3 ms·

This is roughly the same content Brooks posted himself a few years ago, and was covered on HN last week.[1] There are already many companies selling automated

This is roughly the same content Brooks posted himself a few years ago, and was covered on HN last week.[1] There are already many companies selling automated carts that run around warehouses. Search for "automated guided cart". I thought the humanoid thing was silly until I saw the pricing. Here's the Unitree G1, starting out at US$22,000, less than a Toyota Corolla. I though these things were going to cost like Boston Dynamics products. No, the hardware is already much cheaper. This is still a low-volume product and prices are headed downward. Humanoid robots are going to be cheaper than cars. Having more degrees of freedom than you really need for any single job will be outweighed by the cost advantages of mass production and the advantages of interchangeability. The manipulation problem remains tough, but with moderately priced and standardized hardware available, more people can work on it. [1] https://news.ycombinator.com/item?id=45392922 https://news.ycombinator.com/item?id=45392922

ssivark 1y ago
> Here's the Unitree G1, starting out at US$22,000, less than a Toyota Corolla. The sticker price might look cheap, but what I've heard from people actually looking to buy (even in large quantities) is that by the time you assemble the hardware and the tooling and the dev kit, the cost is $80-100k per robot. > This is roughly the same content Brooks posted himself a few years ago, and was covered on HN last week. That Brooks' post is his latest one, from a few days ago.
- matthewdgreen 1y ago
  Seems like the kind of problems that would be amortized by scale. Anyone running a fleet won’t need to assemble one set of tooling per robot.
  - ssivark 1y ago
    Unitree wants to charge buyers $80-100k per-unit (at scale) in order to actually make money on their humanoids. So we're not in the era of $20k humanoids, despite appearances.
    - bluGill 1y ago
      What scale? If Toyota sold 1 car per year they would need to charge hundreds of millions to pay for the engineering. Just making an engine to meet modern emissions will cost more than a billion, but since that engine will be made for a decade we can amortize it and say the cost of the engine is 100 million per car. However because they make many more than that their costs go down. Because they make many cars they have developed assembly lines - to build the whole factory is likely another billion dollars (I don't have real numbers), but if you are only building one car a skilled machinist could do that in a year for much less cost than a factory. Thus my question - what scale are they targeting. At some level of sales their expected costs are right. However if demand is a lot more than they expect it could probably scale down to $10k, maybe less, while if it is less costs go up...
      - datadrivenangel 1y ago
        The point is that the list price of $22k is basically fake if a full system costs $80-100k. It's like saying that a car costs $22k, but you can't buy it without also getting a $60k lifetime support contract.
        matthewdgreen 1y ago
        There is "fake" as in, we offer the thing at $22k but this is a made-up price and we'd be losing tens of thousands per unit if we couldn't upsell the necessary tooling. Then there's "not fake" as in, the hardware is profitable (or at least break-even) at a list price of $22k, and the tooling is expensive because it makes sense to charge a lot to developers. These things put you at different places when scale starts to increase. It's extremely common for dev tooling to be extremely expensive, even for widgets that sell for a few bucks at the retail level.
        bluGill 1y ago
        Why does the full system cost 80-100k though? Can I buy 1 full system at 100k, and then 10,000 at the list price? What is the difference. If I sign a contract to buy 1 million next year (thus giving you time to build an better assembly line) can you get the per unit price down?
- Animats 1y ago
  > The sticker price might look cheap, but what I've heard from people actually looking to buy (even in large quantities) is that by the time you assemble the hardware and the tooling and the dev kit, the cost is $80-100k per robot. I know. You don't get the GPU for AI in the base model, for example. Do you get the articulated hands? A robot package with all those custom parts, motors, and motor controllers, along with cameras and a short range LIDAR. for only $22K is pretty good.[1] As this technology starts to work, it's apparently not going to be insanely expensive. [1] https://robostore.com/products/unitree-g1-robotic-humanoid https://robostore.com/products/unitree-g1-robotic-humanoid
  - lesuorac 1y ago
    Pretty good? It can lift 3 kg / arm. That's like barely enough for a super-sized coffee! What human role are you replacing with this? Some desk job? Why bother with the robot.
matthewdgreen 1y ago
It also tells a story about any US robotics company like Tesla now claims to be. How can you possibly hope to compete against manufacturing costs that low.
Palomides 1y ago
nobody has solved manipulation at any price point, you can't scale your way out of that
- singularity2001 1y ago
  > more people can work on it. His point was that mass production will make it more likely that someone will solve it
  - bluGill 1y ago
    Just be careful. Mass production depends on masses buying it. If they can't get the price low enough mass production may not be worth it. It is entirely possible that enough people would buy this at $1k, as to make mass production worth it, but if you target $10k not enough people would buy as to make mass production worth it. This though assumes that you could do enough mass production to get down to $1k, which might not be possible.

