Technological Unemployment: Much More Than You Wanted to Know | Hacker News Reader

Technological Unemployment: Much More Than You Wanted to Know

slatestarcodex.com

370 pointsgbear605224 comments

Production just requires land and labor. Capital is output from previous production used as input again, and it is nice to have but not strictly necessary for employment of labor.

Labor can be applied whenever labor has access to land. So the question is why can't people just go to work? Grow their own food to feed themselves, if need be?

The Industrial Revolution put farmers out of work, but not because farming became illegal, and not because they all forgot how to farm. It is simply that farm labor couldn't pay the rent of land anymore.

The mechanisms by which the unimproved value of land rises and falls is the key mechanism that we must understand in order to understand wages and the rate of unemployment. Real wages are not a number. They are the lifestyle you can obtain. They are what is left over after you've paid your taxes and your right to exist in a particular location. (Increasingly, healthcare and education are prerequisites to getting by in the same way.)

Simply giving people money, irrespective of the financing method, as most advocates of the "technological unemployment" thesis argue for, ignores the nature of rent. When there's no rent-free land, rents rise and fall based on the ability to pay. People getting free money have more ability to pay, and landlords will definitely notice.

Thing about land is they aren't making it anymore. Higher demand doesn't just stimulate supply. The supply curve is vertical.

This would be a wonderful thesis if it didn't ignore the fact that in the world (and the US in particular) there's a lot of land. So much land that you could fit the entire world's population twice into Texas alone assuming major city densities (assuming 67k /mi^2 and 268000 mi^2 per cursory googling gives some 18B).

Roughly none of that land is freely available.

It's also pretty clear that some sort of resource footprint is more interesting than apparent living density. Just food requires more than a few hundred square feet per person and we consume a lot more than food.

> Roughly none of that land is freely available.

You'll have to qualify that statement a little. Here is 5 acres of land in Texas for $1500 [1], found it in about 60 seconds.

[1] https://www.zillow.com/homes/for_sale/TX/land_type/209109538...

I actually said roughly, I didn't go back and edit that in or anything.

If you want to do brass tacks or whatever, buy it and scrape out a living to prove that cheap scrub desert makes a point about land being available.

What if automation and robotics made homesteading much easier, and more likely to succeed on land that was previously considered useless?

Obviously water and other resources are going to be an issue, but as a labour multiplier, AI and robotics have huge potential.

Unfortunately, AI and robotics are only available to those with great capital and knowledge on how to apply those technologies; and I don't really see that changing any time soon.

West of the 100th Meridian, water gets scarce. Land out here has value pretty much in proportion to how much water it has available, because you can't grow anything on it without water. And you can't fix lack of water by availability of robots.

> What if automation and robotics made homesteading much easier, and more likely to succeed on land that was previously considered useless?

Then that land (if not already privately owned and actually available for homesteading) will be homesteaded, privatized, enter the land market, and not be rent-free land that is available. Just like the land that was attractive for homesteading in the past has been. Sure, its a one-time land-rush just like the initial homestead act, but after that the same forces of centralization of property that always work in a capitalist economy will work on it.

And the cost of plumbing and electrifying that land would be astronomical. Public works have a network effect.

As maxerickson said, that land is not free. And the question is not the amount of land, but the amount of valuable land. Land that is near jobs, near markets, near ports, or near decent hospitals or schools.

Sure there's loads of terrible land in West Texas, which could possibly provide for a lifestyle of subsistence farming, but the US standard of living is such that it would be wiser to live on welfare in a city.

But in theory, if someone could covertly farm some crops on one of the numerous vacant sites in Manhattan, and not have to pay any rent, then that same amount of subsistence labor could provide for a potentially decent lifestyle.

Same amount of effort would bring in a much greater return, just through the better location.

My point about farming was to ask why we can't all share in the fruits of rising productivity. Why can't we just keep doing what we had been doing, and still take advantage of fancy new gizmos? And the answer is that progress causes land values to rise, but the valuable land is generally owned by a subset of the populace who demand a greater and greater rent just because they can.

So really the question is: how can we all share the value of land, without it just funneling back to higher rents? My suggestion is to look at land value taxation.

