Global Debate | Free Debate: Does it have to be humanoid?

[00:00:02]Look at Rob. Look at Rob. Is there something you want to embrace today?

[00:00:05]It's remote control that does dancing.

[00:00:07]Name me one scenario that it can autonomously accomplish today so that you can embrace it into your life.

[00:00:14]>> And that is why we choose the word embrace. As you heard in our definition, embrace we mean make a priority.

[00:00:23]Prioritize our money, our talent, and the policy to make robots that look like Rob able to work in any environment, whether that's the hospital, whether that's the factory, whether that's at home.

[00:00:37]>> Just imagine a couple, both are working.

[00:00:39]One of the member goes away foreign country, and then [music] other partner is at home. One of the partner is collapsed. There's no one to attend to, no one to call. Imagine you have a robot at home. It could do some preliminary evaluation [music] quickly, make the right contact.

[00:00:58]>> Well, if Rob is at home, it will follow upon you.

[00:01:03]>> Let's, ladies and gentlemen, go to the free debate segment. 4 minutes for each team. We start with team A.

[00:01:11]>> We agree with your point that specialized robots can perform better than human form factor robots. And in the long run, form does follow function.

[00:01:20]But in the short run, the world, as Gavin said, is built, the physical infrastructure for humans. It would be incredibly expensive to rebuild that at the same time while we're developing robots. My second point is that sk- single form factors can drive down costs dramatically. And this will be crucial for the global south. So, it's not only that the developed world, the people that can afford robots benefit from this, but when the cost is low enough, the entire world can benefit.

[00:01:51]>> Well, look at Look at Rob. Look at Rob.

[00:01:53]Is there something you want to embrace today? It's a control that does dancing.

[00:01:57]Name me one scenario that it can autonomously accomplish today so that you can embrace it into your life.

[00:02:05]>> And that is why we choose the word embrace. As you heard in our definition, embrace we mean make a priority.

[00:02:13]Prioritize our money, our talent, and the policy to make robots that look like Rob able to work in any environment, whether that's the hospital, whether that's the factory, whether that's at home.

[00:02:27]>> Okay.

[00:02:28]>> Just imagine an example. Just imagine a couple, both are working. One of the member goes to a foreign country, and then other partner at home. It's a real example, actually. And suddenly he get a call. Actually, he doesn't even get a call because the partner collapse at home, no one to attend to. That's because increasing blood sugar levels, many other things because of modern stress life. Actually, she the person, one of the partner is collapsed. There's no one to attend to. No one to call. Imagine you have a robot at home. It could do some preliminary evaluation quickly, make the right contact.

[00:03:07]>> Well, if Rob is at home, it would fall upon your your >> What what it means is to you have somebody which is you can remotely access information.

[00:03:18]>> Why does it have to look like a human?

[00:03:20]Why can't it be a integrated scanning system that gets all your vitals? Why can't there be a drone in with a medic or with a proper system? Why does it have to be >> emotions are like this. They like to see themselves in others.

[00:03:34]>> This guy is like collapsing on the floor, almost dying. Why does he want emotional support?

[00:03:38]>> You You are buying a toy. Would you like to buy a toy look like a monster? Would you like to buy a toy which looks like a reasonably okay >> I want a toy that works. I want something that actually serves my purpose. And here's what you don't explain. So, let's picture paint here for people who wonder where this is in.

[00:03:54]Okay, let's take a factory. In a factory, on an assembly line, you have simple to complex jobs. Those jobs right now are done by single mechanic arms that are reliable, that can paint the same task repeatedly. What happens when you turn this into a human or humanoid robotic? A humanoid robotic is heavier.

[00:04:12]It is insecure. It has so many more junctions and decisions that it's able to make. It can make decisions easily.

[00:04:19]So, the fact is when you put a humanoid robot into a factory, the error rate is something like two-digit percentage.

[00:04:26]But, if you have a single arm, it is able to be secure. So, why replace something that is already working with a new form?

[00:04:33]>> As I said, economics. We have to drive the cost down. If we're making 10, 100 different kinds of robots, the number that can be produced is going to be much smaller than if we had a single form factor for every kind of use case. The other thing that we have to remember is our definition of humanoid robots are not only physical form factor that look like Rob, but also have the brain that can learn, reason, and act independently.

