The Engineering Reason this Robot Feels Human | 1X Neo Factory

0:00 Just imagine like blocks of aluminum going back

0:03 there and then it gets formed into parts.

0:05 Parts get formed into joints.

0:07 Joints get formed into limbs and lints get put up as a neo.

0:11 I walked into this factory expecting to be impressed by the humanoids.

0:15 But what I was not expecting to find was one quietly sorting

0:19 parts in the corner of the same building where it was made.

0:23 Okay, so that's Neo and it's working in its own factory.

0:27 And that kind of detail broke my brain for the rest of the day.

0:31 You produce almost everything in house.

0:34 So why was that so important to you?

0:36 And what are some of the benefits?

0:37 For humanoid robotics,

0:38 there exists no supply chain or systems that actually solves the problem.

0:42 So ideally, you can take existing industrial robot

0:45 systems and just put that together in humanoid form.

0:47 And that might work reasonably well in a factory,

0:50 but it's not going to get you to actually true general labor.

0:53 There needs to be real innovation all the way down

0:56 to the fundamental principles of how the machine works and how it's built.

0:58 If you want inter iteration speed, you have to do everything yourself.

1:01 Pretty proud that we're now going essentially raw materials

1:05 to finished humanoid with our own foundation models within one company.

1:09 Right?

1:09 If you go back to Ford, right, in the river factory,

1:12 they mined the ore, but they didn't just mine the ore.

1:15 They grew the cotton.

1:16 They had the cows.

1:17 The entire ecosystem of that area could be turned into a car.

1:20 If you want to re-industrialize America, this is where we need to go.

1:23 Now, I really want to emphasize this.

1:25 Every single thing built in this building, from the motors,

1:28 the battery packs, the metal parts, 1x made it.

1:31 And I kept on asking Burnt, "Why?

1:33 Why not just buy the components like everyone else does?"

1:36 And he kept coming back to the same word, iteration.

1:39 How fast can you find the problem, fix it, and try again?

1:43 If the part comes from a supplier, I mean, the answer is weeks.

1:47 But if you just made it yourself, the answer is tomorrow.

1:51 That is the whole bet.

1:53 Then I met Vicram and he runs robotics at 1x.

1:56 He wanted to show me something about the arm which I found really fascinating.

2:00 Check this out.

2:01 Because we build and test everything here.

2:03 It's just um

2:05 and the engineers are on the floor here.

2:07 So if there's any issue, it get fixed in like a day.

2:11 It's wild.

2:11 Yeah.

2:12 So the communication is

2:13 the communication, the feedback is for sure.

2:15 So that's why we can keep on improving.

2:17 Yeah.

2:17 So this is the tendon.

2:19 Tendon.

2:19 Call it the tendon.

2:20 The entire architecture as you can see we

2:22 have our own handbuilt custom motors over here.

2:26 Wow.

2:25 So this is all built internally.

2:27 This motor holds the world record in torque to strength to uh torque to weight.

2:32 It's 5x more powerful than any motor in the world.

2:36 We wow.

2:38 Now the reason that matters and Vicram showed me

2:40 this in person is what happens when something hits the robot?

2:44 When the hand hits something, this motor is still spinning.

2:49 Yeah.

2:49 If the motor is still spinning, the hand is going to keep going.

2:51 So, it's gonna hurt you.

2:53 I see.

2:53 Yes.

2:54 In the tendon drive system, as you can see, there are no gears.

2:57 It's basically just these ropes.

2:58 It's incredible.

2:59 And then, and then if you see, it's all back drivable.

3:02 So, if there was a hand over here, for example, I can do this and this moves.

3:07 You will not be able to do this in any gear drive.

3:09 So, you can try it yourself.

3:10 So, not only is it, you know, safer,

3:13 but two, the movements of Neo are so natural.

3:16 So, human.

3:17 Yeah.

3:17 Humanike.

3:18 Exactly.

3:18 They're not stiff.

3:19 They're just organic.

3:20 You know, when he goes to shake your when

3:21 Neo shakes your hand or anything like that.

3:23 Exactly.

3:23 Neo is 30 kg.

3:25 Mass reduction is one of the key tenets of safety.