Your suggestion would be wrong because your analysis is faulty. The land is useless without skilled people to operate it. What the skilled people have done is divide the labour which has created machines that farmers can use. The trouble is they can produce enough for themselves and their machine makers by the end of Tuesday. The challenge for the rest of us is to provide something of value to the farmer so they will work the rest of the week and create an exchange surplus.

Rentiers are just the current solution that forces farmers to work til Friday. The problem is that trickle down doesn't work.

Land value tax just makes everybody rentiers and racks off the farmers because they get nothing of exchange value in return for their week of labour.

Labor is the active force, and land is the passive force. Production cannot happen without both.

As for everyone being rentiers, I'm not sure I see the problem. Everyone being rentiers ensures a minimum standard of living, allowing us all to share in the progress of civilization. But really, land value taxation itself raises the "margin of production" even without any kind of redistribution.

The mention of trickle-down seems like a non sequitur.

>Thing about land is they aren't making it anymore.

SpaceX is going to change it.

If things worked out according to the rosiest predictions, that would raise what classical economists called the "margin of production" and would thus raise wages for as long as there is freely available land of sufficient quality. Although, you'd be surprised how quickly a commons can be enclosed.

Yet, given the harshness and distances in space, and the fact that land generally obtains value by proximity to other stores of human-produced value, I'm not as optimistic.

>Yet, given the harshness and distances in space, and the fact that land generally obtains value by proximity to other stores of human-produced value, I'm not as optimistic.

originally i did mean space, yet even the cheap globally available internet the Musk plans to build using SpaceX would noticeably increase availability of "connected" (an important aspect of "proximity") land.

> as long as there is freely available land of sufficient quality

For any reasonable near-term (centuries scale) prospect, there are no serious space or material constraints to outer space. Quality is a different subject, I want to get into that.

> Yet, given the harshness and distances in space, and the fact that land generally obtains value by proximity to other stores of human-produced value

Density of development is fundamentally different in microgravity. Our cities are all 2-dimensional. We can create a space station that hosts a "metro" area of stupendous population where everyone is in fairly quite commuting time of each other.

Of course, we have to return to the "quality" aspect. More work is required up-front to get this sort of thing working. On Earth, we already have many resources that we take for granted for which we would have to devote effort to producing in space. Consider, Oxygen, fresh water, soil, etc.

But couldn't the same thing have been said about virtually every component that went into creating the industrial revolution in the first place? The production of one thing requires support of another industrial product. Enormous production chains are created. Coal products are used in blast furnaces to make steel. Steel is used to make rail tracks to transport coal. Coal is burned to power steam engines to remove water from coal mines. Steal is used to produce those steam engines. Before long, steam engines are used to move coal over the tracks. All of this network goes to create manufactured goods, which make farming more efficient to support more people. In a superficial sense it would kind of sound easier to stay at prior population levels and live with the less efficient farming that didn't have this wild, confusing, and unsustainable web of dependencies.

Why haven't we made progress toward a space-based industrial network of production? You can hardly say we've tried. Through the Shuttle era, launch economics hardly budged. Space programs were turned into science & jobs programs. Manned exploration beyond Earth orbit stopped, and all robotic (indirect) interaction with the rest of material in our solar system ground to a crawl. A major rover like Curiosity takes years to drive as far as the manned lunar rover did in a day, at a tremendous time and resource cost.

Actual exploitation of space resources has so little actually invested in it. The main value we derive from space now is a perch from which to spy on and locate the citizens of Earth. Spending on space is not, at all, focused on something you can "development". Nor has it been. The one time we passed an international law on space resources (the Outer Space Treaty), we came up with incoherent and inaction-able drivel that undeveloped (and presently useless) resources were the property of everyone, and that only governments are really use them.

It hasn't been done because it's uneconomic with current technology.

Space mining is a nice idea with no practical road map. You can't just send tools into space, aim them at the nearest asteroid, and start drilling. You certainly can't expect to start building viable habitats yet.

The industrial revolution was incremental. A steam engine wasn't a completely new thing, it was just an improved source of power for existing technologies like spinning. Mechanisation could proceed gradually in a relatively comfortable environment with plenty of natural resources.

Not so in space. The problems - energy generation, mining, refining, distribution, assembly, sustainable life support and/or control (and probably defence) - have to be solved all at once, or you get nothing.

