These discussions are always focused on the "robots". What about the spreadsheets, the computers, algorithms, industrial control systems, Python scripts, voice recognition systems, data pipelines, software powered gig economy, algorithmic traders? These types of automation have been in place for decades, and all of them cut costs for the owners of them and reduce the number of workers needed. Many of them can be operated with minimal training.<p>The benefits of those efficiency improvements are already going to the owners of the capital, not to workers, and this has been going on for decades. If in 1980 you were able to replace 4 clerks handwriting transactions in an accounting ledger with one clerk typing them into a spreadsheet but pay that one clerk marginally more than a handwriting clerk, who benefits most in that equation?<p>We're always focused on the some day maybe robots, but automation has been changing society and (I'd argue) increasing inequality for decades, and we've been completely blind to it because we're always talking about "the robots that are coming to take our jobs". We already need to find ways to distribute the benefits of these efficiency improvements more equally. In fact, we needed to years ago.<p>I wouldn't be surprised if this explains a lot of the productivity-pay gap [1]<p>[1] <a href="https://www.epi.org/productivity-pay-gap/" rel="nofollow">https://www.epi.org/productivity-pay-gap/</a>
There's a local burger joint with two screens where customers can ring up their own order. There's always a line at the register and never at the screens. I've walked in, keyed in my order, and received my food faster than people in line. And yet people just don't seem to want to use them. Yesterday I walked in and both screens had an "Out of order" sticky note on it. Minimum wage jobs have a few more years of life left.
>The gloomy narrative, which says that an invasion of job-killing robots is just around the corner, has for decades had an extraordinary hold on the popular imagination.<p>Thanks to the likes of the Economist.<p>Their story on how copilot will make programmers redundant exemplified the way they framed automation - as an answer to their executive readership's prayers - a deus ex machina that will let them fire half of their employees while maintaining their margins.<p>To be fair, there is a litany of trashy economics studies thumping the same drum. E.g. there was a famous one from Ball State U that mathematically conflated automation and outsourcing to prove that the robots already took over and one from Harvard, I think, that measured job automatability through surveys about how "creative" workers thought their jobs were.
Oh my. Econonmists<p>>... a solid microeconomic foundation. ...more profitable and thus expand, leading to a hiring spree. Technology might also allow firms to move into new areas, or to focus on products and services that are more labour-intensive.<p>In my time spent studying economics, this is the sort of thing that micro-economists say all the time and it is not true.<p>The untrue bit is "solid foundation". It is post hock story told to give a quasi theoretical basis for an observation. The article, bless them, acknowledges separating causation and correlation is tricky.<p>Solid foundation, pfft!
Isn't the real development that automation and efficiency improvements move more and more people into bullshit jobs? So instead of becoming entirely obsolete, we continue working, not knowing what value we contribute, if any.
That's why my intuition (and fear) is that in such BS jobs, the rewards will become more and more skewed. So instead of a 9-5 corporate BS job, one gets a highly competitive BS job (evaluated using metrics that are poor predictors and assessors of individual quality of work) in a high-pressure environment.
My take on it is that it is too early to tell. The automation envisaged has not arrived, yet. (Maybe it will take a long while.) Self driving trucks or book keeping systems without data entry are still not widespread.<p>At the same time, automation sometimes produces other results. Not many executives in medium sized companies have secretaries, rather the type their own messages, on a phone.
Jobs are not a necessity. People work because of they need money, country need people to work because they can tax and use the remaining to buy weapons. We don't need jobs, we just need hobbies. I am very optimistic about robots. It will be the time when there will be no slaves in this world, everyone will be truly free.