3:28 We kind of draw upon all the sports research

3:30 that goes on in like in the NFL or wrestling

3:34 and we take all that information and we develop

3:36 our soft goods so that we can prevent concussion.

3:39 And yeah, so I would just uh just uh look out for the pinch point.

3:42 So I would I would just hold it from down here.

3:46 This this structure over here and just lift it like this.

3:48 It's a little heavy.

3:49 So see like here, right?

3:53 Yes.

3:54 Perfect.

3:54 It's not as heavy as I thought it would be.

3:55 It's not

3:56 this.

3:56 It's not bad.

3:57 I mean, I can do some curls.

4:00 It's amazing.

4:02 If your robot is safe,

4:03 your robot can learn faster because it can go into any environment.

4:06 Now, I want to stay on that for a second.

4:08 If you cage the robot,

4:10 I mean keep it in a controlled environment to avoid accidents,

4:13 you cap how much it can learn.

4:16 Safety and intelligence are not intention here.

4:18 Safety is what lets Neo be in the real world.

4:22 And the real world is where it gets smart.

4:26 We can go from raw materials to finished product in less than a month.

4:31 Wow.

4:31 aluminum block, machined part, joint, limb, head,

4:37 the whole thing in under a month.

4:39 And they're already trying to get it down to just weeks.

4:44 The head is where it gets interesting.

4:46 That is essentially where the robots, the humanoids brain lives.

4:51 But when the head gets put on, that's

4:52 when the brain actually connects to the nervous system

4:55 and we light it up with the uh robot operating system.

4:58 It calibrates itself and then it does uh a little bit of a robot,

5:02 we call it the robot gym where it does some squats.

5:05 It it does some, you know, curls and it does, you know,

5:08 uh walking and running and jumping and then and then we do a final check and

5:13 it really is a work of art, isn't it?

5:14 Like when you look closely at, you know,

5:16 all the covers, how that's done, it's it's really incredible.

5:19 And then I went to the machine shop.

5:20 Julian's team makes every metal part in the building.

5:24 You want to you want to talk about in the shop?

5:27 So, this is basically one of these tools.

5:30 Oh, I see.

5:31 Yes.

5:31 Taking it.

5:34 Changes the pool.

5:35 Yes.

5:37 Yeah.

5:36 Is this the hardest or the most most complex or one of

5:39 It's a really complex part.

5:40 I would say this is a 10 out of 10.

5:43 Honestly,

5:44 because you have to machine it from every direction.

5:46 The second time when you clamp it, you lose tolerance.

5:49 It's not as precise anymore.

5:50 You can hold this and see how complicated this is.

5:52 So intricate.

5:54 That's a It's like a watch.

5:57 Incredible.

5:59 Wow.

5:59 Julian's team programs the tool path for every single part Neo is made of.

6:03 That hand piece takes an hour and a half to machine.

6:06 The complexity is frontloaded into the part

6:08 so that assembly downstream is cleaner.

6:11 The hardest thing to build is the thing that makes everything else easier.

6:15 We run this and this runs all night until it breaks.

6:18 So we find we find the failure point.

6:21 Yes.

6:21 We see what caused the failure and we go and fix it.

6:25 And then we come and run the test again.

6:29 Yeah.

6:28 And then we see whether it fixed it and then we will run it to failure again.

6:33 Yeah.

6:32 And then we'll find something else that needs to be fixed.

6:34 And we keep doing this until we have like

6:37 been satisfied that this is a very reliable joint.

6:40 And then we do this for every single joint.

6:42 As you can see, we're getting upwards to hundreds of thousands of cycles,

6:45 approaching a million in some cases.

6:47 So, wow.

6:50 Yeah, that room runs all night.

6:51 Joints are literally cycling until they break.

6:54 Then they figure out why, fix it, and run them again.

6:58 I mean, hundreds of thousands of cycles, some approaching a million.

7:02 Now, I know it kind of sounds obsessive.

7:04 They're going to be in your house, around your family.

7:07 That room is why you can trust that.

7:10 It's kind of wild that what is sitting in front of us becomes Neo.

7:15 Exactly.

7:15 Like it's hard to wrap your head around that.