The steam engine was invented to remove water from coal mines, which was extracting coal to be used for heat. It was pretty darned new at that time, because the first versions operated by pulling a vacuum and experiments on strong partial vacuums had only just really begun not more than 100 years before then within English intellectual circles. It took quite some time and monumental efficient improvements after that for people to start employing coal-fired steam engines for tasks other than improving the production of coal itself.

You are exactly right that an entire network has to be solved together to open up space. This was also the case for the early industrial revolution. Because of this catch 22, right now we essentially have "space for space's sake".

That's also why space development will follow the pattern of a tipping point. As the pre-requisites aren't really there, investors don't have a good option to pour money into the venture. Once those building blocks start to get filled out, options open up, investment comes in, the building blocks themselves get cheaper, and things will progress faster than what people have imagined.

Humans have had, in total, a pitifully scant amount of interaction with extraterrestrial resources. Micro-gravity manufacturing isn't a thing yet, and we don't have the practical knowledge of how that will work. But the first step is lowering launch costs...

Funnily enough, my own (long term) project is aimed at addressing the problem you're pointing out. Given the possibility of remote work, there's no insurmountable reason why the only land worth living in should be concentrated in a few urban centers.

https://sites.google.com/s/0BzzRwqIB2K0dSWhkTXBSTTltTDg/p/0B...

403 forbidden.

My bad.

https://sites.google.com/view/virtuouscycleinvestment/projec...

> there's no insurmountable reason why the only land worth living in should be concentrated in a few urban centers.

IMHO, one very insurmountable reason is that you could never convince everyone that an urban center is worth living in. Quite a few people just don't enjoy that lifestyle.

I think you read that wrong. Paraphrasing what you said: Some people not liking to live in cities is an insurmountable reason that land in city centers is the only land worth living in.

First, your response doesn't make sense (and I think it's just because it's a misreading of the comment you replied to).

Second, I agree, there are people (myself included) that wouldn't enjoy living in an urban center. But that doesn't address what I think the comment was getting at, which is that land in urban centers is more desirable to a greater number of people than land outside of them.

> there's no insurmountable reason why the only land worth living in should be concentrated in a few urban centers.

There's actually a very fundamental reason why land perceived as being worth living in becomes an urban center, and remote work doesn't change most of the climate, water, resource distribution, etc., reasons why particular sites are much more attractive for human habitation.

> Given the possibility of remote work, [...] urban centers.

Even HN doesn't yet seem to be discussing the impact of coming VR/AR on people's geographic distribution.

Perhaps it's still too low profile? Eye tracking; facial expression tracking (researchy); foveated rendering reducing GPU requirements; inexpensive laptops with discrete GPUs; HMDs with monitor-like angular resolution; inexpensive finger tracking (hopefully)... perhaps it's easy to not notice these if you're not tracking the tech? Easy to think "remote work, so video conferencing, so limited impact - far better to do open plan offices"? Instead of "the practice and spatial constraints of software development will change dramatically over the next half decade".

Or for that matter self driving cars. You could commute for 5 hours a day, spend 2 hours in the office and go home again. All while being productive. That greatly increases the commutable range. Then you have the prospect of self driving Motor homes, moving with the work becomes that much easier.

How will we be able to do this without fossil fuels?

People take commuter buses with wifi, from Pennsylvania to NYC, with a 2 to 4 hour round trip. A few commute Boston to NYC on Amtrak Acela, with a 7-ish hour round trip (you can stand up and walk around, get coffee, etc).

I don't understand what problem VR/AR is solving in your hypothetical.

Not the OP, but I've been thinking about this a lot recently.

The problem is that telecommuting is currently very limited by technology. You can currently only transport your voice and a single video view, while collaborating over a single view from a monitor. You end up missing on very many of the modalities available in the real world.

But my vision of what work and study could be like with good VR is The Magic School Bus. Assume that every time you an your colleagues wanted to do work on any domain, you would be transported into a hi-fi all-senses-included representation of that domain, which provides a perfect sense of presence. That technology is still quite far away, but once we're there, it would be significantly better than just being in a room with them. And the world would be forever changed.

so: constricted communication bandwidth?

That's one way to describe it, a better way might be that's its not bandwidth that limits remote working communication, but just a lack of features.

To loop this back to the topic at hand, the idea is that you need people in the same space to work effectively. That idea forces collections of people (cities) and drives up the value of the land (a fixed commodity). Having technology that removes the need to have the workforce physically opens up vast amounts of relatively cheap land to live on.

