Instead of corporations buying robots | Hacker News Reader

Instead of corporations buying robots

3 pointscoroutines9 comments

Sometimes I hear people talking about "the future" where most jobs are automated in sweeping waves of corporate greed.

I was just thinking:

Only large corporations can afford the immediate buy-in cost of a fleet of robot workers. The medium to small companies would only be able to do this incrementally. What if instead of marketing robot workers to corporations, companies making robots marketed their hardware to individuals? I have a naive understanding of what is possible with robots these days - but I'm still of the mindset that it's mostly manual labor with repetitive tasks that are easy to automate.

I think it'd be interesting for "the robot company" to market individual robots to blue collar workers that they can use at their job to do their work for them. The company the worker... works for would essentially be leasing use of the robot from the worker. The worker then acts as the robot technician.

To me this would seem like a much easier way to gradually automate everything without companies having to take the brunt of the costs of buying 2000x robots. Buying a robot replacement for your own labor would be like mortgaging a house. You'd pay off owning that robot over 5-6 years like a car.

Maybe this is silly but I could see there being several types of general-purpose robot workers [[someday]] and you apply for jobs based on what robot you have. "Oh yes, I can offer X company the use of my Masonry Model X7 robot..." You negotiate a rate for the work your robot will do, and you as the owner ensure that it operates at capacity for the needs of the company. There's also a component in here that if a robot can far outpace a human, the owner of a robot would have more bargaining power because the company wants that level of productivity.

I'm not sure how you'd handle the problems of having robots and workers working alongside each other. What legal considerations should be outlined for this, etc etc..

Thoughts?

It's a lovely idea. But in history's many examples of labor being replaced by technology, none of them have ended up with former workers owning the new means of production. Forces against individual ownership include:

  - capital is cheaper for companies than individuals
  - companies can use technology more efficiently than individuals
  - companies have longer planning horizons than individuals

Perhaps the outcome with robots would be different. But I haven't seen a convincing explanation why this might happen.

Physical labor can be replaced with robots. But intellectual and creative labour won't be for a long time - AI notwithstanding.

Even the simplest tangible product requires considerable manufacturing capacity and capital investment. That is why startups are more likely to focus on yet another app or website. So many KickStarter product projects fail to come to fruition. Even Nest, DropCam and Revolv are faltering - in spite of Google's technological and financial prowess.

I think part of this future robot company would be making robots that understand what they're being trained for.

Like, you can make the robot mimic your movements of screwing on a lid to a jar on a production line - but you also need to make it understand that the lid needs to be reasonably tight on there. It can't just perform the motion of the task without knowing what the objective is.

The company making the robots would become competitive by saying "look how easy it is to program our robots compared to x-competitor!"

This is how I see a way to transition into that all-automated future.

I definitely agree that corporations would not want to put power in the hands of the workers.

However, I think a fleet of robots would be expensive for most companies - and the robot-making company might make far more money offering robots to individuals?

But most companies main objectives are the opposite. They invest X$ upfront so they can save $y/yr.

Here's an example with an HVAC system but it's basically the same analysis.

http://www.csemag.com/single-article/economic-analysis-in-in...

Robots need to be programmed. Even the simplest task takes a lot of effort to program and then even more to debug. During that time the robot produces a lot of faulty product that needs to be scrapped.

Large companies can afford the programming costs because the amortise the development cost over a large number of units.

Small businesses generally do not have the resources to market, distribute and sell large quantities of product. Thus you need a lot of capital to play in this game.

I'm sort of imagining a general-purpose robot that looks like the ES5's in iRobot and you could train them like the mimicked movement in Real Steel (2011).

Although uber's doing it with cars so if you had some kind of marketplace.

If ebay or amazon had some kind of centralized manufacturing facility with tele-presence robots you might see a model like that. Probably wouldn't benefit US workers.

Like a mechanic who owns and brings his own tools.

I imagine a look in a mechanic's tool chest every ten years would show some things the same, and some things different.

Now a robot is in your toolbox.

Reply on news.ycombinator.com