Hacker News Reader

Comment by mtct88 | Hacker News Reader

I still don’t understand how this can be considered cheaper or more productive than using a human.

I’m all for automation in industry, but the "human simulation" approach (where a robot mimics a human on a production line instead of using a process optimized for machine operation) just doesn’t make sense to me.

View on HN

Most manufacturing operations have a long tail of stuff that is done by people because it's just not worth buying/designing a specialized machine for. That's exactly the stuff that people do currently that humanoid robots might step in and do cheaper.

Employees cost roughly 70-120K per year. A lot of work in automotive is skilled work that pays relatively well. So, the economic case is easy to understand.

That's why there are lots of companies trying to produce humanoid robots. If they get good enough, lots of companies would buy them. A humanoid robot costing about 100K that can work around the clock (minus charging, battery swaps, servicing, etc.) doing work that otherwise would be done by a person could earn itself back in well under a year. Maybe it will cost a bit more or a bit less.

Will they be able to do anything? Not right away probably. But they'll probably be able to do useful things which means people don't need to do these things and can do more valuable things with their time.

I work in manufacturing and one of the driving forces here is the proposed cost of these humanoids. The target price range given by some companies is somewhere between 10k€ and 30k€, which would make them insanely competitive vs a human or a custom automation (which is easily over 100k€).

> which would make them insanely competitive vs a human or a custom automation (which is easily over 100k€).

But the humanoids are not competing with custom automation.

Judging by some of the footage from BMW and the humanoid manufacturers themselves, they very frequently boil down to pick n place tasks, which is a field where lower to low cost automation solutions have been available for a while. Often times with significantly higher throughput as well.

Its been a while since I was dealing with shopfloor stuff and I am not an expert, but I do not see these humanoids anywhere near as compelling as many people pretend they are

well, I imagine they have to start somewhere

and they will gradually get complicated tasks

actually sounds like a trainee

Employee costs vary greatly depending on where the factory is located. Here in Italy, a line operator costs around €60,000–70,000 per year, and half of that goes back to the government for public welfare redistribution.

Robots costs have a fixed CapEx that humans don't have. If they become expensive you can move the factory to a cheaper nation.

People keep saying that a humanoid robot will cost around €30,000, but is that just for the hardware, or does it include all the additional services required to operate it? Will they be as interchangeable as humans, who can be reassigned to a different task in 30 minutes without notice?

Honestly, it still doesn’t make sense to me; to use an analogy, it's like you're building and "horseless carriage" instead of a car.

It makes sense when you find out that there's only so much humans who want to do and are capable of doing that type of work. Also, robots are cheaper, and they are the opposite of humans - theyt are not emotional, lazy, feeling fatigue or any other spectrum of feelings or behavior ...

But why humanoid robots, instead of robots optimized for the task at hand?

You have to rebuild the factory and manufacturing process I guess. Humanoid robot is a drop in replacement for human requiring no intervention

Because humans can do any task. A fixed robot can only do one task.

A cheap humanoid robots will do wonders to alleviate the need for migration. And they can be plugged anywhere you have existing humans working.

Reply on news.ycombinator.com