I think the efficiency gain from cities is more than just some people working in a single company, but all sorts of logistical savings.

Perhaps my imagination isn't big enough, but I just don't see VR replacing dasein as it were.

> efficiency gain from cities

Sure. It's just that "tolerable commute" would no longer need to be one of those gains.

A fun conversation went by with someone from the Sierra Club. It was pointed out that the built environment which best matched their resource conservation/efficiency goals, was Manhattan. Their response was something like "yes, but we don't emphasize that, because it wouldn't appeal to our membership". ;)

> dasein

Picture brainstorming in front of a hallway whiteboard. Picture grabbing some nearby colleagues to collaborate. Now picture how relatively ineffective and unpleasant it would to do that collaboration over video conferencing and such. It's suboptimal, but that's a price you're sometimes willing to pay for non-locality. But now imagine that given a choice, you would much prefer to be brainstorming in front of a virtual whiteboard.[1] And collaborating there. Simply because it's a more powerful and helpful environment. So your physical whiteboard languishes, gets taken down for wall painting, and is never rehung. Then non-locality is no longer something you pay for. Locality constraints are just gone.

There's a fun decades-old metric on communication technology: Given some communication task, and a choice of technologies, how far would you walk to get access to them? Physical letter, email, phone, in-person conversation. Down the hall, different floor, different building, across the city.

Already, there are some tasks where even if you are sitting in the same office, you would use some tech rather than turning around. VR/AR seems likely to greatly increase that set of tasks. Reducing the importance of being able to physically walk to people.

[1] Here's a limited existing virtual whiteboard: http://store.steampowered.com/app/585040/Dry_Erase_Infinite_...

Or maybe even just an always-on video setup. Basically use the following innovations to connect an office in SF with one in Fresno:

1.) MIT Stanford Wormhole http://archive.boston.com/business/technology/articles/2011/...

The designers of the installation, who call it the wormhole, say it is meant to encourage random encounters between students and staff at two of the country’s premier technology-oriented universities. Someone at MIT’s Stata Center for Computer, Information, and Intelligence Sciences, hungry for lunch and conversation, will be able to chat under the dome with a stranger who’s grabbing a morning coffee in Stanford’s Huang Engineering Center. Casual meetings can be held at the wormhole. New ideas may be spurred, new science inspired.

2.) Google Fiber

3.) 200 inch LED video wall https://www.engadget.com/2018/01/07/samsung-4k-146-inch-micr...

> Or maybe even just an always-on video

Or perhaps video in addition to XR? HMDs don't entirely play well with eyes (at least until we get adaptive optics or light fields), so I want to give my eyes a break periodically. A video wall in the kitchen would permit a "let's take a break" conversation with a non-local colleague.

The power of the XR argument, is that if your ideal development environment is XR, and your ideal collaboration context is via XR, then locality constraints have simply gone away. Rather than video's challenge of being an impoverished medium, which may or may not be worth some non-locality benefit.

> MIT Stanford Wormhole

I spent a lot of time in Stata, over may years, and I'm not sure I ever saw the "wormhole" used. But perhaps there was some pattern of use, like late lunch, which I systematically missed. And one can walk by the cafeteria without easily noticing.

Can you provide a working link please? This is something I've been thinking about a lot, would like to learn more about what you are thinking.

oops, thanks for the heads up.

https://sites.google.com/view/virtuouscycleinvestment/projec...

"The Industrial Revolution put farmers out of work, but not because farming became illegal"

Not completely true, see the Inclosure Acts https://en.wikipedia.org/wiki/Inclosure_Acts

It's not unique to England either, it's just well documented there. Generally speaking the process of removing subsistence hunters, herders and farmers off of land they have historically lived off takes place with violence - not because they can't pay the rent anymore.

Additionally, the claim on the land of the people who are generally doing the displacing is usually questionable at best...

This isn't a matter of history either, taking land held in common by indigenous people with little if any recompense is a process that continues today in many parts of the world.

>>The Industrial Revolution put farmers out of work, but not because farming became illegal

>Not completely true, see the Inclosure Acts

I don't think that's disputing the parent's point. Seizing land isn't "banning farming" any more than repoing my car is "banning driving".