I continue to believe in the idea that automation/AI will destroy entire sectors currently requiring labor. It is something that happens gradually, then suddenly.<p>I also believe there's "safe" sectors where this is unlikely to ever happen. Good luck making a robot plumber. By the time you solve that, you've solved everything and we're entirely obsolete.<p>Package delivery, cash registry, retail jobs...there's plenty of high employment sectors at danger. And perhaps many office jobs are at even more danger.<p>One idea is that "as always" we simply move on to "higher order" work. This is nonsense. We don't need billions of AI programmers, it would be really crappy AI if that was the case. Nor is it realistic that everybody can do such a demanding job. We're talking about billions of relatively lowly educated workers here.<p>Another idea is UBI, or even better...post scarcity. For the optimists, I guess.<p>I expect that neither will happen, instead we're already experiencing the solution. You just inject trillions of stimulation money into the economy, which creates (artificial) demand, hence allowing for jobs that otherwise might not exist.<p>Right now, somebody is inventing potato chips flavor #31,122. Nobody asked for it, the world doesn't need it, and I hope we can agree that it's unimportant. But somebody invents it anyway, somebody designs the packaging, and somebody ships it. You buy it, and might just like it.<p>All these activities aren't really demand or need driven, they're more like "because we can" jobs. Our world is largely supply and marketing driven, not demand driven from the bottom up. Fluff and endless meaningless luxuries. When I walk through a mall I wonder...do people really buy all this garbage?<p>Put harshly, you can call all of this bullshit jobs. Keeping each other busy. And that's what we'll continue to do no matter the state of technology. It's just a few dials in the financial system.<p>Have you never wondered why everybody deeply cares that one has a job (at any and all costs) yet there's never any serious inquiry about the point of the job? Nobody cares if you have a meaningless job, the point is that you have a job. Any job. This reality gives away how the system works.
The thesis of this article seems to be that robots won't eliminate jobs, but it really doesn't adduce evidence to support that:<p>> a recent paper ... put forward a “new view” of robots, saying that “the direct effect of automation may be to increase employment at the firm level, not to reduce it.”<p>"At the firm level" is quite the caveat! If a company can produce two widgets per worker instead of one by adding robots, it'll succeed and grow as a result - at the expense of its competitors. But if every widget maker adds those robots, and there's no more demand for widgets than before, a lot of factory staff will be out of work.<p>They touch on a related point near the end:<p>> Mr Aghion and his colleagues add that even if automation boosts employment at the level of the firm or industry, the effect across the economy as a whole is less clear. In theory robot-adopting companies could be so successful that they drive competitors out of business, reducing the total number of available jobs.<p>...but it seems to me like a major problem with the whole framing of the article, not just a "by the way".
What drives the adoption of automation is not "labour is expensive" but "labour is unavailable", "labour is too slow to obtain/train", "labour is imprecise", "labour is dangerous", "labour is untrustworthy" or "labour is unreliable".<p>Generic industrial automation solutions are still very expensive to integrate unless simplistic. Furthermore, if you automate one unit operation, the rest of your process line also needs automating or the adjacent operations will simply become the new choke points.
This is missing the obligatory reference to Bertrand Russell who said in 1932 that working hours will reduce to 15 hours/week [1] and this is good.<p>[1] <a href="https://harpers.org/archive/1932/10/in-praise-of-idleness/" rel="nofollow">https://harpers.org/archive/1932/10/in-praise-of-idleness/</a>
>The gloomy narrative, which says that an invasion of job-killing robots is just around the corner, has for decades had an extraordinary hold on the popular imagination.<p>This "gloomy narrative" first and foremost served well for decades to depress wages and wage expectations and ensure a considerable supply of cheap labor exists.<p>The political argument was even spoken out quite freely from time to time, and it went like "well, if you [politicians] don't ensure labor is cheap, we [rich corporate owners] will simply automate away all the jobs and your voters will be left unemployed". The narrative of this development seemingly being "inevitable" and the necessary tech always being "just around the corner" for the automation of practically everything played so well into this, that it leads me to assume that interested circles knowingly pushed this narrative for exactly these political reasons. And a lot of technologists helped hugely in this push, often not because of political or economical gains (though quite some probably own tech stocks or work in the tech sector and thus stand to benefit potentially, but I also know tech enthusiasts for which these both don't apply) but because of tech fascination.<p>The truth however is that automation is a hard and costly business with clear limitations, and even many simple jobs are probably not automatable for decades to come. Automation often just helps to make certain jobs more efficient, but the efficiency gain is immediately being consumed by extensions in service and/or complexity that would not have been possible before, but leave us with a net zero in the change for the need for human employees (or even an increase!). But that did not stop the automation zealots from pushing their political agenda, because it's just too helpful in ensuring that most of the monetary value of the efficiency gains ends up with them and their shareholders instead of the human workforce.<p>So the newest paragraph in the book is "AI will automate away even all the complicated, brainy work!" and a surprising number of otherwise smart people fall for the same trap that has already been successfully used to "manage" the lower-paid untrained workforce for decades.
Automation will probably hit information workers before it hits manual labor workers. The reason is that information workers are expensive because it's harder to outsource to a poorer country, and we've got a long way to go before robots get good enough to compete against child labor (joking).
I cringe when I see an analysis begin with a trite conflict of interest argument like "Warning people of a jobless future has, ironically enough, created plenty of employment for ambitious public intellectuals looking for a book deal or a speaking opportunity." Yawn - that's charmingly glib, if factually dubious, public intellectualism is not a growth industry in general, let alone for bot-doubters.<p>1: <a href="https://archive.md/L6TDA" rel="nofollow">https://archive.md/L6TDA</a>
Slightly off topic, but unemployment rate is one of those all but meaningless metrics that politicians and economists love to parade around (as meaningful). For example, you can be ineligible for unemployment (i.e., you are employed) but your income is so low you still get public assistance. Or you're living with 20+ other people.<p>It's also possible to stop being eligible for assistance, so they stop counting you as unemployed.<p>The point is, the unemployment rates rides a lot. Which is why politicians love it so.
The article and headline asserts that economists view robots and automation as a threat to jobs. They don’t.<p>The conventional wisdom from most economists (defined as professors of Econ at major Universities) is the opposite, that automation leads to more jobs. This article is based on a strawman.
Why do we care about the number of jobs? We use this as a proxy for individual well being, but having a job is not intrinsically valuable. What matters is that everyone in a society has their needs met.<p>We should not be optimising the number of jobs but the total output of society, while maintaining an equitable distribution of the said outputs. It may be the case that automation permanently makes some classes of workers unemployed, but with the increased benefits from automation we should still be more than capable of providing for them.
As someone with a bit of experience in robotics, I've always viewed these economists as "useful idiots", unknowingly doing the bidding of firms who sell us the dream of advanced automation (e.g. Tesla, Uber). The economists were clueless themselves, but the companies selling the dream profited immensely thanks to them.
"Automation" propaganda was always a threat aimed at labor and labor unions. The only significant "automation" has been 1) moving factories to countries with lower wages and to localities with lax regulations on working conditions, and 2) the gig economy (the culmination of Manpower's et al. temp revolution from the 90s) along with independent contracting in name only to dodge benefits and the costs of downtime.<p>Actual automation is a comparison between workers, who are infinitely flexible, know their jobs better than their bosses, and are responsible for taking care of themselves, - and machines, which are morons that break down and require expensive workers to fix. Workers are better the vast majority of the time, and when we can get a machine to be cost-effective against them <i>it's a good thing.</i> It is just a rare, difficult thing.
As someone who develops automation software for the capital markets industry its too early to state what the impact will be on jobs.<p>For sure what I have personally seen is backed by this article in that automation has not resulted on job losses.<p>It has however resulted in task losses - meaning tasks that humans would do are being done done by software autonomously - this has freed up time and the theory is this time is then deployed to more valuable tasks.<p>In my mind the jury is out how this is really working out but if the gain is shift from work -> leisure time without impact on growth then I also see that as a win but as we know the distribution of these gains is far from even across society which is a growing problem.
