Hacker News new | comments | show | ask | jobs | submit login
If the Robots Come for Our Jobs, What Should the Government Do? (nytimes.com)
26 points by raleighm 5 months ago | hide | past | web | favorite | 87 comments



The article features three views. 1. A paper from the Roosevelt Institute [1] 2. the McKinsey Global Institute and 3. Michael Strain, a scholar at the conservative American Enterprise Institute.

Summarized ideas introduced in the article:

    * Allowing for inflation to tackle unemployment (from 1.)
    * Less monopole granting intellectual property laws (from 1.)
    * Work-sharing programs (from 1.)
    * Publicly funded higher education and training (from 1.)
    * Subsidized education and training (from 2.)
    * Making job benefits like health insurance and retirement funds more “portable,” (from 2.)
    * Expanding the earned-income tax credit (from 3.)
Interesting pre-assumption from [1] that might be dominant in economics but maybe not on the internet:

> This set of proposals is based on the idea that the emerging wave of digital disruption won’t result in a permanent loss of demand for workers, but rather shifts in what types of work the economy needs.

[1] http://rooseveltinstitute.org/dont-fear-robots/

[2] https://www.mckinsey.com/mgi/overview/2017-in-review/automat...


The article talks about how globalization caused automation to replace millions of factory workers. Factory jobs, aren't the greatest, especially before the 1970s, and in many parts of the world, are still awful. Similarly, If we could replace every mining job with robots we'd remove a lot of human abuse in many parts of the world.


You’d also remove the way for lots of people to make a living.

The unemployed only know that making a living in a lousy industry is better than not making any living at all.


That's exactly the point. Humans rely on jobs, what happens when robots take over?


We don't actually rely on jobs, we just force one another to work to satisfy our principles of fairness. In theory we could agree on a different notion of fairness that doesn't require everybody to work, we just choose not to.


FWIW, I think you may be under-selling "We just choose not to."

Human principles of fairness are based very heavily on social constructs of tribe and group survival that predate civilization by 10^5 to 10^7 years. Human thought is extremely malleable, I think the human psychology jury is still out on precisely how malleable "fairness" is.

That having been said, it feels like if we could start by ratcheting most people from a 40-hour week to a 30-hour week without cutting their pay, that'd square with the gut feeling of fairness. ;)


I agree that fairness will pose some short-run concerns. But, consider that it probably didn't make sense to early hunter-gatherers to work themselves to an early death in order to accumulate or bequeath hoards of any of the materials (raw or processed) or tools at their disposal. It seems certain they still had a sense of fairness, which suggests the triggers of that gut feeling are themselves quite malleable.

One legacy of agricultural and industrial production has been the devaluing of just about all "non-productive" activity. A broad re-valuation of what we spend our time on might be a good place to start re-calibrating our sense of fairness.


Yes, before industrialization, the US was filled with entrepreneurs in the form of farmers and ranchers. Here's some land, go live. Families would grow what they needed to survive, then sell the rest for money to buy stuff they couldn't make themselves. It was fairly self sustaining, but vulnerable to drought, etc.

I think one of the problems today is there really isn't a way to self sustain anymore with just land and some hard work, so we need jobs to survive.


What are some potential alternatives?


I think we'll find, if we stop hitting them with stupid-sticks in childhood (I'm looking at you organized educational systems) and feed them well that most people are in fact interesting, intelligent, curious, and driven to express themselves in glorious self-directed vibrant creative constructive expression. (And if some of them turn out not to be, well then, we can toss 'em in a volcano.)

A "job" is not something you need. You need to contribute to society and you need support from society as well, and we currently structure that relationship primarily through jobs and "the economy" but that structure is what's going away. The question is, what to replace it with?

Automation is a good thing. From the point of view of the machines Moore's Law means that humans are becoming exponentially more valuable! (Thanks to Mark Miller for that gem.)

Unless you can do something that a machine can't be programmed to do you have a serious problem, though. What will your children be? And their children? We are deciding today what our distant descendants will be doing.