To respond to the general point: if yeoman farming were actually more productive per unit input, then even enclosure wouldn't be enough to stamp it out. Displaced farmers would still have been able to buy back the original land and continue: the new owners would see the offer price as being better than the capitalized returns they could make, and some bank would have seen the business plan and interest payments as being enough to merit a loan.

"if yeoman farming were actually more productive per unit input, then even enclosure wouldn't be enough to stamp it out."

I'm not making the claim that yeoman farming is more productive. I do claim however that such laws are sufficient to crush indigenous agriculture because enclosure is a violent and effective process for doing so.

"Seizing land isn't "banning farming""

Enclosure laws aren't a ban (no society has ever banned farming to my knowledge), but they effectively put some farmers out of work if they engage in a particular type of subsistence agriculture. That is farmers that utilize a community owned shared hunting ground/pasture/cropland on land without a title recognized by the state.

"Displaced farmers would still have been able to buy back the original land and continue"

Your typical indigenous person or tribe who is having their land enclosed or confiscated doesn't have the capital to buy it back, assuming they are even have the knowledge, language and legal framework on how to go about with that process. They are likely not even interested in agriculture for the purpose of making money, and if they were, do you really imagine people like the Orang Rimba are going to buy back their land? http://www.bbc.com/news/world-asia-41981430 It's going to be palm oil plantations no matter how effectively they were using the land - assuming anyone is even measuring.

> Simply giving people money, irrespective of the financing method, as most advocates of the "technological unemployment" thesis argue for, ignores the nature of rent.

Its not giving people money, its redistributing it. The total amount of money chasing land rents won't change by the full amount redistributed, so the whole amount won't go into increased rents. Who can rent how much will change, and that's pretty much (in the reductionist view you advance where all purchase of economic output amounts to land rent) the point.

> Thing about land is they aren't making it anymore.

This is false. Humans have been making new land for a long time, and aren't about to stop:

https://en.wikipedia.org/wiki/Land_reclamation

Matter is neither created nor destroyed, and locations are the same.

If you did make an island off the coast of Manhattan, then you'd only be able to do so through the ownership rights of the location.

Titles are granted by the government, usually at a price, but sometimes as a handout to special interests. The US Federal Government leases the rights to deep sea oil rigs.

If you obtained such a handout to build an island off Manhattan, that would be a transfer from the commons into private hands... but not a net increase in locations.

You could easily take away the exact opposite conclusion from that. The source of funding for the money you are giving away should be 100% of all land rents.

>"Increasingly, banks recognized the value of tellers enabled by information technology, not primarily as checkout clerks, but as salespersons, forging relationships with customers and introducing them to additional bank services like credit cards, loans, and investment products."

Was this written before the Wells Fargo scandal came to light?

Hard to be optimistic when the "silver lining" is that people who were once employed to do necessary work have since been reassigned to trawl for opportunities to make easy money at the expense of the gullible and vulnerable.

Sadly this appears to be yet another theme in technological advancement--the proliferation of scams. Just look at how much social media advertising comes from multi-level marketing schemes.

Turn the competition knob to 0, everyone slacks, turn it to 11, everyone cheats. In a slowly evaporating pool of jobs, it will head to 11 very quickly.

That behaviour describes my biggest pet peeve with banks. I don't want product offers from you. You take my money, make money with it, and pay me interest (a whole < 1%). That's our relationship, fuck off with your "credit card protection free trial period".

That "credit card protection free trial period" is how they're making money.

originally they made money by the margin between the profit they make on wise investments and the amount they pay in interest for the privilege of investing money for them. those days are long gone.

The Wells Fargo scandal does not represent the banking industry as a whole. Believe it or not, but the products sold by the banking industry in general have an investment return greater than 0%.

Investing human capital in providing those products results in a net increase in overall human productivity.

It feels like we're moving away from jobs that are purely productive, to jobs that are more creative. I don't mean arts etc, but 50 years ago a company like Facebook/Snap/Gaming/Any digital entertainment really, could not exist because people were occupied work that was required for society to function. Now humans can start focusing on spending their day building things for the entertainment of other humans, and also choosing careers based on what they feel is fun/satisfying for themselves.

Maybe. On the downside; maybe those jobs won't be available to everyone because there is a skills gap or there aren't enough of them.

If there are people who are left behind; what do we do about it?

