Robot automation will 'take 800M jobs by 2030' (2017)
bbc.com
bbc.com
Humans are flexible. Machines typically aren't. The more automation exists in a facility, the less flexible it becomes.
Until we hit AGI, I think there is a cap on how automated a facility should (not can) become.
A critical piece of this is that highly automated systems are expensive, and difficult to upgrade in a piecemeal fashion. Thus, using humans to tie systems together limits the amount invested in automation, and can save you money in the long run. US manufacturers are already aware of this, as some are dealing with the effects of aging overly automated production lines.
All of this depends on what you're making and what quantity, but until we reach AGI, over-automating is a mistake manufacturers have made and will make again.
In the US and similar, we've automated about as much as we could given existing technology. Newly available technology raises the automation cap a bit, but not that much.
Contrast this with nations which are just getting to the point where mass automation is available and have a spiking per capita income. They're much more likely to see massive amounts of automation, proportional to their increasing wages.
When the majority of the population is essentially living extremely efficiently on a vegetarian diet without so much as personal transport, one should expect very little economic shock from issues as remote and abstract as gradual technology change. In addition, in many parts of India electricity is not yet reliable enough and corruption too endemic to allow the confident adoption of industrial automation.
highly automated systems are expensive, and difficult to upgrade in a piecemeal fashion
Highly automated systems are cheaper and more capable than ever, and getting cheaper. They are easy to justify on a budget because they pay for themselves in saved wages, 24x7x365 operation, and reduced error. Two examples are the dozens of companies now producing (1) industrial robot arms; and (2) warehouse goods transport robots at scale in China.
You can design automated systems for modular upgrade, it just means accepting the costs for that up front. Often it doesn't make sense to do so, because by the time it matters you can probably design a new facility to pilot new off-the-shelf technology before retroactively updating prior sites. I believe major western companies use this approach for iterating data center designs, for example.
In the US and similar, we've automated about as much as we could given existing technology. Newly available technology raises the automation cap a bit, but not that much.
Strongly disagree. Just a few examples of areas the US has dropped the ball on despite access to technology would be: mobile payment, sustainable energy, transport, agriculture, health care, retail banking. From outside this often seems to be the case because the US domestic macro-economy is against change for reasons of monopoly, litigation, regulatory configuration, educational focus or similar.
It seems that the technologies and social configuration being deployed in China today are likely to be the forebears of living arrangements for the majority of urban populations globally in the near future: package dropboxes (HiveBox), mobile payment (WeChat), high speed rail, nationalized critical infrastructure, etc. This is largely because they have, versus the US, what Tony Fadell recently described as "last mover advantage" and a population comfortable with socially mandated solutions. https://overcast.fm/+Lkc1bqy8c
That does not make sense...
The phrase doesn't make sense logically. You don't have to be an avid consumer of technology to be affected by gradual technology change.
The further I penetrate into the detailed history of the achievements of Chinese science and technology before the time when, like all other ethnic cultural rivers, they flowed into the ocean of modern science, the more convinced I become that the cause for the breakthrough (occurring only in Europe) was connected with the special social, intellectual and economic conditions prevailing there at the Renaissance, and can never be explained by any deficiencies either of the Chinese mind or of the Chinese intellectual and philosophical tradition. In many ways this tradition was much more congruent with modern science than was the world-outlook of Christendom. - Joseph Needham, page 2, Part 2 (Science and Society in East and West: General Conclusions and Reflections), Volume 7, of Needham's Science and Civilisation in China (2004)
This posits the question: are the social, intellectual and economic conditions prevailing in the US the root of its problems? Perhaps that swamp should be drained first.
There are a large number of decisions to weigh when automating. Wages are only a piece of this, and 24/7 operation isn't useful in all scenarios. Again, it depends on what you're making and how much of it.
Each facility is limited by its own constraints on production. Lately, at most companies I work with this has been a shortage of skilled labor. They want more smart humans right now, not more robot arms. You need even more skilled and expensive employees to set up a robotics cell than to run a typical automated facility. When your robot guy leaves and you lose all ability to adjust how the cell works, you end up in trouble. I've seen this happen, and it's an example of over-automating.
My knowledge of India is basically non-existent, but I've heard (on HN) that companies will make up for a lack of skilled labor by hiring a large number of unskilled workers. I don't know if this is actually true. As per capita income appears to be rising there, my theory was that this is an environment ready to be hit by the same wave of automation the US has already seen. While my guess on India and similar countries may be way off base, my opinions on the US come from direct experience and I would stand by those.