An A.I. koan:

Q: What can you do that a machine cannot? A: Could a machine answer that question?


I do not believe we are at a special tipping point in this history of automation in the economy. Automation has been removing certain jobs from the economy for thousands of years. All-in-all (certainly in the 20-21st century in the US at least) unemployment has ebbed and flowed. When automation occurs in an industry it generally allows for greater production and cheaper products which raises the standard of living of everyone. There are many new jobs that exist today that didn't exist 100 years ago or at least were in very small number. Here's a nice podcast on the topic https://itunes.apple.com/us/podcast/episode-3-i-want-the-rob... :)


Why do we assume that robots will replace all existing work, or that the jobs being automated won't be replaced by other forms of work or income that don't rely on jobs that can be automated?

I think it's a bit sensationalist to assume that all work everywhere will be automated.

Even if we get to the point where that happens, capitalism as we know it would crumble if there is literally only automated jobs and robots owners.


Well, my reasoning goes like this:

1. Industrial revolution made obsolete physical work. The physical work that is done today (for instance, blue collar workers, waiters) is about mental capacities, not physical.

2. The brain is a physical object that process information (no magic).

3. At some point we will able to create machine that process information better than the human brain, making humans obsoletes.

If we accept that premises, the only question left is "when?".

>>"Even if we get to the point where that happens, capitalism as we know it would crumble if there is literally only automated jobs and robots owners."

I agree with that.


"capitalism as we know it would crumble if there is literally only automated jobs and robots owners."

Exactly! What happens after capitalism?


The way things currently are, I agree. But I think it's a bit restrictive to limit yourself to that, when you are projecting decades into the future anyway.


This is the reality ... it's happening. Industry will always optimize, and as technology marches forward, they will replace human workers with technology when it makes financial sense to do so.

The discussion should be, "what do we do when that happens"?


I think countries around the world are starting to realize that globalization while having lifted millions out of uttter poverty has now also turned its power to extract wealth toward those to whom it promised to bring greater opportunity which never came (I’m looking at you, Michael Moore’s castaway workers the industrialized world over; people who can no longer raise a family on a blue collar job because they were shipped over seas).


On a long enough timescale (assuming we don't nuke ourselves, or we choke on smog, first) ... total and all-encompassing globalization is the end-state. At that point, it will be no different than employers in the US moving from one state to another. Of course, that doesn't bode well considering we have problems even just in the US with some states having booming economies, and others having massively lagging economies ... but that's how it's gonna be


You can say that about every efficiency ever invented.


I've heard this expressed as the "Washing machine paradox."

Person comes home from a coalmine job, grumbling because there are rumors going around that the mine is going to buy a bunch of automatic conveyors and lay off half the miners. They come in through the basement entrance straight into the laundry room (people who've lived in coal country, you know how that works ;) ) and toss their grubby clothes directly into the washing machine. How are they gonna make ends meet if they get laid off, they wonder?

They do not for one second stop and consider that a scant hundred years ago, a person would have been the one doing that laundry the machine is doing right now.


>> If we could replace every mining job with robots we'd remove a lot of human abuse in many parts of the world.

Reducing the value of human labor never goes well. As people become less valuable they become easier to abuse. There will always be a dirtier and more dangerous job to be done.


Don't think of "employing" people, of making sure people have jobs. Instead, think of a society "spending" people.

We used to spend most of our people growing food. Then we got tractors and combines, and we could spend only 2% of our people growing food. We spent the newly-available people on factories - some to make tractors and combines, but more to make cars and dishwashers and tons of other things. It was an overall improvement to society - we could spend our people on more productive uses.

We need to do the same in this situation. We can automate factories. Great! We don't have to spend as many people making stuff. The trick is to find better things to spend our people on. If we can, those new things, being more useful to society than factory work, should also pay better than factory work.

But what are those "better" things? Aye, there's the rub...


This is a great way of thinking about the problem. And it's not hard to think of better jobs for people. There are a lot of unsolved problems in fundamental research across all fields of science.


True. But unfortunately, not all people laid off from factories are qualified to do fundamental research...