As long as it isn't profitable to help the people left behind there won't be much done. Hopefully the people left behind will start making it unprofitable for the elites to ignore them. Probably won't happen though due to the vast amount of surveillance we're subjected to along with possible robotic police force in the future.

Yes. What is being suggested by the zeitgeist is that we are in a transition period; and that the scale of the problem is going to be so large that it cannot be ignored.

That is, if 75% of the population has no job in 2150 you have a huge ethical and societal problem.

This article is about unemployment rates, but wage drops in lower paying job sectors is already a huge ethical and societal problem. It can already not be ignored. It is, nonetheless, still being ignored.

In my mind large corporations like Facebook are a good example, because not all of their 25,000 are programmers or "high skilled workers". Their will always exist an ecosystem around large groups of people that are diverse enough to satisfy most types of people.

I'm not saying that this is the end-all solution to the larger problem described by the article, but I do feel some hope that it will work itself out. However, the advent of self-driving cars will be interesting to follow, considering how many people work in human transportation. But I'm optimistic!

>also choosing careers based on what they feel is fun/satisfying for themselves.

I don't think this is ever going to change in human society. If a job is fun to a large group of people, then it will pay poorly because many people will want to do it. And poor pay decreases the satisfaction one gets from such a job.

The only people who win here are those who find a job fun / satisfying that most other people cannot handle.

There's a reason professions like doctory, investment banking, and consulting pay so much, they involve lots of sacrifice and/or job unhappiness that you have to pay people a lot to get them to commit to the job.

Lots of people would love to play in the NBA and would find it fun and satisfying. It's just that very few of them are 6'8", absurdly athletic, and skilled at shooting a basketball. However, if something like nanotechnology made it possible for all those people to play basketball at that level, then the entertainment and spectator aspect would fall away and the desirability of that occupation would go away as well.

Exactly.

Lots of people don't want to play in the Eastern European basketball league. They specifically want to play in the NBA. It's not about basketball, it's about the perks.

Sidenote: Other than the height, the absurd athleticism and skill at shooting ball are grown over a long period of hard work. Nanotechnology cannot substitute for hard work and years of learning the strategies and techniques of playing the game...

> Nanotechnology cannot substitute for hard work and years of learning the strategies and techniques of playing the game...

Of course it can. You don't even have to go nano. Give me a team of people from Boston Dynamics for a year, and I guarantee you they'll be able to make me a wearable exoskeleton that is better at throwing the ball on-target than the best NBA player.

It's just, like GP said, with such tech allowed, "the entertainment and spectator aspect would fall away". Sport disciplines are designed to ensure scarcity of good athletes.

> There's a reason professions like doctory, investment banking, and consulting pay so much, they involve lots of sacrifice and/or job unhappiness that you have to pay people a lot to get them to commit to the job.

That's partly true, but there are other dynamics at work as well.

Expert sewage/waste management workers get paid a lot less than those trades, though their trade is necessary and involves lots of sacrifice/job unhappiness.

On the other hand, private military contractors and undersea welders make a premium, despite being nominally "service" or less popularly lauded jobs.

There are other factors at work in addition to this.

Private military contracting and undersea welding are both dangerous and require fairly specialized skills. Waste management, though critical for the proper functioning of society, is a low skill job and not particularly dangerous, which keeps the wages low.

And medicine, finance, and consulting, require a fair amount of intelligence and the barriers to entry are high, keeping wages high.

Those kinds of jobs are only sustainable when the audience is 100-1000x + the size of the pool of creators. It's not a broad panacea.

Just look at patreon. It is fairly unreasonable to expect even your fans to spend more than $1 a month on you, so if you want a reasonable income (and I'd consider reasonable to be peak happiness if you want success in life) than you want around 100k USD in yearly earnings. Take out the 40% self employment tax, the 3% Patreon takes, and you need roughly 12,000 people giving you a dollar a month to "succeed" in terms of income happiness as a creator on Patreon.

Now, some of your audience are going to be whales that have vastly more money and are willing to treat $100 like $1. But you also have a huge portion of your audience that pays nothing, by nature of the default state of Patreon being a free main product with only supporter bonus content rather than paywalling your main focus (because how are people going to discover your stuff if its paywalled?). I imagine the best conversion ratio is like 20% and that is wildly optimistic. Twitch subscriptions are comparable, where most streamers float around 20% of their average viewership as subscribers.