[00:05:04]>> Well, let's embrace it until we have it.

[00:05:06]We don't have it today. And then, we don't want a toy at home. We want something that works. It simply does not work today. But, in other form factor robot, vacuum robot, you talk about volume. China ships 20 million vacuum robots per year. That's volume. We ship how many of them? 3,000?

[00:05:23]>> Like, I don't understand.

[00:05:25]>> 20,000.

[00:05:26]>> It It seems like your argument is that the human body is the peak of evolution.

[00:05:31]But, that is not the case. There are many things that robots can do right now that we can't do. I can't climb under my bed in the crevices. I can't go up on the roof. I can't fly. So, why don't we invest in infrastructure that is able to solve all of these problems? You haven't told us, even in 10, 20 years, but which by the way is a immensely long-term that you're trying to win on, why is it that human is the ultimate form and it's not something else that we don't know and we need the infrastructure for.

[00:05:56]>> You misunderstand. We did not say long-term ultimate. We said now. The long-term we agree with you. Form follows function. This is why jet airplanes don't look like birds, but that takes time. And if we want the benefit and as we said, we is all of humanity and the benefits are too great to receive. We should be delivering them as quickly as possible and that is the humanoid form factor today.

[00:06:25]>> Gavin and exactly it's a harder way for robots to use man-made tools. They should use robot-made tools. And you talk about the short-term [snorts] impact, so we should forget about whether AI will get a breakthrough in the near future. We are talking about now. Now it's the machines can only do pre-programmed fixed trajectories, so they should use tools of their own. And also you said that robots is a universal time. No, humanoids are just a special form that mimics human. It is also the specialized robots. And by doing it, we need many small motors and joints that dramatically increase the cost instead of reducing it.

[00:07:03]>> No, it's like Gavin and Andy, you can't pick and choose your side and time frame and your stance however you feel like.

[00:07:08]On one hand, you say we want it right now. We're talking about the now, we're not talking about the long-term. On the other hand, you're like, "Oh, I know Bob can't really dance, so like we're going to wait until he can magically rescue someone dying on the ground and not fall on them." You have to pick a stance.

[00:07:22]>> Well, that's the whole point is that the question is not 10 years, 5 years. The question is literally the question is asking, let me read it for you in case you forgot. The question is, should we embrace humanoid robots now?

[00:07:33]>> Are you going to embrace Bob now? How do you know Bob is not going to fall on this old lady instead of saving her?

[00:07:37]>> a question for you. How do you expect us to go from zero to 100?

[00:07:41]>> And that's long-term. And that's long-term.

[00:07:43]>> Thank you for answering the question.

[00:07:45]>> So wait, wait, you literally just said long term. That's not >> It's not now. The question is not but in order to get to long term, you have to start somewhere.

[00:07:51]>> You just said it's [laughter] not now.

[00:07:52]You just said it's not now.

[00:07:54]>> Yours is not now. You have to start somewhere being that somewhere. Rob, is this somewhere to start for us to understand how robots work, how to embrace >> 20 seconds left for team eight. Go ahead.

[00:08:03]>> I'm sure you've heard of Moore's law.

[00:08:06]Doubling every processing power doubling every 18 months. People say AI is doubling in power every six months.

[00:08:14]>> In China we use Tao's law, not Moore's law.

[00:08:17]>> [laughter] >> But in any case, we want to train humanoid robots for the human space.

[00:08:23]A classical example is in Beijing recently.

[00:08:25]>> Sorry, we have to lecture you for 39 seconds.

[00:08:28]>> [laughter] >> Well, >> [clears throat] >> you said now.

[00:08:32]That I disagree. This guy has no brain, the no proper software to make it work, no proper hardware to make it work. So our stance is make it wait. Let's get those components ready so that we can move forward. Then we have a thriving humanoid robotic sector, but not today.

[00:08:48]>> Yeah.

[00:08:48]>> I'm going to jump in and say, right now we are on an empty, clear, and safe stage. You are trying to embrace them into complex real-world environments, factories, hospitals with people, noise, and obstacles. How are you trusting this person to be embraced now?