This problem is much deeper than just automation and has been spiraling out of control since at least the 1970s. https://thecurrentmoment.files.wordpress.com/2011/08/product...

Years of tax cuts for corporations and the rich combined with the destruction of Unions and workers ability to collectively bargain has created this problem. Start taxing wealth and corporate profits at post-ww2 levels, increase protections for workers. Encourage co-operatives, profit sharing, and other forms of worker control and ownership.


Except that this problem is strictly american problem that more/less requires nation-wide therapy from the red scare and acceptance of positive sides of socialism and how through grass-root pressure, trade unions and political activism, it actually shaped the western world for the better (an 8-hour workday was a blasphemy before socialism, as just one blatant example).

Automation OTOH, is a global problem that requires wider-reaching systemic changes. And while the western world will likely cope by inventing more white-collar bullshit make-believe jobs, people in developing world will likely be hit very hard.


The government’s job should be to keep an eye that gains are shared broadly enough, and ensure increased productivity doesn’t cause excess environmental harm. (If the government funds Robotics research, all citizens should share in the windfall)


Based on your comment I think you'll really enjoy this book - I'm about 75% through it. https://www.penguin.co.uk/books/280466/the-value-of-everythi... The Value of Everything by Mariana Mazzucato. One of the best economics books I've read in a very long time, well argued and not too heavy in Jargon.

The thrust of the argument is that we've misconstrued value and thus don't properly account for all the activity in the economy, overcounting things which aren't 'productive' in the historical economic sense (like finance, FB's ad market), while undercounting other activities, including those of govt - before everyone jumps on me, both Marx AND Smith agreed that finance didn't contribute to productivity, merely enabled it.


My understanding of economics is that capital and love are inputs to the productivity formula, but don’t raise it.

There are lots of externalities (pollution) that are hard to measure.


"all citizens should share in the windfall"

Maybe we all get assigned a robot as our "work" and we get the robot's pay.


Figuring out ownership of robots is interesting.


The simplest solution is an automation tax which is revenue neutral. Any time automation moves in to an industry, the tax's revenues can be used to help transition those workers into other work, or give them at least an automation relief stipend.

If we don't tax automation, what's to stop increasingly cheaper automation from replacing increasingly expensive first world workers? Especially when businesses are expected to increase salaries yearly, provide expensive benefits like healthcare - businesses would rather pay up-front for automation.


Another solution is to claw back mineral rights by taxing raw materials. You burn the sky, take the water, dig in the ground, that is something the citizens don’t get to use. You owe them.

The thing I like about this (its not my model but it’s my favorite) is that it happens at the input side so it codifies the idea of Reduce Reuse Recycle by setting a base cost where it’s cheaper to recycle and purify than it is to use virgin materials.


I like this idea, as long as it extends to all land use. So, people renting from others are actually receiving a dividend from the landlord, for the utility of owning that land.

I've read that land/material-based taxes instead of income-based could raise enough to fund a GBI, and be sort of organized like Alaska's oil fund.


Property taxes are a thing already. It funds infrastructure which we all need more when we are employed, but we all need it anyway.


I am wondering, why wouldn’t this logic apply to outsourcing, too?


How are automaton owning plutocrats supposed to get rich if they don't sell their robot made wares to the plebs? It's silly to think that automation makes the poor poorer. It makes them richer. Only logical conclusion is that stuff will be cheap enough that real wages go up even if nominal ones dont.Look at any rich country, it's got an order of magnitude more automation than any poorer country and yet it's not the country with less automation that is better off. Why do barbers in rich countries earn more than those in poorer countries for essentially the exact skills? Baumol's disease anyone? Luddites need to think through this harder.

PS: if the evil robot owners produce stuff and keep it for themselves without selling them, the rest of the world still has needs that are unmet and we end up in exactly the world we have today


Humans used to gather resources via hunter gathering, then Agriculture tech came along and automated that. Humans now gathered resources via farming. Then industrialization came along and automated Agriculture. Now we gather resources via 'jobs' instead of hunting gathering or agriculture.