I don't understand people who argue that robots create jobs because they need to be programmed, built and maintained.<p>Of course those jobs are created, but many more jobs were destroyed in the process.<p>Not to mention education cannot make engineers out of thin air, it requires a lot of resources to educate people.<p>And for those who don't get educated, it's a net loss since the gains of productivity are not shared.<p>That is what Antiwork is about: there's enough shelter and food for everyone, so work can't be mandatory. What's the point of working in fast food or taxi driver as a permanent job?
We are seeing a boom as cloud and data spend is going up as internal IT struggles to go without it, which in turn makes a virtuous cycle for even more. Likewise, once you start doing things like GPUs for graph neural nets and visualization, it's like highspeed internet: our users don't want to go back. For folks doing real things with data, it's been busy times.<p>My issue is more like "we and our F500 users are asking for more GPUs than azure can supply", vs job loss and stagnation.
This is not at all surprising. The scholarship positing the opposite was a deviation from what has been the standard line of economic thought since more than a century ago.<p>Automation is the main cause of a 20 times growth in inflation-adjusted wages since 1820. Per capita GDP growth is mostly just labor saving innovation, which is primarily automation, and secondarily greater coordination, like division of labor and specialization, and that is mostly through trade.
> Automation might help a firm become more profitable and thus expand, leading to a hiring spree. Technology might also allow firms to move into new areas, or to focus on products and services that are more labour-intensive.<p>This is a good argument as long as machine learning doesn't get too good. Unfortunately, improvements in ML are accelerating, so we'll see if that holds up.
I’ve always laughed at this idea that automation will take all our jobs. If it were true, farm automation would have left all of us unemployed at least a century ago. Most humans used to barely survive trying to eke food out of the earth. Yet somehow there’s not mass unemployment even though now a small fraction of the population is in agriculture.
How is automation different from industrialization? This has been going on for 200 years more or less.<p>Maybe, if lower prices stops increasing consumption (and investment). So far, better automation has just made us build more.<p>But why should that all change now? If it does, the transition will probably be gradual and take decades.
Not while we place very little value on peoples time ( their sole existence on earth ever ).<p>Capitalism is happy for people to waste there lives doing things they don't want to do.<p>A society that values peoples time as it should would encourage automation.<p><a href="https://www.newyorker.com/magazine/2019/05/20/if-god-is-dead-your-time-is-everything" rel="nofollow">https://www.newyorker.com/magazine/2019/05/20/if-god-is-dead...</a>
If we had perfect robots creating almost limitless "free" labour, what would we as a civilization do? Ban the technology? Descend into a struggle to own the technology. Or forge a better and different society for ourselves
The more we automate jobs that people hate doing, the more time people will have to do jobs they are really good at and enjoy doing.<p>Hopefully creating TikTok videos won't be the majority of the creative jobs humans will focus on. There is always that danger.
All this is wrong. The causalities are:<p>Low unemployment -> uppity expensive workers -> increasing rate of automation<p>Positive rate of automation -> more workers
Tim Worstall said it better than I ever could:<p><a href="https://www.timworstall.com/2011/03/peter-wilby-really-really-doesnt-get-economics-at-all/" rel="nofollow">https://www.timworstall.com/2011/03/peter-wilby-really-reall...</a><p>snippets:<p>Average wages in an economy are determined by the average productivity in that economy.<p>Individual wages are not set by what that individual does: but by the wages paid by the next possible alternative use of that individual’s labour.<p>. . .<p>Finally, as Marx pointed out, wages rise to meet average productivity, not wages falling as cheaper labour becomes more productive. This happens because capitalists compete for access to the profits that can be extracted from that labour. As the labour becomes more productive more profits can be extracted and the competition means that wages are bid up.<p>. . .<p>All of the things that we consume, now being made by these cheaper providers of labour, become cheaper, thus our real incomes rise. Their wages, now that they are becoming more productive, rise. Our wages on average, determined by our average level of productivity, move in step with our productivity, not the changes in the productivity of others. And our labour can go off and do those other things which will satisfy yet more human desires and wants: another way that we get richer of course, for satisfying two or three needs and desires instead of only one means that we are of course richer as long as we define wealth in any rational sense at all.