7:17 It is.

7:17 It is.

7:18 Yeah.

7:18 It It's fascinating.

7:19 It really is.

7:20 Now, I seen Neo a few months back before coming into this factory,

7:25 but standing on the factory floor having

7:27 a conversation with it that felt different.

7:30 And Burnt has one word for what

7:32 makes every Neo different from every other humanoid.

7:35 What truly makes every Neo unique is the memory.

7:40 Yeah.

7:41 The only parallel I have is kind of like it's it's my hobbs in Calvin and Hobbs.

7:46 So that means it's not like getting a dog.

7:50 It's not going to replace my kids.

7:52 It's not going to replace my wife.

7:54 It's something new.

7:55 Now, I've covered a lot of different hardware on this channel.

7:58 A lot of different futuristic tech, but I don't think I've ever heard a founder

8:02 describe their product as a Calvin and Hobbs companion.

8:05 And I found it really interesting because he wasn't trying to be poetic.

8:08 He was being precise.

8:10 pretty confident like a lot of people now will agree

8:13 that in 10 years we will actually have true abundance across society.

8:18 We will have what I would say is a hard takeoff meaning robots building robots.

8:23 We're already working on that.

8:24 But it's not just about the robots building the robots.

8:26 You actually need the robots building the data centers,

8:29 the energy infrastructure, the chip fabs.

8:31 We're going to run our chips building out servicing and running

8:34 the mining and refining to ensure you have the raw materials.

8:38 You can actually get this fully self-sufficient kind of on Newman machine.

8:43 Yeah.

8:43 Where it's producing itself.

8:47 Yes.

8:46 From like Adam to sand to intelligence.

8:49 The entire loop automated.

8:51 Who knows maybe like 200 years from now

8:55 my grandkids grandkids Ara will tell them about me.

9:02 whether the 10-year timeline is right,

9:04 and BERT will tell you 10 is more honest than say five.

9:09 I don't think the direction is really any question anymore.

9:12 I mean, the factory is running, the humanoids are shipping.

9:16 I have a 10-month-old, almost 11-month old son,

9:18 and I genuinely have no idea what the world looks

9:20 like when he's old enough to understand any of this.

9:23 And to me, that used to feel, if I'm being honest,

9:25 a bit of a scary thought, intimidating, if you will.

9:28 But walking out of this factory, out of this building,

9:31 it felt more like a wide open one.

9:33 Like there was so much more possibility for the world that he will be living in.

9:37 You know, I feel like I need to talk about

9:39 Neo sitting behind us here and the case Neo is in.

9:42 It's so interesting to me.

9:44 So, is this the case that Neo will be

9:46 shipped in or tell me about this case in particular?

9:49 So, this is currently the case Neo is shipped in.

9:52 Okay.

9:52 Currently, we will see.

9:53 We will see.

9:54 We're having a big discussion about this internally.

9:56 My personal opinion is that Neo just knocks on your door.

10:01 But how do you do that?

10:02 How how do you do that from the shipping standpoint?

10:05 Should we partner with Whimo?

10:06 I think that would be pretty cool.

10:07 That actually would be really your robot comes delivered with an autonomous car.

10:11 I would like to just Neo to just call a Whimo and uh just get to someone's home.

10:16 Okay.

10:16 So you're you're you and Brent are on the same uh wavelength with that.

10:20 We don't need cases, Neo.

10:21 Just just go just go uh just go to your home.

10:23 It's amazing.

10:24 I know.

10:24 Go home.

10:27 Okay, so 1X starts shipping this year and I

10:30 still haven't decided which delivery I would want.

10:32 I mean, the case, the Whimo, or having Neo show up at my door

10:36 with his charger already in hand, ready to work.

10:39 I think a combination is what's going to happen here.

10:41 Now, if this is your first time here,

10:43 I make videos about the technology that's actually building the next thing.

10:47 Make sure to subscribe and I will see you in the next one.

10:50 Oh, and I linked 1X down below.

10:51 So, you know, you can go buy your own humanoid today,

10:54 which is so wild to say out loud.

10:57 I had to give it a shot.

10:58 All right, I'll see you in the next one.