How will humans gather resources after the 'job' method is automated?

Or will we even need to gather resources anymore and will be able to focus on art and scientific pursuits and advancing the human race?


I've heard universal basic income may be a viable approach, but I think it's unclear how it scales up as of yet.

Regardless, I haven't seen better suggestions yet.


What sold me on BI was the concept that every decade the GDP per capita goes up. But we don’t need that product as much which leads to consumerism. If one guy can produce all the stuff for five people what’s the point of the other four working? We could switch to an experience or service based economy, and/or we could give people a cushion that lets them try out a lot of different things.

Not everybody knows what they want to do at 9 years old like I did. I watch friends and family struggle their whole lives with this and recognize that I was very fortunate.

That said if I had health insurance and programmer culture was a bit different I would work 30 hours a week, get nearly as much done and have more time for family and hobbies.


Imagine how great society would be if humans could just focus on scientific and artistic endeavors instead of grinding it out for basic survival?

Think of how advanced the human race could get if no one was starving and people's attention could be focused on science and art and building cool new technology to better the human race.

One day people will look back at how we live as extremely primitive.


Who owns the robots?

http://calroc.blogspot.com/2012/12/blue-collar-coders.html

> In your lifetime the current economic paradigm will become as quaint and bizarre as a Renaissance Faire.

> The handwriting is on the wall, plain to see for all. With the rapid sweep and accelerating scope of technological development, most people will be left behind. Just sitting there you are becoming relatively more ignorant as a sort of inflation of the mind occurs all around you. There is no way you can keep up with it all, so how is a normal person going to do it?

> These are the options for normal people:

    1 Exterminate them.
    2 Enslave them.
    3 Pay them to stay home and play video games.
    4 Trick them into make-work and drudgery.
    5 Something else...


I like this article a lot except the confusion about unions. Without some kind of organization of workers (you don't have to call it a union) the people with money will own the computers and they will use them to do plans 1-4. A union is a way to harness the political will get the future described in the article.


Do guaranteed minimum income. Give everybody a computer and lots of free time. If 0.001% of those people become hobbyist software developers it will pay for the program in a year. And we will probably get AI in two.


The total US tax revenue in 2017 was $3.422 trillion.

If you split that evenly among 300 million people, each would get $11,000, which is half of the poverty line of $22,000.

Even if you thought that little money was acceptable, to afford it you'd have to cut:

1. all military programs

2. all existing social security programs

3. all government infrastructure projects

4. all education programs

5. ...


Is this an honest reply? The middle sees no real change in total income, the only the extremes see changes in total income.


My post is describing why minimum income would not result in everyone getting $50,000 a year.


Not that I would advocate for such extreme lack of incentives but US gdp per cap is over 50k. Doesnt this mean that you could in theory give everyone 50k?


Well, yes, it would require significantly increasing tax revenue to pay for it.


To get everyone up to the poverty line and not have to cut government projects, you'd have to triple tax revenue.

Since the average tax burden for those making $500k+ is around 30% already, that would effectively mean taxing them at 90%.

I don't think it's politically feasible.


I know enough to know that I don't know nearly enough about economics to say if it's fiscally feasible, my gut says probably so but I can't imagine how such a thing would even be modeled. I 100% agree it is not politically feasible. Even the poor folks in America think that employment is a moral issue and being poor is a reflection on your value to society.

I benefit from lack of competition in my field leading to significantly higher-than-average compensation, but I have to say that I'm intrigued by the idea of giving everyone the freedom to find something to do which meets their aptitude rather than wasting what might be a lot of really smart, capable people, or artists, etc, who have to piss away precious hours of their possibly productive time flipping burgers or something else equally valueless. It may not fit well into our current "meritocracy" ideal but I wonder if we wouldn't build a much better world if we upended our existing incentives a bit. Enough money for everyone to survive (with roommates and such) but little enough that there remains an incentive to go earn more.


You're only looking at personal income taxes. Payroll taxes account for about 2/3 as much government revenue and are currently one of the most regressive tax types. Removing the limits on maximum taxable earnings could bring in a lot more money.

