Hacker News Reader

Comment by alexgmcm | Hacker News Reader

Robotics still seems a long way off being able to automate many physical jobs - would you trust a robot to look after your aging parents without hurting them?

Likewise jobs like computer programming seem pretty 'robot-proof'.

It seems only repetitive, predictable work can be automated and we shouldn't waste human labour on those tasks in any case.

Personally, I think the threat of automation has been greatly exaggerated in the media, it is true that self-driving cars would have a huge economic effect but progress appears to be slowing down in that area.

View on HN

would you trust a robot to look after your aging parents without hurting them?

Having seen how little many humans care about other peoples aging parents under their care, I'm not so sure robots would be a huge step down.

edit: but as others have pointed out, robots will/should be there to help humans do the hard parts of their jobs not completely replace them. So in the future we'll probably see a human and a robot doing what used to take 3-5 humans, not simply a fleet of robots replacing all humans.

> Having seen how little many humans care about other peoples aging parents under their care, I'm not so sure robots would be a huge step down.

Which is an issue. It's the old addage: I don't have to outrun the bear, I just have to outrun you.

If the robots are only maginally better than the humans, then they will out compete the human workers and we will have robots that may still be very bad at doing the task; potential failure all the same and with bad patient outcomes.

> So in the future we'll probably see a human and a robot doing what used to take 3-5 humans,

In Chess, these amalgams are called 'centaurs' and they consistently outpreform humans and/or the best computer programs. It's a good idea in chess, as most centaurs are made with very good human players. It's when the humans start to rely on the computers that you get into real trouble.

True, productivity growth is very low compared to the past. It seems that the long recession coincided the hyping of certain new tech (self-driving cars, drone delivery) leading some people to wrongly believe that we were on the verge of robots causing mass unemployment.

On the other hand, I think we might not have really dealt with the changes from mass automation in the past. If you look at things like "bullshit jobs"[1], or large sectors of the economy that provide negative value to society, it seems like previous automation freed up huge amounts of labor that simply haven't been used to improve society. Since that's already happened, though, it's much harder for people to see it as a problem.

[1] https://www.newyorker.com/books/under-review/the-bullshit-jo...

Taking care of parents without hurting them is very easy. The bigger problem is whether they can take care of them at all.

Also the real threat of automation is wealth inequality. If workers own (a share of) the robots that replace their jobs then there is no problem. If all the robots are owned by corporations there is a massive problem.

Actually, wealth inequality is a big threat /to/ automation because lower wages means automation isn't economically efficient.

The first places where automation is put in place is where labor is expensive. That is, both geographically and industry-specifically.

It's not the factory workers so much as the accountants, lawyers, and doctors that are marginalized first. Software is cheaper and the reward greater.

And with respect to geographically, Scandinavia employs the most robotics to augment manually labor jobs. They will continue to lead in this area.

I found that interesting. I've also heard it said that unemployment is higher in France compared to England. But their (France) productivity is higher. I wonder which is better...

Productivity in my opinion - there is no point in employing people if it isn't efficient.

But it seems most societies still seek jobs for jobs sake.

Agreed. We need to migrate to worker owned and run businesses ASAP.

I think there's some pretty significant blowback coming. Poor people don't just disappear.

Work is going to be here for a very, very long time. There are aspects of it that can be automated, like some parts of manufacturing.

I think the bigger problem we are going to encounter is a retiring workforce. And the lack of knowledge transfer in a lot of industries. I recently walked into an office with a PBX and no one knew how it worked. The servers were installed in a basement and the documentation was damp, moldy, and crawling with bugs. What should have been a simple transfer to VoIP now requires tracking down lines and real downtime.

> I think the bigger problem we are going to encounter is a retiring workforce.

Or a workforce that is "replaced" by young, fresh-off-university people that are desperate for a job that pays anything to survive with no benefits, while at the same time the old guards with real, proper incomes are fired as they are vastly more expensive...

Isn't it strange that his has become such a problem, people who can't or won't retire, once pensions were destroyed?

This is another, separate problem - but yes, real indeed. This kind of stuff happens when government services get defunded or outright destroyed...

I was talking with our 401k company guy about this. But basically, he told me most 401Ks who were invested before 2009 are just becoming whole now. If you were planning to retire before 2009 and converted to bonds before 2008, like you were supposed to, you were okay. And if you started contributing after 2008/2009 you did well. But for a lot of folks, who contributed on the riskier side of things (basically anyone under 55) you lost 10 years. Anyone that planned to retire at 65-67 is now having to do so at 75, otherwise they will be underinvested at retirement.

I anticipate you are going to see a large wave of workforce retirement because there is such a large demographic who had to delay retirements. And that bill is now coming due.

It is easy to imagine an elder care robot that is humanoid in features and attempts to directly replicate what a human does. But that is rarely how it plays out. Instead you get technology that solves a single practical problem. For example, a washing machine, ready meal, dishwasher, or hi-tech toilet. All replace activites that require a lot of labour to do by hand. Maybe there are other practical technologies that would be of utility to the elderley and help maintain independance.

Yes. We already have retirement complexes where eg chairs have sensors which alert someone nearby if someone is sat down for too long. Net result, you have less need for someone to 'check in' on people (but then you need to find another way to give people social contact). It's about using human labour more efficiently, not replacing humans.