Which is why India is banning things like self driving cars. https://www.bbc.com/news/technology-40716296
There are also a range of things that might have to be done. Including taxing agricultural returns, defragmenting land ownerships, and changing nonsensical laws like letting only the farmers progeny be farmers. Then of course letting import of equipment to industrialize food production etc etc.
But India being India, I expect these things to happen at the very last moment. Like when everything is about to melt.
https://www.indiatimes.com/news/india/new-vehicle-will-soon-...
This is what a minister in the Central Government has to say: https://www.hindustantimes.com/india-news/there-are-no-jobs-...
This situation in India has been pretty much anticipated since the late 1980s.
I'm guessing eventually India will pass laws and enact reforms to increase agricultural productivity and some kind of UBI scheme on the longer run.
http://www.wtec.org/loyola/nano/NSET.Societal.Implications/n...
But now the machines are replacing service and mental labor. There is no where for a human to shift to this time.
As far as I know, computers on their own are not very good engineers, programmers, physicians, etc.
I also believe creativity in general is a kind of unsolved problem. The art/music computers made so far ... not worth mentioning. (except as awesome research projects)
The OP didn't say all jobs by 2030, it said 800M. Obviously there will still be lots of jobs, but not enough. Too many people unemployed and broke, combined with not enough of a social safety net for food, housing, and medical care, will result in social unrest and, if that's not dealt with well enough, it will progress to social breakdown, which either means a revolution (to institute a new government that does implement appropriate safety nets) or a martial-law type of scenario, neither of which is pretty.
Below-average doctors, engineers, and lawyers will be at risk of being out of work. On-site blue collar jobs are safer, as a percentage. Robots can prefab things in factories, but it will be a long time before automated robots can build or repair anything on-site with the efficiency of humans + power tools.
ETA:
I think you overestimate the number of people required to handle most of the problems you think need work. And the current free market doesn't agree that they need more work, or it would price those jobs higher and more people would jump into those labor pools. In some cases, the jobs require serious specialization (maybe not advanced degrees, but a fair amount of training) and not just anyone can do them... in which case even the optimal labor pool size would barely scratch the surface of the coming AI-caused unemployment crisis.
Even a massive public works effort to employ tens of millions of workers to deal with aging infrastructure, or to build green tech, is not a permanent solution. Most of those jobs would be gone in a decade, and then you're right back to where you started, only worse because AI will have gotten better.
Massive government spending (including on temp public works projects) is essentially UBI for a lot of able-bodied people. I don't object to that, but it needs to be limited to public works projects that are really necessary, and then you have to figure out where you're going to get the money to pay for them when the govt is already running trillion-dollar deficits. Is 5% debt inflation per year not good enough? Should we try 10%? 15%?
Thats the point. Robots are and will be for quite some time if not forever, pretty dumb amd needs human supervision for any complex task that cannot be encapsulated.
Also .. with climate change etc. etc. ... there is so much needed work to be done.
And population gets older ... and needs much more medical therapeuts, physicians etc.
So much pollution and garbage.
So much ugly things all around.
There is an endless amount of needed work around. And if the current economic system does not adjust for that, then there is something very wrong with that and not that robots doing shitty jobs.
Thats my point. I do not consider the current markets very free. Very distorted by many factors, but that is another topic.
"Most of those jobs would be gone in a decade, and then you're right back to where you started, "
That is probably the most optimistic asumption about what it takes to change the whole energy base of our current economy, I heard so far.
Anyway, even if: Space for example is huge. I see some (infinite) potential there. And on infinite other projects to make this world more beautiful instead of the opposite.
It is hard to replace someone like that, but you can make their work meaningless. You could e.g. make electrical wiring that is trivial to install and self-diagnoses. So now it is "Uber Electrician".
Instead, perhaps UBI makes better sense. I know the economics part won't make sense until we're past the need for the current system, but if we don't need people to work, why have them waste their lives with suboptimal labor.
My biggest issue with UBI is that you'll probably just get millions watching Netflix all day. What would be awesome is if those millions were writing books, doing physics, or creating works of art instead. You'd have machines doing the labor and humans being freed up to focus on new contributions to society.
Of course, not to be elitist, but I'm not sure how many truck drivers will pick up work on the grand unified theory of physics, but if even one makes a new and important discovery that would be amazing. Also, I wonder how long until even jobs held by the highly educated (healthcare...etc) are largely automated?
There is a really easy solution to this that I have never seen suggested. Instead of giving money to everyone for free, why not raise the budgets of research and art institutions? You can get as many people as you want doing physics and making art as you want, all you have to do is give money to CERN.