And the other side of this is to try to bring down costs. Hopefully a basic income will encourage more regular doctor visits and fewer expensive emergencies. We should also be able to reduce some kinds of welfare spending if food and rent are being paid from the new basic income. And there are already plans underway to increase availability of cheaper housing.


What should the tax rate be on a robot?

If the tax rate is too high, then people won't replace jobs with robots. That's a negative feedback on the behavior we're worried about, so that's not a bad thing.


It still amazes me that so many people don't get this.


Another data point: GDP per capita is around $50.000.


Invest heavily in retraining programs, tax concentrated wealth (of individuals and companies) more deeply, and prevent mergers of already huge companies - especially mergers from the same industry.


Raise everyone up a level to make sure they can perform jobs that robots can't do, like get an advanced degree or incorporate a business. Make it a requirement for citizenship.


You run into diminishing returns pretty quickly that way. Having an advanced degree doesn't prevent a robot from taking your job. Maybe we should let the robots do the work and find something cool to do with our time instead.


We have computers doing diagnostics now.

Remember that problem with the glassblower? He had no one to replace him because there was no market for a journeyman chem glass blower so nobody followed behind. We will have the same problems elsewhere.


Start from first principles, jobs are costs. We wanna reduce costs. Also, exports are costs, imports are the benefits. It is counterintuitive to most people, however


Robots don't come for government jobs?


Solution: the government should hire everybody. They've already got a good start with about 2% of all workers directly employed and another 2% working for contractors. Add state and local government with another 8% of the working population and we're well on our way to full employment. I'm kind of joking but it's interesting to think of job creation projects like WPA taken to a further extreme.


And despite all of these concerns, the NYT frequently and consistently advocates for increases in low-skill immigrants, the very people who will the government will need to support when the 'robots come for [their] jobs'. The question is which of those 2 arguments are disingenuous?


That's not the case at all. It's very cost-ineffective to automate low-skill jobs because the human capital has historically been extremely cheap. There's a reason we don't broadly have robots cleaning our houses, landscaping our yards, providing in-home care, picking fresh produce, or providing general construction labor.

It's the precisely the lack of these low-skill workers that will drive the necessity of automation, because the cost of hiring people will become too expensive.

Quartz published a well-research article last year on the same topic: "Low-skill workers aren’t actually the ones most threatened by robots" https://qz.com/1010831/the-middle-skill-job-is-disappearing-...


Do you think it's possible that the NYT has authors with different perspectives and that every article written may not come from a consistent viewpoint?


The strongest argument for increased immigration is that most if not all of the western world has a declining population. Also its also not yet clear what types of jobs will be easy to automate. For example I think an assistant nurse would be very for a robot to do. A radiologist less so.


I don't think either are disingenuous. It's more so short-term vs long-term thinking. The influx of low-skilled immigrant labor is a "right now" issue whereas The level of automation needed to make them obsolete is optimistically 10-15 years away.


Make robots pay the equivalent income tax as the people they displaced.


Build a big, beautiful wall.


Robots are doing that too.


The government should do precisely NOTHING.

The jobs don't belong to workers, so the premise is entirely wrong to begin with.


Right, but the country these companies operate in, the infrastructure they rely on, the laws that allow capitalism to succeed, and the battalions needed to protect it does belong to the citizens, so ignore the citizenry at your peril.


This is correct. It's not the government's job to provide you a job, it's a personal responsibility to find one, if you want one. If you seek one, the government might provide some minor worker abuse and discrimination protections, but this in no way implies you have entitlement to a job or even income.


So you're saying that even though the government promised us that exalted "job creators" deserved tax cut to create jobs and increase wages; and even though those same corporations took those tax cuts and paid themselves and bought robots with it instead of meaningfully boosting wages (where is my $4k every worker was promised?); and even though most of that money stayed at the top; and even though we'll be paying off that money far into the future; and even though the government continues to erode worker's rights; and even though this system has created people who have more money than they could possibly spend in 10 million lifetimes while others are starving and homeless; you're telling me that despite all that, it's our own damn responsibility to find a job, and the government plays no role here in a future where the game was rigged the entire time so that capital owners could control everything?