> would you trust a robot to look after your aging parents without hurting them?

I often see these kinds of examples with 'non-automatable' work. I agree that the full extent of this might not be automatable, but there's probably opportunity for efficiencies through automation, which won't make all people reduntdant, but might decrease the number of people needed to provide this care.

One thing I could think of is camera systems with anomaly detection to watch over people to make sure they're not hurt or something. Or biometric devices that provide feedback to a centralized system. Sure, people might be weary, bot 15 years ago people were stil weary of mobile phones too.

Like computers, robots should be seen as workforce multipliers. A robot that helps lift elderly out of the bed with human supervision makes a two person job a one person job.

The typical way automation takes job, is not by replacing them but by increasing the efficiency of the task being done, therefore needing less workers to do the same job.

>would you trust a robot to look after your aging parents without hurting them?

Physical robots are not there yet, but you can already get an improved efficiency by camera monitoring. Using computer vision and pose tracking you don't have to always be present, you only need to be able to dispatch someone if a problem is detected.

> would you trust a robot to look after your aging parents without hurting them?

My grandparents certainly could use a proper self-driving car if it were able to drive on the Croatian "magistrala" (central road on the coastline, very dangerous as it is), and my parents are really satisfied with the Roomba my sister and I got her for xmas two years ago. Probably what could be very useful to both is a robot that does laundry and dishes, basically the two tasks that are most demanding on their backs.

For people in retirement homes or with disabilities, robotic assistance in basic bodily functions (=showering/keeping clean/eating) or cooking probably would be a godsend.

This is the kind of robotics we need to free up (caretaker) resources, but unfortunately these tasks are simple for humans but exceptionally difficult for robots.

For older people doing dishes what often really helps is simply a dishwasher that's installed at desk-height so that nobody has to bend down to use it. A simple (though possibly costing some $$) change that makes no difference to young people but all the difference to elderly folks.

One of the problems with these type of robots, is misuse. People who could become capable of doing for themselves, instead become more and more disabled because the lean on automation and won't work to develop their own ability. We already see some of this with mobility scooters and such, or elderly people who end up in nursing homes for physical therapy and end up never leaving.

Likewise jobs like computer programming seem pretty 'robot-proof'.

Depends on how you look at it. I see programming as the single most automation-friendly job there is. Even if all you do is add a few lines to your Makefile to run the test suite when you type 'make test', you're still automating a repetitive task you'd otherwise do by hand.

Looking at the bigger picture: programmers are constantly developing new programming languages, libraries, text editor plugins, build tools, deployment tools, continuous integration servers, SCM tools... These things are all designed to save programmers time and if we are successful it means we're increasing our productivity, which should mean we don't need to hire more programmers to do the same job otherwise.

Maybe the last point seems hard to accept in light of the fact that there are more working programmers than ever. Just try to imagine what it would be like to build an app like Uber if you still had to write code on punch cards and batch-feed them into a giant single-tasking computer.

I think there is a similar thing with electricity production, having a fusion generator would give "huge amounts of electricity" until someone thought of a use for all that electricity. There's always something bigger you can find that is limited by current amounts of resources.

Another good one is roads. The old way of thinking was straightforward: "we've got too much traffic, let's build more roads!" Now we realize that building more roads encourages more people to drive. There seems to be practically unlimited demand for roads.

Programming is automation, and that's why it's a job that can't be eliminated by automation.

Every line of code ever written is literally automating something. Yet no programmer ever worries "will this be the line of code that puts me out of work?"

> would you trust a robot to look after your aging parents without hurting them?

Robots should be introduced into caring for the aged as soon as possible.

I can't say this strongly enough: It is a really terrible world for old people who have reached a certain age or level of incapacity.

If you've spent any time with someone caring for an older person at home, or in a nursing home, or in a hospital, there is SO much room for the improvement of people's lives.

home care needs automation.

nursing homes need automation.

hospitals need much much better automation.

Things that people need: help getting someone out of bed. Help getting someone into bed. Help getting someone to go to the bathroom. Help giving someone a bath or shower. Help getting someone some food. Help going outside.

Getting help in these ways is so expensive they basically don't happen.

I could go on and on.

> Robotics still seems a long way off being able to automate many physical jobs - would you trust a robot to look after your aging parents without hurting them?

You're only seeing what's left after centuries of automation/mechanization. Thank of all the tools people doing physical labor everyday use, power drills, various power saws, nail guns, angle grinders, air compressors, etc. These were all done via manual labor not long ago, many within living memory.

Even repetitive, predictable work like making sandwiches to order or picking up litter are tremendously difficult to automate. The tasks seem simple but when you break down all the motions and error conditions there's a tremendous amount of mechanical complexity. I predict that automation will have only a minor impact on those types of jobs by the time most of us reach retirement age.

Sandwiches may become more robot-friendly in their structure, without losing their taste.

Picking up trash is harder. But much of the high-precision picking can be replaced with coarse vacuum-cleaner-style action. The biggest problem likely will be determining what is trash and what is nit.

Machine Learning is augmented human computer programming

Reply on news.ycombinator.com