ssivark 1y ago

> Here's the Unitree G1, starting out at US$22,000, less than a Toyota Corolla. The sticker price might look cheap, but what I've heard from people actually looking to buy (even in large quantities) is that by the time you assemble the hardware and the tooling and the dev kit, the cost is $80-100k per robot. > This is roughly the same content Brooks posted himself a few years ago, and was covered on HN last week. That Brooks' post is his latest one, from a few days ago.

matthewdgreen 1y ago
Seems like the kind of problems that would be amortized by scale. Anyone running a fleet won’t need to assemble one set of tooling per robot.
- ssivark 1y ago
  Unitree wants to charge buyers $80-100k per-unit (at scale) in order to actually make money on their humanoids. So we're not in the era of $20k humanoids, despite appearances.
  - bluGill 1y ago
    What scale? If Toyota sold 1 car per year they would need to charge hundreds of millions to pay for the engineering. Just making an engine to meet modern emissions will cost more than a billion, but since that engine will be made for a decade we can amortize it and say the cost of the engine is 100 million per car. However because they make many more than that their costs go down. Because they make many cars they have developed assembly lines - to build the whole factory is likely another billion dollars (I don't have real numbers), but if you are only building one car a skilled machinist could do that in a year for much less cost than a factory. Thus my question - what scale are they targeting. At some level of sales their expected costs are right. However if demand is a lot more than they expect it could probably scale down to $10k, maybe less, while if it is less costs go up...
    - datadrivenangel 1y ago
      The point is that the list price of $22k is basically fake if a full system costs $80-100k. It's like saying that a car costs $22k, but you can't buy it without also getting a $60k lifetime support contract.
      - matthewdgreen 1y ago
        There is "fake" as in, we offer the thing at $22k but this is a made-up price and we'd be losing tens of thousands per unit if we couldn't upsell the necessary tooling. Then there's "not fake" as in, the hardware is profitable (or at least break-even) at a list price of $22k, and the tooling is expensive because it makes sense to charge a lot to developers. These things put you at different places when scale starts to increase. It's extremely common for dev tooling to be extremely expensive, even for widgets that sell for a few bucks at the retail level.
      - bluGill 1y ago
        Why does the full system cost 80-100k though? Can I buy 1 full system at 100k, and then 10,000 at the list price? What is the difference. If I sign a contract to buy 1 million next year (thus giving you time to build an better assembly line) can you get the per unit price down?
Animats 1y ago
> The sticker price might look cheap, but what I've heard from people actually looking to buy (even in large quantities) is that by the time you assemble the hardware and the tooling and the dev kit, the cost is $80-100k per robot. I know. You don't get the GPU for AI in the base model, for example. Do you get the articulated hands? A robot package with all those custom parts, motors, and motor controllers, along with cameras and a short range LIDAR. for only $22K is pretty good.[1] As this technology starts to work, it's apparently not going to be insanely expensive. [1] https://robostore.com/products/unitree-g1-robotic-humanoid https://robostore.com/products/unitree-g1-robotic-humanoid
- lesuorac 1y ago
  Pretty good? It can lift 3 kg / arm. That's like barely enough for a super-sized coffee! What human role are you replacing with this? Some desk job? Why bother with the robot.

matthewdgreen 1y ago

It also tells a story about any US robotics company like Tesla now claims to be. How can you possibly hope to compete against manufacturing costs that low.

Palomides 1y ago

nobody has solved manipulation at any price point, you can't scale your way out of that

singularity2001 1y ago
> more people can work on it. His point was that mass production will make it more likely that someone will solve it
- bluGill 1y ago
  Just be careful. Mass production depends on masses buying it. If they can't get the price low enough mass production may not be worth it. It is entirely possible that enough people would buy this at $1k, as to make mass production worth it, but if you target $10k not enough people would buy as to make mass production worth it. This though assumes that you could do enough mass production to get down to $1k, which might not be possible.