Now, a pessimist might say that some people can't advance science or art because they lack the capability. I counter with the fact that anyone can do it at some speed, some are just faster than others. If you greatly increased the amount of money the Institute for Advanced Study had, they would be able to bring people on to their staff that were less efficient on a dollars per fundamental insight scale, I.E. you or me.
Is it a good idea to educate people who gets squeezed out of society? Sure. But most societies are going the opposite way even now. Like many other things it isn't a technical issue, but a social one. There isn't really a need to wait for society to collapse before giving people affordable access to housing, education, health care etc.
At the same time, it is worthwhile to decouple the problem of power from other problems - so that we can think clearly about how we might go about controlling power. This opens the possibility for solutions which would involve funding the arts and sciences. If we don't think clearly about both problems, we'll never be able to solve either.
If power can be made to represent community values, then it becomes possible to use Machine Learning and Robotics to provide all the basics for everyone. And it becomes possible to work on fixing the sustainability problem. So each of these needs to be examined and potential solutions to one impact the other. Things don't have to be the way that they currently are - we live in a time when extensive automation is possible for providing "basics".
Simplified, basics might cost $500 a month, but to live in New York you are paying maybe $5000 a month. The income isn't the problem, the cost of goods isn't the problem and the premium is the problem.
It doesn't really matter which way you do things unless you can remove the premium on success, progress, prosperity or whatever you want to call it. And if we do remove the premium we don't necessarily need these esoteric solutions.
Most, or at least many, people today already have money, it just doesn't go very far. So how is giving people a small amount of money going to change anything? It probably isn't, unless there is social change. Which the lack of is therefor the problem, not providing income as such.
For the record I do think a mixed market economy that keeps the cost of living in check and taxes automation is the most obvious answer. But Sweden already sort of tried that in the 1970's. Unsurprisingly very unpopular.
https://www.jacobinmag.com/2017/08/sweden-social-democracy-m...
> If machines produce everything we need, the outcome will depend on how things are distributed. Everyone can enjoy a life of luxurious leisure if the machine-produced wealth is shared, or most people can end up miserably poor if the machine-owners successfully lobby against wealth redistribution. So far, the trend seems to be toward the second option, with technology driving ever-increasing inequality.
-- Stephen Hawking ( https://www.reddit.com/r/science/comments/3nyn5i/science_ama... )
> The frightening coincidence of the modern population explosion with the discovery of technical devices that, through automation, will make large sections of the population 'superfluous' even in terms of labor, and that, through nuclear energy, make it possible to deal with this twofold threat by the use of instruments beside which Hitler's gassing installations look like an evil child's fumbling toys, should be enough to make us tremble.
-- Hannah Arendt
Another counter is that this may be caused by our current educational system. With UBI, it may make sense to employ e.g. 20-30% of all workers in education, to enable 1:1 tutoring for all students.
This also seems to ignore that research science itself is increasingly becoming automated, and such positions that are less efficient on the dollar per fundamental insight scale may disappear at around the same time as our hypothetical truck driver's.
Personally, I think that we as a culture need to change our thought processes on the necessity of everyone doing work. So what if some people just end up in the sitting around watching Netflix all the time category? Not everyone's going to do that, there are a ton of different outlets I would pursue if I didn't have to worry about working or money. And they're mostly things that I'm only money limited on because if I didn't work and pursued these things then I wouldn't be able to eat or have a place to sleep.
Work on giving people quality educations and the open-ended opportunities to explore, play, and pursue, and I think we'd all be surprised at what people will end up doing with their time.
They could work on automating their own jobs like software developers and engineers do. Or research for adverse effects of self-driving.
If you are already willing to pay someone's living expenses for the rest of their life, then there's no such thing as a gap due to not being remotely qualified. No how many years it takes for them to learn how to do research, if you are willing to pay the UBI, you are willing to pay their stipend for this time.
>such positions that are less efficient on the dollar per fundamental insight scale may disappear at around the same time as our hypothetical truck driver's
I think this rests on an incorrect view of human nature. People aren't born with a list of jobs that they can fit in to, they just have aptitudes for various things. In a purely capitalistic society, the aptitudes translate into a list of jobs because profit-seeking enterprises will not pay you more to do something than the net dollars you bring in - which is itself determined by your aptitude at that specific task. However, once we're talking about the UBI, this vanishes.
There does not exist a person that is capable of exactly washing beakers and driving trucks. This only appears to be the case because there do exist people who are capable exactly of turning a profit while washing beakers or driving trucks. There is a disturbing view of human capabilities running though this thread, that some people are just truck drivers "by nature," and that like the trucks themselves they must be retired if we no longer need their services. However, this "fact" is only a result of the financial realities that are specifically overturned by the UBI.
The amount of intelligence that constitutes gifted is an accident of history and economics. You can't divide people into those who can advance human knowledge and those who cannot, because everyone has some capability to do it. Presently there appears to be a "cut line" of competence that separates intellectuals from regular people, but that only exists because of economic forces that require us to limit the number of professional intellectuals. If budgets contracted the smart-ness of the average intellectual would rise as the lower performers were cut, and if budgets were expanded the smart-ness of the average intellectual would fall as hiring dug deeper into the baseline human population.
That wouldn’t be too different from the system today :( It would be a different set of people, sure. But it would actually be a net positive if this subset is smaller than the current subset (of people who don’t want to do x). I’d expect the number of people who are able to find their niche in arts, science, philosophy,etc would be quite large.
> Prosperity Comes From Eliminating Jobs, Not Saving Them
Other automated changes for manual jobs would require substantial changes to infrastructure. For example, refuse collection using underground tubes to suck away rubbish is already a reality (e.g. Envac system designed in Sweden) but it's unlikely to be retro-fitted to existing housing developments. On the other hand, things like housing construction and crop picking I can imagine will continue to see improvements in automation.
Robots right now are bulky and incredibly expensive devices which don’t even come close to human dexterity, intelligence or flexibility.
Look at self driving cars. First DARPA driving challenge happened in 2004 and self driving cars are still very dumb and have killed a few people on the road. Simulation hours don’t count. Real driving hours without human engagement count.
Sure in soft intelligence like speech recognition, language translation, vision we’ve done big advances. Deep Surveillance is going to be a real thing where every individual/car is tracked through the city.
But replacing a ~billion jobs with robots. Not going to happen.
Unless we have huge advances in bio manufacturing and computing. Doing it like DNA does it. Building machines from the same fabric we’re made of. Grow the machines from watery goo at scale. Make artificial muscles that are strong, fast, accurate, light and quiet.
Now that would be the stuff of science fiction brought to reality.
Work is one of the major sources of purpose people have. Earning an income and putting food on the table is fundamental to one's self-respect. Like educational achievement, it gives the impression that there is some way to advance beyond your lot. Take that away and it will cause a mass existential crisis unless there is something to fill the void. All the blights of the rust belt and other economically overlooked areas should be expected to spread.
Eh, that's a wild assumption. If anything, automating away crap labour that barely keeps you sheltered and giving people the economic stability to focus on something long term should enable various forms of self-advancement for masses of people who are now just stuck doing a crap job until they get to retire.
Such as?
Jobs that humans do are being altered as a result focusing on more cognative and social tasks.
Retooling / reskilling policies should be getting more attention.
Capital is making more money than labor. The 'hard working man' ethos makes less and less sense. You can't work your way out of poverty anymore.
No sane person is arguing for full-blown communism here, but a shift towards a more socialist system is going to become the only way to keep society from breaking down. We need the benefits of this revolution to benefit all and not just those who were born wealthy.
And it's going to be a very long road to get there, as political power today in most nations is most easily accessible to those with wealth, those who would lose out if such a shift were to take place.
Politically, I'm fairly conservative/libertarian, and I generally accept the economic realities of things like creative destruction. And even with those overarching views, I am terrified of the social impact of autonomous vehicles.
When you take a step back, it is nothing short of disturbing how aggressively and proudly Google and Uber (et al) are pursuing the destruction of the livelihoods of those 3.5 million households in this country. Just a bunch of mostly honest, hard-working folks who live in the exurbs or rural towns, with absolutely no backup plan when a robotic truck either creates tremendous wage deflation or takes their job completely.
And for what? So we can all save a few cents at Target. So a couple billionaire founders can toss some more on the pile. So a handful of engineers can make more millions. So paper-pushing fund managers can make a few more bucks from their stock going up. It's like watching a horror movie in slow motion.
Creative destruction has been a fact of life for centuries, but overall, societies have thrived. My sense - and I have no data to back this up - is that we are approaching a tipping point where the gains from technology are accelerating way beyond society's ability to reallocate those resources to other productive places in the short and even medium-term. That steady-state of frictional unemployment caused by technology is about to be majorly upended, and it just scares me. Societal problems are going to get very, very bad.