AI, Robotics, and the Future of Jobs | Hacker News Reader

AI, Robotics, and the Future of Jobs

pewinternet.org

52 pointsrpm432141 comments

Reading articles like this I'm often reminded of Amara's Law:

"We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run."

In terms of overestimation in the short term, it seems to me that although we are trending towards automative technologies, 11 years seems like far too short a span for those technologies to have "displaced significant numbers of both blue- and white-collar workers."

Meanwhile, in the long term, the time frame these technologies actually need to have fully permeated, we don't know what types of jobs or how many will be brought about by their existence.

Exactly my feeling. Although I can't blame someone for overestimating the impact of automation on his jobs and preparing for it. Those who underestimate AIs will pay a hefty price.

Anyway, I don't think the "when" is really interesting to think about, especially when it's an arbitrary date. I prefer to ponder on much more significant and inevitable events/thresholds such as when AIs will surpass humans in every aspect. In my opinion, as long as AIs are inferior to humans in terms of intelligence, there will be jobs for most people, but past this point the concept of "work" is not event relevant anymore. And don't tell me "people will always need interaction with other people". That's clearly underestimating the capacity of future generations to identify with highly intelligent and sentient robots.

> Throughout history, technology has been a job creator—not a job destroyer

This is one argument that I have always really disagreed with. Technology doesn't inherently create jobs. Rather it makes people able to be more productive and efficient. I think that will longer stand the test of time.

Hopefully someday soon we will go beyond this need for everybody to have jobs so that they can pay for food and rent. Instead we will be able to do human things. Do projects, cool hacks, make art and music, explore the universe. Why spend all our time compensating for scarcity when robotics can do it for us?

That however is going to be a long transition. Especially one for the people in the US, who have such a heavy burdened work culture (i.e. if you aren't working you aren't valuable to society). Also shameless plug here for basic/guaranteed annual income.

I'm not sure about the art and music making. Computers will make this too and not just better, but also personalized for your brain. I think, there will be a great depression, when even creativism becomes boring.

Being creative doesn't become boring because someone else can do it better. Just trying to express yourself is fun. If it weren't, nobody would participate in amateur hour or open mic nights.

I agree 100%. The idea of automating creativity is ludicrous. Creativity is what being human is all about and the universe we live in is infinitely vast and full of endless paths to explore.

>The idea of automating creativity is ludicrous.

No it isn't. Not anymore ludicrous than automating anything else, just far more complex in its implementation. It's not impossible.

>the universe we live in is infinitely vast and full of endless paths to explore.

Yes, but human beings are not infinite, nor is their capacity for creative thought. We're bound by finite time, space, perception, culture and neurology. Therefore, I believe human creativity is a process which can be modeled, replicated and mass produced just as human labor can. There's nothing special or magic about it, it's just a nut we haven't cracked yet. But then, there was a time when the concept of a machine doing math was unthinkable as well, and a machine doing cryptography, and a machine playing Jeopardy, etc.

Consider how formulaic much of modern popular culture is. Do you really think it's impossible for an AI, with access to the entire web and, theoretically, working neural models, to churn out an acceptable dime store novel or movie or video game? We're not talking about genius, here, we're talking about commerce. People will buy it. And if people will buy it coming from a machine, there's no longer any reason to pay humans to generate it.

It doesn't even really have to be better than what humans would produce - just acceptable.

I think the point the above posters were making is that people will continue to create for fun, even if machines are better at it.

This is exactly my point. Creativity is not a mindless endeavor in need of automation. It is the heart of human existence, no matter how finite your world view considers it.

I agree that creativity is as important to the human experience as you say. But that still doesn't mean the industry of creativity won't suffer the same drive for automation as every other, or that it can't be generated by machines. I don't think there's actually a contradiction here, people will still be creative, but more and more popular culture will be algorithmically generated - the jobs for creative work, and the support work, may not be there.

Although that does also suggest that human-generated creative works may become more valuable for their rarity.

Since the invention of mass media. most folks in the "creative industry" already work for next to nothing. Automation would only threaten the .1% of artists, musicians, and actors at the top. My band, for instance, barely makes any money, but we're in it for fun. Hatsune Miku won't put us out of work, but after a few more updates, she might replace Katy Perry.

@krapp

Here is an expert who agrees with you, and has the results to prove it:

http://vimeo.com/7441291

Being creative can also be doing things like this: http://hackaday.io/project/2021-3D%28ollar%29-Scanner

Yep. It's mine.

> I think, there will be a great depression, when even creativism becomes boring.

The entertainment industry has never been so big and it doesn't stop growing, so I guess the trend hints at the opposite. Electronic games weren't a thing 30 years ago, online games weren't a thing 10 years ago, both are now big industries.

I bet more in technology introducing new medias (e.g. Oculus) and aiding on better / cheaper / faster content creation than completely automating creative work.

You could pretty much eliminate all mcdonalds staff already with a vending machine that has a conveyor belt that runs under a heating element(like they do at quiznos), and sprays some condiments on a bun. All you'd need then is someone to replenish the vending machine.

So artificial intelligence isn't even necessary.

Has to be self-cleaning, sense outages, package up the food and bag it. Respond to disruptions (conveyor broke; ketchup dispenser leaking etc). AI could be useful.

Or just sensors? I doubt that the McDonalds grill is cleaned at all during the day. If that were the only issue, you could get someone in every night and clean it/restock it. You certainly don't need AI to bag a meal (or even just _I_ for that matter). I'm just trying to say that really a lot of the work can be replaced already but they haven't done it yet. Why?

I think you want to automate the whole 'food chain', end to end, to make it profitable. Else you have robots AND employees, which costs more than one or the other.

Still cheaper to pay someone for 1 hour to clean & re-stock then to pay them for 8 hours to cook, as long as the machinery costs & maintenance is cheaper in the long run.

You'd need supervising staff with real technical skills, whose wage is the clear barrier here, not the tech involved. A sufficiently high minimum wage might do the trick.

To supervise what exactly?

Automated factories exist but human-less do not (yet). Making a machine that works almost all the time and thus requires occasional human intervention is much easier that making one that can deal with all possible circumstances.

Considering cars are made by robots I always wondered why a big company like McDonalds couldn't do similar when making its cheeseburgers. Never researched it until now but looks like there is a machine capable: http://www.zerohedge.com/news/2014-01-12/meet-smart-restaura...

I look forward to completely automated fast food. Drive up, wave my RFID tag near a sensor, get a menu of last 5 thinks I ordered, touch screen, plop! Bag drops out chute, clean and hot and just the way I like it.

We're always hearing "there will be new kinds of jobs", but never what those jobs will be , how many they could be in number or any relevant detail to that claim.

Why is that ?have we no sense of imagination ? Or is it that the claim is just an easy comforting truism(that preservers the current order) and nothing but ?

Do you think people in Turing's time could have possibly envisioned the existence or proliferation of the role of "developer"?

Not saying you're right or wrong, but our inability to envision future jobs, or lack thereof, doesn't mean much one way or the other.

It's hard to see developer jobs (as awesome as they are) replacing displaced jobs on a massive scale. I was listening to an NPR podcast where they estimated that only one percent of workers in the United States are employed as software developers.

> Do you think people in Turing's time could have possibly envisioned the existence or proliferation of the role of "developer"?

Yes, I do. Before that there were already people that spent most of their time creating algorithms, and programming machines.

Perhaps there is net job creation but not for the people who are immediately displaced. The jobs figures are treated like some pool, and it ignores the fact that many people can't adapt their existing skill set to the jobs that were created.

For example, 100,000 truckers are put out of work by a truck that drives itself. There might be 100,000 other jobs created - we might all get to work less one day - but for those 100,000 truckers, the number of them becoming automation developers is probably pretty low.

It seems to me that the new jobs will require more and more education, which could be a problem for the uneducated. I am optimistic though, because when I look at the past people have always been able to adapt to the new jobs and benefited significantly. Just think of all the poor hunters and gatherers who were put out of work by the advent of farming.

I believe as the robotics industry grows, an appropriate support network will need to grow with it to handle the inevitable growing pains. For instance, I believe a "robotics mechanic" will be a job of the future, analogous to an automotive mechanic. Today there are experts/specialists in different automotive systems like transmissions, electrical, etc. Likewise, robots will need specialists in perception, locomotion, cognition, etc.

But robots of the future will not be designed to do one thing, as cars are (press throttle, move forward). They will be complex, autonomous decision-making machines, something the world really has never dealt with. I imagine the support infrastructure needed to deal with such machines would more closely resemble the healthcare industry than the automotive industry.

For example, you would have your highly trained specialists as I mentioned, but also generalists skilled at diagnosing and repairing common issues, and of course support staff needed to assist them in their jobs. Thus you have tiers of skill so that if a robot displaces a housekeeper(for instance) he doesn't have to get a PhD in order to find a new job in the robotics industry that made him obsolete.

Its unlikely that disruptive changes in the robotics situation will magically leave room for everyone, or even for most people, to still have meaningful work. We'll have to decide what to do about that. And it will take more than hopeful trickle-down theories. Probably a basic income or work program solution of some kind.

"robotics mechanic" - sounds like a job for fully automated robotic robot service centers.

This question intrigued me, so did some research and started designing "future-proof" business plans and brands that people can use right away.

It's still a work in progress, but it's here if anyone wants a look: http://startjumper.com.

It's simply a historical observation. If it was obvious what kind of jobs there would be in the future, companies would have already sprang up around those ideas and would be hiring, which would make them the jobs of today. Anything less would be equivalent to the whole world having an understanding that Apple stock will rise by 100% in the next two years but nobody buying it.

Basically, the future is unknowable. But as Marc Andressen likes to say, human needs and wants are unlimited. As long as people can provide for the wants and needs of others via jobs, there will be jobs.

I'd take it one step further, in fact. If we imagine the robot dominated future that so many people fear, it stands to reason (given human nature) that robot-produced goods and services will be viewed as inferior, cheaper, or less personal than human-produced goods. Robots and AI are by nature easily copyable, thus fungible. Sure, AI can be trained and tailored to specific needs, but competitors can simply copy the successful training methods once they are apparent.

A widespread perception of AI/robot produced goods/services as inferior or less personal will happen irrespective of the actual merits of of the products and services in question. There will probably be a few field-specific exceptions to this, but primarily in areas where exact precision, scalability, or replicability are the primary competitive advantages (precision machinery, automobiles, aircraft, etc).

In industries that are very cost or precision-sensitive, robotic automation will do well. In industries sensitive to quality of service, humans will still be useful.

Ordinary products such as food & clothing, and especially services, can be marketed based on a "human touch and taste" differentiating advantage that robot-based corporations will be unable to compete with. You can go into RoboDonalds and order your hamburger with all the ingredients you want, and that will be fine. Or you can go into a traditional steakhouse, talk with the chef about the cuts of meat available, select one, tell him/her how you'd like it prepared, and enjoy conversing with the waitstaff. Until AI can hold human-like conversations and understand all the nuances of human communication, any product or service that differentiates itself based on a non-trivial communication with the customer will be robust to AI/robotic disruption.

Yes, it's an historical observation. But it would be very interesting imagine/inquire how it could manifest in the future.And BTW - science fiction authors did predict something generally similar to the iPhone.

OK ,it's possible that robot produced goods will be perceived as inferior, but judging the popularity of today's hand made goods vs machines made goods it might not be a big factor - but who knows.

It's also possible that people would be interested in being serviced by a human and not a robot. It's interesting to think such a world, where more than 50% of your wages are purely going towards buying relationships - and most of the work time of people is purely selling relationships. It seems pretty lousy - but it's possible.

We can't say what kind of jobs there will because we are horrible at predicting the future. The most certain guide we have for the future is the past, and the past isn't as mysterious / scary as the future.

One thing is for damn sure though, we can be quite certain of the jobs a robot can do right now and that we are just waiting for price is right.

Accuracy shouldn't prevent us from telling stories about the future, stories that could inspire, and give us some direction about the possible future, even if some of them are wrong ,because some could be right.

And heck, i'm pretty sure there would be decent readership for such stories. But there aren't none as far as i can tell.

I don't get this issue. Let the robots have all the jobs.

Disruption of the economy. What if they had Your job? And you couldn't find another one.

The real question that intrigues me is why have a job in the first place? Obviously the current system derived from capitalism requires that we either work or own, but this doesn't make it the right path for mankind.

Reply on news.ycombinator.com