he built self driving fork lifts

[00:00:00]Imagine if you could have anything you needed [music] the second you knew you needed it. A fleet of autonomous machines grabs the item, places it [music] on a hypersonic highway, and delivered straight to your home without ever touching one of us much slower humans. Once we [music] do that, it means that every single industry can progress 10 times faster. We think we have a better solution, and it all starts with autonomous forklifts.

[00:00:26]>> [music] >> Yo. What's up?

[00:00:34]>> [laughter] >> What's up, man?

[00:00:35]>> Take me inside.

[00:00:36]>> A few years after my [music] grandma died of breast cancer, I was talking to some cancer researchers, and they were telling me, "Okay, every [music] time we need a lab sample, we send it off, it goes through multiple different facilities before it gets there not being processed [music] just from truck to truck, and it takes 2 weeks." I'm like, "Okay, how often would you do it otherwise?"

[00:00:55]>> [music] >> We would maybe do it every day, and it spends about a day in a lab. But, compound that over 20 years or so, >> [music] >> they would have cured it. My grandma could be alive right now if this wasn't a problem. Anytime I do a project, you know, you order everything online, it gets there, [music] you work for a few hours, realize you forgot something, or you're learning for the first time you need something new, and then you basically repeat this process of like ordering, working, waiting [music] until you accomplish something that should have taken you an evening in like 2 weeks. This is a problem for the most important people working on the most important problems. We can cure every disease. We can solve every problem, but we can't do it if we don't have that capability. [music] So, that's what we're working on. So, this is our fully custom-built autonomous forklift. This is the drive motor. This is the steering motor, controllers. This is actually a power switch. Then, we have the the drive and steering controller right here. There's a few parts on here that actually come from other forklifts. We were trying to leverage existing supply chains for this. So, like the forklift assembly itself, the wheels, they actually all come from other forklifts. The main thing that's custom is actually the body itself and the electrical system.

[00:02:04]>> And those are the most important things?

[00:02:06]>> By far. We were originally trying to just modify existing forklifts to turn them into autonomous vehicles, but standard forklifts are almost entirely analog. It's not like a car. [music] They don't come with the sensors and controls that you would expect to see in a car. They're not digitally presented to you. You actually [music] have to build hardware to replicate the sensor values that allow the control in the first place. [music] I didn't want to do a bunch of different cameras, have to use a bunch of different configurations.

[00:02:36]>> [music] >> So, I just did depth camera spam. I put depth cameras everywhere instead of doing like a bunch of mono cameras or something. When it comes to getting as much information as you can possibly get and in a slow-moving environment that is pretty static, [music] I don't have fog in warehouses most of the time, cameras are much better than lidar. To make an extremely reliable autonomous machine, the most important [music] thing is that you actually have all of the data so that you know exactly where every piece [music] of the machine is at every second. We have a custom sensor stack that allows us to know the height of the forks within 10 microns. I was talking [music] to the founder of SRI. They created the world's first teleoperated surgery robot. And he was telling me, "10 microns, you could technically do surgery with that forklift." Yeah. I mean, once you get like an effective autonomous system, >> [music] >> then the bottleneck to scaling is like how fast can you pump it out? Like how many of these can you make [music] not a year, a day? I mean, that's why we made it out of sheet metal. It's lightweight, it's cheap, [music] it's really easy to work with. Look Look how ridiculous this is. People who know anything about sheet metal know that this is >> [music] >> a very, very weird bend to make. This is too big to be bending. But, [music] I it does introduce one problem that's very specific to forklifts. They are usually carrying thousands [music] of pounds.

[00:03:57]So, you need some weight to counterbalance that load. And then, when you're working with sheet metal, you don't have that. [music] So, our solution, we just decided to put these two giant lead blocks in the back. How much do you [music] think those weigh?

[00:04:10]>> I don't know.

[00:04:11]>> 2,000 lbs each.

[00:04:13]These operate in warehouses across the states. One of our customers is one of the world's largest generic drug manufacturers. A single pallet's worth like a couple [music] hundred thousand dollars here.

[00:04:22]>> Why is that?

[00:04:23]>> They They call it pharma elegance.

[00:04:25]>> [music] >> They polish the floors two times a day.

[00:04:28]They make sure that every pallet has not even a chip on it. If there is, they got to replace it. So, we literally can't mess up. Like [music] it it's not toilet paper, man. It's a lot of money. What it does for us is like if we work here, we can go to any customer and say like, "Yeah, man, we can move your dog food.

[00:04:43]It'll be fine." Half the time it's autonomous. The other half of the time is teleoperated with one of these. We basically realized, you know, the customer doesn't care if it's autonomous or not. It's all about the work getting done.

[00:04:57]What do you say?

[00:04:58]>> What's next?

[00:05:00]>> I mean, we have 15 forklifts coming in, hopefully, next [music] month. They're all already spoken for by multiple customers. We just moved into this space like three or four months ago.

[00:05:10]And we already have to get a space like three times [music] bigger than this.

[00:05:13]>> I can tell you've been fundraising lately.

[00:05:15]>> No, you can't.

[00:05:17]>> [laughter] >> I mean, the goal is to get anything [music] anywhere in a few hours. This means like 50,000 tons from Cape Town >> [music] >> to New York few hours. Needs to be a possibility. If we want to be a multi-planetary species, we want to cure every disease, we want [music] to solve every problem, you need the infrastructure to solve every problem. If we want to get to these massive problems, we've got to solve this today.

[00:05:41]>> And are are are you going to?

[00:05:43]>> Yeah, of course.

[00:05:45]>> Hello.