Strictly speaking, you're "right" ... however, it belies an incredibly inhumane approach to life. Although I concede that reality doesn't always reflect intention, _my intention_ for government is to provide services and protections for all of us; every citizen.

While yes, that shouldn't mean one is necessarily entitled to anything, I want my tax funds going towards helping people who have unforeseen things happen to them. If they lose a job, or get sick, or their head-of-household/primary income passes away, or they are the victim of abuse ... I want my funds to help those people.


We're talking about a world where automation has consumed say 20%+ of jobs. You presumably don't want to let these newly unemployed people starve, so what's to be done? It's naive to think that continued rising mass unemployment while a small amount of robot owners get richer and richer is a stable model of society. The automated economy would end up in public hands soon enough, one way or another.


>>" It's not the government's job to provide you a job"

Maybe now things are how you say, but it's not a divine mandate.

If you see the government as a tool of society, it's the society who should choose the role of the government through the democratic process.


The government should do precisely NOTHING.

Exactly. It is not the job of the State to make sure everybody has a job. Arguably all the State should do is protect property rights, enforce contracts and provide some notion of a "justice system". And really, all three of those things can be debated.

The best way to think of it, IMO, is the way Bastiat[1] put it by describing "the law" as "no more than the collective extension to our individual right to self defense". (paraphrased slightly).

[1]: http://bastiat.org/en/the_law.html


The job of the state, ultimately, is to preserve itself. In practice, that means upholding a social contract in order to maintain order. If we see a widespread outbreak of technological unemployment then more effort on the part of government is necessary lest public order breaks down.

It's that simple. No moral arguments are necessary.


Widespread technological unemployment is definitely a concern, no question there. I'm just not convinced that government is the best tool to try and solve that problem, on both moral and practical grounds. But when I say "government" I really mean "Government as we know it today" - big, corrupt, invasive, inefficient, etc. Possibly a better model could be developed which would be better suited. That, to me, is the challenge for our time.


I don't see government corruption as being wholly determined by models. Models are a factor, no doubt, and you can affect the rate at which corruption encroaches by changing the model, but time and complexity are also major factors.

Within a given model, complexity tends to increase with time and thus so does corruption. Perhaps another way to look at it is just to consider corruption as one form of complexity. Eventually complexity increases beyond the point of diminishing returns into negative territory. Then the system collapses. This is the position of Joseph Tainter, one I find particularly compelling.


The problem is that when you have a recently minted, largely unemployed underclass who are not having their basic needs met, you frequently get rebellion. At that point the state's ability to enforce contracts, protect property, and dole out justice becomes severely hampered.

The best way to think of it, IMO, is the way Marx described the State which is a group built to protect the bourgeoisie.

It should (will? could?) eventually reach a point in which the State represents a social contract between all members of a civil society to work in the group's overall best interests.


The best way to think of it, IMO, is the way Marx described the State which is a group built to protect the bourgeoisie.

It's funny, I've got a whole pile of books by, and on, Marx, piled up waiting to be read (even though I'm a Libertarian in American terms. "know thy enemy" and all that) but finding time to get to that stuff is always the challenge.

I did not realize that Marx described the State as such. That is a somewhat interesting take on things.


The basics of the most Marxism and Conflict Theory are that 1- There is an imbalance in the availability of goods/services relative to needs/desires (capitalists call this "supply and demand"). 2- Society is built in such a way that those bourgeoisie are rewarded unfairly for their contribution (ownership and capital) relative to the contributions of the proletariat(physical creation of goods, performance of service) 3- The current state exists in ways that allow for that status quo to continue, through the very existence of a "State Class" (politicians) who act on behalf of the Owner Class but create the illusion of a second group acting against the workers thus dividing their attention


The state not only protect property rights, it defines them.


>If the Robots Come for Our Jobs, What Should the Government Do?

Nothing


Would you answer this question differently: If the Robots Come for Our Jobs, What Should Society Do?




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: