, I'm going to
26:47
get into that runway. I'm going to cut it.
26:48
I'm going to get to the end of the line, turn the machines
26:50
around the all the machines of the flock turn around and then
26:54
the human operator gets into the next runway.
26:56
But. What about the Foley autonomous
26:59
one? Do you just, is it as easy as
27:01
like you back it out of the ramp off the truck and then just all
27:05
right, go or how would that work?
27:07
So this is the dream, Alex. You know, my dream is that one
27:11
day you can just take a mower is supposed to be autonomous and
27:16
you put it on the grass and it just magically knows where to
27:20
mow, when not to mow, how to plan its path and so on.
27:25
Best that's. What you're working towards?
27:26
Right, best knowledge would almost be like robotic vacuum
27:28
cleaners because. Like a room or something.
27:30
Yeah, exactly. Because nowadays they can mop at
27:33
your house, right? The mop, the see your walls, the
27:35
mop your house, and they just randomly figure out how to
27:38
correctly maintain your space. Today, it's very different for
27:42
commercial landscaping applications because there is
27:45
often problems like there is no clear boundary between the space
27:50
you're supposed to mow and the space you're not supposed to
27:52
mow, like somebody's backyard, for example, right?
27:54
And it's just a lot of more edge cases to handle as to how you
27:59
set it up. And that's the reason we have to
28:01
generate path plans, which are essentially definition of how
28:04
far the MOW can go in each direction, where it's going to
28:07
turn and so forth. And so you said customers are
28:09
what saving money, like what's kind of the business impact from
28:12
doing this? Are they not?
28:13
Are they, hey, we were going to hire three people, now maybe we
28:16
just hire one or two because we've got these mowers that kind
28:19
of can like double up on the work.
28:21
Is that how it's been or what's been the the feedback?
28:24
Alex Yeah. So the typical use cases for the
28:29
fully autonomous, for example, is that if you had a crew of
28:32
three people, it can now easily become a crew of two people, or
28:36
even the one person crew who operates multiple autonomous
28:39
machines, right? So instead of you coming to the
28:42
site where one person is doing the open fields and another
28:46
person is doing tight areas and trimming, you now have one
28:50
person who just deploys the mower.
28:52
The mower can do the open fields and in the meantime the human
28:55
does the tight areas, the trimming around the site and so
28:58
forth. With the semi autonomous psychic
29:01
product, this has been even more interesting because now you have
29:07
a human who is operating a mower at a typical speed that we
29:12
usually operate at and you have another driverless mower that
29:16
mimics the motion at the same speed.
29:19
So essentially they breakdown complex, they breakdown complex
29:22
big sites and reduce the time by almost half on how long it takes
29:27
to do it. We are seeing some excellent
29:31
production rates from our customer runs something on the
29:36
order of three acres per hour per more for a 74 inch machine.
29:41
So as you can imagine, they go to a site that has 6 acres to
29:45
mow, they cut that site in one hour with one human, and they go
29:50
to the next site. The it's almost like hiring on
29:56
another human without actually hiring the human, right?
29:59
Right. You just use the technology to
30:01
do the job that the human would need to do.
30:04
Yeah, wow. It's kind of crazy.
30:06
So if I'm listening to this, I might be like, this is all
30:08
exciting stuff, but hot like this is not a business that just
30:11
anyone could get into randomly like you.
30:14
What's the reason that you got into autonomous?
30:16
And then even to take it a step further, like why lawn care, you
30:19
know, but I'm curious, what was your path to kind of start this
30:21
business? Alex, that's an excellent
30:24
question. So I love technology and I've
30:29
been in the development of software for autonomous vehicles
30:34
for some time before starting the company.
30:37
So I had a little bit of knowledge of things that
30:39
geospatial mapping, even knowledge of how engines
30:42
operates and how electrical vehicles operate.
30:47
And then I decided that I am going to look into starting my
30:52
own company that's somehow related to that space, right?
30:57
And then back then, I was living in South Naperville and I just
31:00
bought a house that has a huge yard.
31:04
I thought I got an awesome deal for that house.
31:05
I'm like, the house comes with a huge backyard.
31:09
Great. It was my first home.
31:11
I was an inexperienced homeowner and I thought the bigger the
31:14
yard, the better. After the first time I mold the
31:17
grass in that yard, I was like, Oh my gosh, what did I just do?
31:20
This thing is going to take my whole weekend to maintain.
31:24
And this is kind of when I started looking into that
31:26
market. It was back in 2013 and I began
31:28
researching the market of autonomous mowing and they've
31:32
been following that market for some time until one day I
31:35
decided this is it. Was there much in that space?
31:38
I feel like there's probably not a ton.
31:39
Yeah, you know, so the first autonomous mower idea as an idea
31:43
itself has been patented back in the 1980s.
31:46
Really. People started looking into that
31:49
back then. By the early 2000s, so this is
31:53
25 years ago, there were quite a few models of mowers that
31:58
navigate randomly. They navigate in the pseudo
32:00
random motion inside the perimeter wire, which is kind of
32:04
like similar to like a dog fence, right?
32:06
So you bury a wire and then you put them all there and it just
32:08
bounces back and forth until it somehow molds your yard.
32:11
Yeah, OK. And then right now this next
32:16
evolution of technology that we are seeing is autonomous
32:19
machines like autonomous, how small was that navigate using
32:22
GPS or using some kind of predefined right?
32:25
Right, much more complex, yeah. So you would you build like a
32:29
prototype for yourself basically and then launch that like take
32:32
that to launch the business? Yeah, You know, I started
32:34
building a prototype for myself. This this was a company I
32:38
needed. We started out of my garage.
32:39
So you're. Working full time?
32:40
You're starting this as a side hustle or did you leave full
32:43
time to say I'm going to quit my job and go do this full time?
32:46
Yeah, what I did, Alex, is I started as a side hustle until I
32:51
realized that it was way, way too much time consuming and I
32:55
could not and I could not keep my full time job going.
32:58
As every business owner goes through at some point, there's a
33:00
crossroads at some point. Exactly, exactly.
33:03
I just realized it's not going to work as a side hustle if I
33:05
don't do it full time. It was very high risk to jump
33:08
into it full time, yeah, but I decided that I want to do it and
33:13
that the risk is worth it. Were you worried about like, you
33:15
know, financially, did you have to go get investors?
33:18
Because this is not one of those like you know, you like this
33:21
isn't I would think of financially intensive operation,
33:25
right? You need technology.
33:26
You got to build. You're building the mowers
33:28
yourself or you're using mowers and embedding the software
33:32
technology inside the mowers. Do you guys actually make the
33:34
mowers or do you just software that goes in them?
33:36
Exactly, Alex, we do not make the mowers.
33:38
That's what I thought. So we actually not as hardware
33:41
intensive as you. OK, OK.
33:43
So we don't make the mower itself, right?
33:45
We do build the boxes that contain computers and
33:49
controllers, OK. And our core intellectual
33:52
property is the software right that powers the autonomy.
33:56
Got it. But you still gotta go find a
33:58
mower manufacturer, whoever else to put these things on there and
34:01
all that type of stuff, right? Absolutely, yeah.
34:04
Absolutely, yes. So we went through iterations of
34:06
actually trying it with smaller machines and putting into larger
34:10
machines. So I was.
34:11
Thinking about your customer base, like, yeah, you serve like
34:13
technically lawn care providers, but like, isn't the customer the
34:16
lawnmower companies? Because aren't you really trying
34:18
to sell to them? And then they sell direct to
34:20
consumers. Exactly.
34:23
Our customers are public institutions like park
34:26
districts, OK colleges and commercial landscaping
34:30
companies. So you package the mower up with
34:33
the technology and sell it as one unit.
34:35
Is there a reason you wouldn't go to the manufacturers and just
34:37
sell the technology or the software or do a partnership
34:40
deal there? Is that not what you want?
34:41
To do? Absolutely.
34:42
There is pros and cons to each approach.
34:44
Yeah, it's all about the supply chain dynamics, right, as to how
34:50
much, what percentage of the cut every portion takes, yes, and
34:53
and distribution model of dealerships and and so forth.
34:57
But ultimately, yes, we are ultimately seeing ourselves as a
35:01
technology solution provider that makes the best autonomous
35:06
and semi autonomous technology in the market and that any
35:09
manufacturer, not just normal manufacturers, but in the future
35:13
snow removal machines for skid steers, construction,
35:16
agriculture, you name it, that anybody can integrate that stock
35:20
into their machine and provide real value for their customers.
35:24
How long do you think it'll be before humans don't mow our own
35:28
lawns anymore? Or it's just an autonomous like
35:30
like an everyday. Obviously now it's you start
35:33
more commercially bigger yards, of course it's that makes total
35:36
sense. But at some point, everything
35:38
kind of comes more commercial and mainstream, all, if you had
35:41
to guess in the future where you'd like, you'd walk down the
35:44
street and you'd just see like mowers going.
35:46
Yeah, So I don't think such a day would ever come, Alex,
35:50
really. I don't think there would ever
35:52
be a day where humans don't mow grass.
35:54
The reason for that is not because the technology cannot be
35:57
advanced enough, but because some people just simply enjoy
36:00
lawn mowing. There are people out there
36:02
that's true, who just simply enjoy to cut grass.
36:05
And if you make all lawn mowers autonomous and say you don't
36:08
need to cut grass, they will say no, no, no, I'm going to pay you
36:10
money so I can cut grass because I enjoy doing it.
36:14
So there will always be people who enjoy driving, who like to
36:17
but race and so forth. I see that, but also there's no
36:21
option now to not mow the grass yourself besides paying some
36:24
another human. So I don't know.
36:25
Do you think that changes when there are legitimate safe
36:29
options that are well adapted and affordable?
36:31
At that point, you know, maybe you still got a small percentage
36:35
of people that are like, I'd rather just mow it myself.
36:37
But how long do you, do you think I'd be able to say, don't
36:39
you think 9095% of people are gonna be like, yeah, just mow
36:41
the damn lawn like robot? You know, right, right.
36:45
Also depends on the cost, right. So right, for sure.
36:47
So the ability for robots to replace people is right now it's
36:51
innovative right now it's something that's exciting and or
36:55
adopt as a jumping on it. In the long term, we have to
36:58
show that it's less expensive to have the robot perform the
37:03
specific task, yeah, than to have a human do it right.
37:06
And in other words, when you look at the minimum range, you
37:10
have to say, like, look, it's much more affordable to have a
37:13
robot do this task than to pay a human to do the exact same task
37:17
at a minimum wage. And this would really depend,
37:21
Alex, on the type of sites and complexity of sites.
37:24
Yeah. So for example, I think we are
37:27
still at least five years away from having a package that's
37:31
sufficiently cost efficient that would be able to detect
37:35
everything a human operator can detect on the field.
37:38
So, for example, if you are a commercial landscaper, let's
37:41
just say for a second you decide to quit your job there and
37:43
become a commercial landscape. Sheridan Landscaping, we're in.
37:46
Business. Yeah.
37:47
And then you are driving down the field and you see a glass
37:50
bottle in front of you. Yeah.
37:52
And you're like, oh, OK, I need to slow down my machine and pick
37:56
up the glass bottle, put it in some trash bag on the mower and
37:58
it can keep going. Yeah, it would likely be my
38:05
estimate about five years before commercial mowers that operate
38:09
at 10 miles an hour have the capability to detect and remove
38:13
a glass bottle. And it's not because the
38:16
technology is impossible to make.
38:18
Yeah, but it's not. It's simply because it's not
38:20
cost effective enough. Right, right.
38:22
A lot of things you could do, but it's like, does that, do the
38:25
economics make sense to really replace that at scale?
38:28
And it's like, not really at this point, but yeah, just like
38:31
flat screen TV's used to be like super expensive and like bulky
38:34
and big. And now it's like you might have
38:35
flat screen TV for like 400 bucks or less and like a decent
38:38
size on a name brand 1. You know, that's exactly the the
38:42
cost. And the same with the electric
38:43
cars, right? Like I think Tesla and Lucid and
38:46
a lot of these other ones are the the costs are seemingly
38:49
coming down a little bit. Well, that's that's exactly it.
38:51
And This is why way more is making robotaxis and not selling
38:55
autonomous cars. Yeah, because if they were to
38:57
try to sell the autonomous vehicles, it's only the ultra
39:01
wealthy, ultra rich could afford to buy a car that cost
39:03
$5,000,000, right? But if you're running a robotaxi
39:07
business model, right, then you can keep the inventory in house,
39:10
you can keep ownership of those assets and just lease them out
39:14
essentially, which allows you to monetize on the usage on a very
39:19
high usage of them over time. Right.
39:22
So if cars, trucks and mowers even are being driven
39:25
autonomously, what does that mean for jobs for humans?
39:31
I think it's a good evolution. It's we as people.
39:35
We were meant to innovate, yes, we were meant to make our life
39:38
easier. We were meant to become more
39:40
efficient. Yeah, it's kind of what we do.
39:42
It's kind of what we do, right? I mean, if you look at what has
39:44
happened for the past thousand years of human evolution, right?
39:48
Yeah. We went from using horses to
39:51
building cars. Yeah, right.
39:53
A lot of jobs which involved taking care of horses, grooming
39:56
them or removing the poop. Yeah, right.
39:58
Those jobs have been eliminated. Yeah, because they're no longer
40:02
needed because we use cars. As a result of that, new, many
40:06
new jobs were created for the car industry.
40:09
When you look at autonomous mowing, it's just another
40:12
innovation. It could be a very big
40:14
innovation, but it is just another innovation, meaning we
40:19
could have had a lot more workers working in landscaping.
40:22
If we stuck to if we made a regulation, for example, that
40:26
lawn mowers should not exceed 26 inch in cutting deck width in
40:31
size, all mowers should be within 26 inch.
40:34
And that way you have a huge, huge industry that needs to be
40:38
lawn mowing, right? Because big parks would require
40:41
10 people to show up at any time with small mowers and to condom
40:45
the invention of zero turn commercial mowers took a lot of
40:48
jobs away. I don't think it's a bad thing.
40:52
Yeah, we made the industry much more efficient.
40:54
One human can now cut has the same production rate as ten
40:59
small push modes. What autonomous modes do is they
41:03
take it to the next stage where we're saying one human now can
41:07
supervise 2/3 even for autonomous machines cutting the
41:11
grass at the same time. Yeah.
41:13
And yes, it reduces the number of jobs available in
41:16
landscaping, but it increases the number of jobs that are
41:19
needed for the software programmers to develop those
41:23
things. It increases the number of jobs
41:25
of mechanics who will be building these autonomous
41:27
systems. It helps companies who run it
41:31
become more efficient and elevates their people to the
41:35
next level of of responsibility. So it is almost like a career
41:42
path for the companies that are adopting this technology.
41:45
Yeah. So in other words, if you were
41:47
running a landscaping company, instead of telling your
41:49
employees, hey, you have no career path at all, you will be
41:53
cutting grass in this exact way for the next 40 years, right?
41:57
Instead of you saying that, yeah, you can tell them, look,
41:59
you have a career path in front of you.
42:01
You are no longer a blue collar laborer.
42:04
You're now a supervisor of autonomous equipment.
42:07
Yeah, you can get paid more. Your job is is going to require
42:12
a little bit more thinking sometimes, but you would now be
42:16
able to have the efficiency of two or three people.
42:20
Makes sense. How would you describe the
42:22
relationship between autonomous and AI?
42:26
Just so people understand like. Yeah, I think they're very
42:31
closely tied together, right? I wouldn't say they're one to
42:34
one. Correlation, right.
42:35
Yeah, no. AI is a much broader field that
42:39
has everything from natural language processing right to
42:46
object recognition. Yeah, I think the biggest level
42:50
up there is around image processing.
42:54
Yeah, right. So it's about looking at the
42:58
environment around you, understanding that environment,
43:02
correlating it to things you've already seen and making
43:06
decisions based on that, right. So more like a bigger data
43:09
science field? The way I've always thought
43:11
about it was AI is needed to be in autonomous things, otherwise
43:16
it's kind of hard to read and predict and analyze data, right?
43:19
But AI obviously, to your point, is a huge field.
43:22
I think the jobs thing makes sense for what you guys like.
43:25
You know, I see, you know, there's not as many taxi drivers
43:28
or Uber drivers or lawn mowers, right?
43:31
And that's probably a good thing because it allows us to do other
43:33
things. It's kind of replacing a
43:35
physical thing that we do. Where I think it gets tricky,
43:38
though, is when AI replaces a lot of the mind jobs where you
43:42
needed to be intellectual. And it was, it was not labor per
43:46
SE or a physical thing. It's like the industrial
43:48
revolution. Everyone's like, oh, all those
43:50
jobs got eliminated, but the new ones came up.
43:52
But it's like, yeah, but that replaced like, instead of 50
43:55
people plowing a field, like there was a tractor or two.
43:58
And then those people went to the city and got jobs or worked
44:00
on the tractors or designed them or whatever it is.
44:03
I'm curious to get your take on how you see AI disrupting
44:07
employment because for your stuff it makes sense.
44:10
But what about the the mind jobs, as I would call them like
44:13
that, where now it's not just a eyes taking over or eliminating
44:18
a lot of the physical labor jobs, but now it's like you're
44:21
an analyst, you're a lawyer, you're a marketer.
44:24
You know what I mean? There's a lot of jobs out there
44:26
like that that are going to be disrupted.
44:28
Yeah, Alex, absolutely. So the bigger question there is
44:32
ethical, right? It's on the high level
44:34
philosophy of what we as humans would do once our jobs have been
44:38
eliminated. Yeah.
44:40
And I feel like I may not be the most qualified to give you the
44:45
answer on what would happen, but it would be more along the lines
44:50
of like, hey, I tried to replace myself many times.
44:54
I tried to have AI replace me on daily basis.
44:57
Yeah, I automate what I can. Yeah.
44:58
So that I don't have to work as much.
45:01
Yeah. And look, if one day you can
45:04
come to me and say, hey, yeah, look, I developed an AI that can
45:07
run your company for you, which means that you can go on
45:10
vacation for a month and it's gonna be just fine.
45:12
Yeah. I'll be like, cuz this is a
45:14
genius invention. Thank you so much.
45:16
I'm gonna go on vacation. Yeah.
45:17
So the question. You'd wanna go build stuff
45:19
though you'd be tired of after a week.
45:20
You'd be like, I wanna start building.
45:22
Stuff that's right for me like two or three days ago, yeah,
45:24
yeah, I'm like, OK, I'm gonna need to get back to the
45:26
building. But yeah.
45:27
So the question becomes a little bit more philosophically is if
45:30
you were to replace ourselves, if we could replace our jobs,
45:35
both the manual labor jobs and the intellectual jobs with AI,
45:40
then what will the government do to make sure that people don't
45:46
just all become homeless and starve because they don't get
45:48
paid? In other words, it's a little
45:51
bit more around politics of making sure that everybody has
45:55
some basic living standard regardless of their job has been
45:58
replaced with AI or not. Yeah, the universal general
46:02
income I think is what they're calling it or something, right?
46:05
Because yeah, I do think it's not going to be like AI comes in
46:08
and then replaces this job, but then one pop, one job pops up
46:12
and replacement of that. And it's like, I don't think
46:14
it's a one for one game. I think there's going to be over
46:17
the next 5-10 years, whatever it is, there will be some, some
46:20
effects from this, you know, and, and overall, I think we'll
46:23
be fine and we'll evolve like we always do, right?
46:25
There's always been these, you know, moments in history.
46:28
We're humans, we're resilient, we build shit, we figure it out.
46:30
You would think, right? But they're certainly going to
46:33
have to, there's certainly going to be some, some issues with
46:37
income, with even purpose. Even if you give someone a
46:39
check, is that, are they happy? Are they just about to collect
46:43
the check now and sit at home? Like, what do I do with my life?
46:45
You know, there's, that's more philosophical, right?
46:48
Or political or whatever, however you want to call it.
46:50
But it's interesting. Exactly, Alex.
46:52
Yeah. Those are some of the most
46:53
difficult questions for society as we adopt and embrace AI.
46:59
And I feel like it would take time.
47:01
Yeah. I mean, we just had a discussion
47:03
about Waymo, and as much as I love autonomous vehicles, I just
47:08
want to remind you that 99%, way more than 99% of vehicles on the
47:12
road today are still manually driven.
47:14
Yes, yes. They're still not way more than
47:17
they're not Tesla's or anything else.
47:18
Right. So with that being said, it
47:21
would take time for us to get there.
47:23
It's not going to happen in one day.
47:25
Yeah. And by the time we get there, I
47:27
think it's pretty important for us to think of, yes, like, once
47:31
people's jobs have been replaced, what are they gonna be
47:35
doing on daily basis that would help us create a better world
47:39
for the future? Yeah.
47:40
Right. Yeah, I know.
47:42
If I was analyzing data every day and that was my job, and I
47:45
was crunching numbers and putting things on, I'd be
47:47
nervous. I'd be thinking, well, I
47:49
wouldn't be there. I would just be thinking, how do
47:51
I start using AI to replace myself?
47:54
What type of job, career paths can I learn?
47:56
What type of skills can I Start learning now?
47:58
I think the danger is if you just sit back and say, it
48:01
doesn't affect my job today. And then one day the company
48:04
comes in and goes, hey, we've got this awesome technology.
48:07
We don't need 100 of you. We need 10 or five.
48:11
And it's like, oh shit, I wasn't thinking about this.
48:13
So it's like you can't just bury your head in the sand and like
48:15
hope it's not going to affect you.
48:17
I mean, we use AI just like you. I use it every single day.
48:20
Like we're going to use AI had to chop up this video content.
48:23
It'll detect who's talking or what's happening and it'll go to
48:26
that camera, right? I mean, that saves an hour or
48:28
two of an editor going through and manually doing it, you know
48:31
what I mean? So all types of stuff.
48:33
But yeah, it's going to be, it's going to be a wild decade.
48:37
I feel like, you know, like it's just, there's a lot of change
48:41
that's coming. And my advice, just like you
48:43
said, is like be adaptable, be try to put yourself out of the
48:47
out of business. If you can figure out a way to
48:49
use it, get aggressive with it, understand it, use it every
48:52
single day. And then I think leverage the
48:54
other side of that is leverage the non AI stuff like
48:58
relationship building, maybe some networking, emotional
49:01
intelligence, learn how to lead and coach and develop people and
49:04
teams because I think those things are going to be more
49:07
important than ever. Exactly.
49:09
Alex and I have to admit that not all societies are structured
49:13
the same way. But in the United States, we are
49:17
driven by profitability of companies.
49:19
Whether you like it or not, this is how the society operates
49:22
today. Which means that if you work for
49:25
a company where you are a bottleneck on them being able to
49:29
generate profitability because you don't want your job to be
49:32
replaced, you have to realize that it's only a matter of time
49:35
before your job will be replaced.
49:37
So instead of looking at an AI as an enemy that's trying to
49:40
take my job away, you can think of AI as your best friend that
49:45
makes your job easier, that allows you to do more for your
49:50
company while receiving slight higher compensation.
49:54
Yes. Right.
49:55
That's right. And that's the moment you switch
49:58
IT mentality, the moment you understand that the adoption of
50:01
technology in general and the adoption of AI is not there to
50:05
replace you, but to make you more efficient.
50:08
That's the moment you feel that, yeah, my job security is
50:12
guaranteed forever. Because I'm going to continue to
50:15
look at the latest trends of technology.
50:17
I'm going to continue to adopt the most efficient trends and
50:20
I'm going to generate even more value for my company, which
50:22
would make my company more profitable.
50:25
In terms of AI in general, are you optimistic about AI?
50:29
And as a result, like just from a human civilization standpoint,
50:33
you know, if you flash forward 102 hundred years, are you all,
50:36
are you at all worried that AI could potentially take over
50:40
humans at some point? Or are you?
50:41
Do you just not go that route at all?
50:45
Yeah, You know, I think that AI still has a while to go to catch
50:50
up to human intellect in some areas.
50:52
Yeah. And in other areas, it already
50:54
exceeded human capabilities a long time ago, for sure.
50:57
Perfect example I think of is chess.
50:59
Oh yeah, right. When computers first started
51:02
playing chess, people thought that it would be impossible for
51:05
a computer to defeat the world champion in chess.
51:07
Yeah, until, until that happened, yes, right now it's.
51:12
Impossible to beat a robot. And once it happened, right?
51:14
Yeah, Nobody has doubts anymore. The computer can play chess
51:17
better than people can, right? Because it's a well structured
51:20
task that can be well defined. Yeah.
51:22
And this there is no competition.
51:24
And play out every scenario humanly possible in a fraction
51:27
of a second. Exactly.
51:28
I mean, what are you going to do with that?
51:30
Exactly. It's, it's just a way, way more
51:32
advanced than the human brain can ever process.
51:35
But when you go to other tasks which require a lot of human
51:39
intellect, a lot of human thinking, even something as
51:42
simple as lawn mowing, I think we're still many, many years
51:45
away before AI can do the task of lawn mowing as best as the,
51:51
as the most experienced human operator.
51:54
And that just simply because of everything that can go wrong,
51:56
from flooded sites, the glass bottles being in the field, to
52:00
anticipating the kid who is hiding behind the tree so he can
52:03
ride his bicycle across the field.
52:05
Yeah, right. But you remember the like
52:07
watching the Terminator movies, right?
52:09
Sure. And remember, like, I don't, you
52:11
know, decades ago when those came out, it was just like you
52:14
would watch it and be like, that's so like crazy, you know
52:17
what I mean? Like there's not fucking robots
52:18
that are going to like, we're going to have to go to war with
52:20
robots. And I'm not saying that's going
52:21
to happen. But as you think about it today,
52:24
though, it just like it's a little more plausible.
52:26
It's a little bit more like, yeah, I mean, there's a version
52:29
of where we could go down that path where the AI could get so
52:33
good, fifty 100, whatever it is. And we're not there today,
52:36
obviously. But at some point it keeps
52:37
getting better and better and better.
52:39
And the AI just kind of looks around.
52:40
It's like, why do we need you guys again?
52:43
Yeah, I don't know if you're needed.
52:45
I don't know if this is the most efficient way to run the planet.
52:48
And so I don't know, it's, I'm optimistic and overall, but I
52:52
think it's interesting. You'd be almost lying to
52:54
yourself if you didn't think there's a possibility it could
52:56
go that way. Or it gets in the hands of the
52:58
wrong people and there's gonna be that too.
53:01
For sure, yes, AI could become a weapon if it gets in the hands
53:05
of people who misuse it for that purpose.
53:07
Right? Anybody who builds AI for
53:11
applications that do good needs to think of having kill
53:15
switches. Yeah, and shuttle switches,
53:17
right? In case things go wrong.
53:19
Because they would go wrong. Yes, they would go wrong at some
53:21
point. Mowers attacking everyone.
53:23
You got to make sure you have a kill switch that's integrated
53:25
with multiple levels of redundancy.
53:27
Yeah. So you have to build redundancy
53:29
in different stages to make sure that if that's if this kill
53:32
switch fails, we still have that one.
53:35
Yeah, Yeah. It's the moment AI starts to
53:40
make bad decisions, Alex. That's the moment that I
53:44
hopefully, hopefully we have built enough protection and
53:48
safety systems not to allow it to happen.
53:50
Right. It's kind of wild to think
53:52
about. I'm gonna share something funny
53:54
with you. So I watched the Terminator
53:56
movies when I was a child. Yeah.
53:57
And I really like those movies. I love them.
54:00
And I didn't really kind of pay attention to the back story of
54:04
how Skynet takes over the world and stuff like that.
54:07
Yeah. And a couple of weeks ago, me
54:10
and my daughter decided to watch Terminator movies.
54:12
My daughter, she's 12 years old. OK, cool.
54:14
She's like, I have an 11 year old and an 8 year old or a nine
54:16
year old now. So yeah.
54:18
And she and I decided we're going to spend Friday evening
54:21
and we'll first watch Terminator One, and then the next Friday
54:24
we're going to watch Terminator Two.
54:25
Yeah. And one of the things I noticed
54:28
about that show, the excuse me, this the movie is the vendor of
54:32
the back story of how the war began with the with the robots.
54:36
And I was surprised to see something very, very small in
54:41
the geospatial contact them is that the robots once they
54:46
decided to rebel against the the US forces, they did not go and
54:51
attack US. They went and they attacked
54:54
Russia because they knew that by attacking Russia, Russia will
54:59
automatically assume that they were attacked by the US.
55:01
Yeah, right. And there you have a nuclear
55:03
catastrophic right. Wow.
55:04
I didn't, so I don't remember that part about it.
55:06
Yeah, it's been a while since I watched the first one.
55:08
Yeah, I didn't even know that. And it's it's it's kind of like,
55:12
yeah, this is an interesting back story that.
55:14
They happened to go attack, Yeah, China, Russia, just blow
55:17
it up. Exactly.
55:18
Just attack them because they would automatically assume that
55:20
they've been attacked by the US and they have a nuclear war,
55:22
right? But wouldn't that also destroy
55:24
the AI if they're just launching nukes though?
55:26
They must have had a plan for that.
55:27
They must have had a plan, yeah. They must have knew exactly
55:29
which targets would be destroyed and so forth.
55:32
Yeah, not the data centers. Keep the data centers.
55:34
Alive keep the data centers alive exactly so it's kind of
55:37
like a as we build the AI it's also important to some extent
55:40
right because we're trying to compete with China on AI and
55:42
when we try to compete with Russia you.
55:44
Have to, otherwise we're really screwed if we won't go
55:46
aggressive with that, yeah. Exactly.
55:47
And at the same time, you can almost think of it as being
55:50
like, OK, what kind of security protocols we have in place that
55:54
if the AI in one of the countries goes bad, Yeah, that
55:57
we know it's the AI that's going bad and other people.
56:00
Right. Oh.
56:00
Man, that's a great point. Wow, that's.
56:02
Just a lot to it, yeah. How do you know if it's actually
56:05
the the country attack and you are some malfunction or AI or
56:08
some malfunction? Yeah, wow, Yeah, man, that's
56:11
crazy. So for your business, what's
56:15
kind of the vision? I mean, what do you want to have
56:17
happen over the next 5 to 10 years?
56:18
I know obviously you want to advance the technology and you
56:21
know bring in more customers and but what's kind of the end game
56:23
for you if there is one? Yeah.
56:26
I mean, I'm excited to see the company grow.
56:29
We are heavily focusing on expanding our psychic
56:32
deployments because of the efficiency they offer and the
56:35
safety they offer. It's kind of like the best of
56:37
both worlds is you have AI, but it's human supervised AI where
56:41
the human has access to an enabling switch and can shut
56:46
everything down in a split of a second.
56:48
So that's that's one of the things we're looking at heavily.
56:51
And then we are expanding into hopefully into more complicated
56:56
sites, smaller sites with deep shade, a lot of trees.
57:02
And then we are going from there into additional applications
57:04
such as potentially snow removal, agriculture
57:07
applications, fertilizing, you name it.
57:11
That's really cool. What's what's been the most
57:13
challenging part about growing a business like this?
57:17
It's been like a low point where you're like, man, I don't know
57:19
if this is gonna work. Yeah.
57:20
You know, for a company like ours that's building where it's
57:25
core business is building relatively complex technology.
57:27
Yeah, it's hard. When the technology doesn't
57:30
work, it's hard, right? Yeah, I could see that.
57:32
Some days you're like, Oh my gosh, so much effort has been
57:35
invested into it and it just doesn't work right.
57:39
And the worst is when you cannot explain why it doesn't work
57:43
right. And then you start blaming the
57:45
AI because it's trying to do something there, right?
57:47
Yeah. But yeah, that's been, I'm sure
57:51
you have such days in every company that builds anything for
57:54
sure. But for a company that's
57:56
developing something new and when things don't work and you
58:02
cannot explain why they don't work, that's usually one of the
58:06
lowest. Points, right?
58:07
Were there ever moments where you thought about shutting down
58:09
the business or getting out of it?
58:11
Yeah, sometimes. What were those moments like
58:14
when that technology doesn't work?
58:16
I mean all, all kind of reasons. I mean, yeah, it just most days
58:22
I'm very excited about building this company and doing a
58:24
business. But just like any other
58:27
entrepreneur who is, are kind of constrained by what they can do
58:32
financially, constrained by what they can do, by how much time
58:36
they spend with the children, right?
58:38
Yeah. You, you keep thinking about a
58:40
lot of things, about a lot of scenarios and how they could
58:43
play out. But ultimately, the way it's
58:47
looking so far, Alex, I am super excited to grow it.
58:50
And who knows, maybe make it a company that operates across
58:55
multiple countries at some point.
58:58
Yeah. Is it, do you have a team,
59:00
investors like what's the team look like now?
59:03
Yeah, we are a team of four people full time.
59:06
Yeah, a couple people part time. We do have some investors.
59:10
We are definitely gonna be looking to raise more as we have
59:15
attraction and expanding both sales and deliveries and yeah,
59:20
excited to be to see what comes out of it.
59:22
That's cool. If you were talking to you of
59:25
several years ago when you first started the business, you're
59:27
talking to a brand new entrepreneur.
59:29
What advice would you give yourself as a new entrepreneur
59:32
Day? One that you know now that you
59:33
didn't know then? You learned a lot, as we do
59:38
every year. It's like, man, you feel like
59:39
you just went through another MBA or something, you know?
59:42
Yeah, that's right. And that's a good question
59:44
because I actually don't come from the business world, right?
59:47
This is the first company I founded.
59:49
Yeah. And I've learned a lot of things
59:51
on the fly, everything from the legal aspects down to sales,
59:55
marketing, you need it, right. So yeah, I would recommend
1:00:01
surround yourself with people who are smarter than you.
1:00:05
Trust more. Dedicate.
1:00:08
Give more responsibilities to others who want to help you
1:00:11
manage them to make sure that they deliver those things.
1:00:15
But the sooner you learn to not do everything yourself, and the
1:00:20
sooner you learn to hire the right people and give them the
1:00:24
tools they need to be successful, the faster you'll be
1:00:27
able to grow your company. That's good advice, really.
1:00:31
I appreciate you coming down here and doing this.
1:00:33
This is a really great conversation.
1:00:35
Where can people go find you, connect with you, learn more
1:00:38
about what you do, and maybe even describe like a good
1:00:40
customer for you guys right now? What would they look like?
1:00:43
Yeah, you can find us on our website, havenshine.com.
1:00:48
Yeah, we have contact information there.
1:00:51
You can reach out to me through the website.
1:00:53
You can also connect with me directly on LinkedIn and
1:00:56
definitely follow the Havenshine, the LinkedIn and
1:00:59
Facebook channels for updates. And what about a good customer
1:01:03
for you guys? If someone's listening right
1:01:05
now, what do they look like? You're asking me like what is
1:01:10
the ideal type of customer? Yeah, just so if someone's
1:01:12
listening, yeah. Yes, if you are a commercial
1:01:15
landscaping company or a park district that maintains large
1:01:20
open land. If you have large open sites
1:01:24
that and you feel like you've always wanted to get in
1:01:27
technology and never knew how, or maybe you're very startled
1:01:29
deploying small robot models and they're not quite doing what you
1:01:33
want them to do, then give us a shout and we'll come up to your
1:01:38
site. We'll evaluate it, and before
1:01:40
you know it, you'll be using the most efficient and safest
1:01:43
technology available in the market.
1:01:44
Love it. Have you talked to Saint Charles
1:01:46
Park District yet? Yeah, I think I spoke to them at
1:01:50
some point. I probably should reconnect with
1:01:52
them. Yeah, we made a lot of progress
1:01:54
since our last conversation. We'll touch based on them.
1:01:57
We'll have some good content to show them from this episode so
1:01:59
you can get that in front of them.
1:02:00
Yeah. Yeah.
1:02:00
Awesome. Cool.
1:02:01
I appreciate you coming on the show, man.
1:02:02
Thank you. Alex, it's been a pleasure.
1:02:04
Thank you. My pleasure.