So you need 60,000 fans, at least, to make happiness income creatively. Now, each fan would probably support more than just one creator per month, and there are additional revenue sources including ads, merch, cons, kickstarters, etc. But no matter how you subdivide this economy the ratio of creators to consumers has to be drastically biased towards consumers over creators.

Facebook/Snap/Gaming/Digital-Entertainment don't employ all that many people.

I'd be surprised if the ratio of entertainers / audience is much different than it was for, say, stars of silent movies 100 years ago, even though Charlie Chaplin wasn't doing something that could be termed "needed for society to function".

Ah, but he did. See [The Great Dictator](https://en.wikipedia.org/wiki/The_Great_Dictator) for example. For us to function we need people like Chaplin to show us (in a way or another) what's going on in the world.

The Great Dictator is one of the few films with obvious value beyond entertainment, because it's a vaccine against Triumph of the Will. Triumph of the Will is a skillfully made propaganda film, and if you have no awareness of how it works then it would be easy to leave with positive feelings about the Nazis, even if only subconsciously. But once you've seen The Great Dictator it becomes a joke. E.g. the low camera angles don't make Hitler look powerful, they just remind you of the extending chair scene. His exaggerated gestures no longer project confidence, and instead make you think of slapstick routines.

John Cleese's silly walk had the same effect on the goose-stepping scenes. It's art that makes people more rational, by transforming emotional manipulation tricks into comedy. It's beneficial even if you don't agree with all of the politics of it (e.g. Chaplin's call for open borders).

Oh, that's an interesting question. I've seen a graph of (US) advertising spend vs GDP which seems to hold constant over the last 70 years.

But to lump all FB/snap/twitter into entertainment+advertising and get % of the population working for those companies would be interesting.

But then you start doing violence to the categories. Twitter/Facebook certainly have entertainment functions, but they also have tentacles reaching deep into publishing and news. Now, suddenly a like for like comparison needs to include newspapermen, typesetters, printers and papermill workers.

Well get to work then : )

But seriously, I hope data-minded grad students/sociologists are investigating this more.

> Now humans can start focusing on spending their day building things for the entertainment of other humans, and also choosing careers based on what they feel is fun/satisfying for themselves.

How long before we get to a situation where this is a realistic proposition for many people? Most people on HN live in a comfortable bubble. The reality right now is that the vast majority of jobs are not creative, fun, or satisfying for the people that perform them. Shop work, changing car tires, cleaning roads/offices/windows/people, mining, warehouse work, digging holes, erecting scaffolding, checking invoices, data entry, driving trucks, painting houses. Dull and exhausting: that's the reality of work for most people right now.

Or worse. At least those dull and exhausting jobs could formerly afford you a reasonably content middle class lifestyle with a modest home, but that is increasingly out of reach of even some of the most successful in society at least in some countries, if you'd like to live where there are actually jobs.

As we continue down this road to a winner takes all society, I'm not so sure this is what I'd call progress, and I'm easily in the top 20% of income.

Both a middle class lifestyle and a modest home are moving targets. People eat out a lot more than in the 70’s, they have larger homes and smaller families, they have safer and more reliable cars, they live longer.

As far as being able to live where the jobs are this is a problem caused by NIMBYs. This is not a problem they have in Japan.

The root cause is fewer desirable jobs. Fewer businesses with a larger share of the market, in fewer and fewer places.

Desirable jobs is also a moving target. I’d bet a lot of money that virtually every job is safer than in the 70’s, more are indoors, fewer are outside or demand physical labour. If you mean that the supply of desirable jobs has increased slower than demand, sure, I can buy that.

Fewer jobs offer job security and decent pay. Because of less job security, you need to be in a market with many employers (aka cities), and this results in more people living in fewer places, which means higher rent.

That translates to needing higher pay, or at least a career where you can outpace rent and insurance costs. There are some manual labor jobs that pay well outside of cities, but you have to watch out for the condition your body will be in since people live longer and there will be less of a social safety net for when you are disabled and/or are too old to work.

Bottom line is more and more people are less and less valuable.

Japan doesn't have an artificial demand problem like some other countries. They're also not mentally disabled by skin color.

> How long before we get to a situation where this is a realistic proposition for many people?

the FIRE (Financial Independence / Retire Early) movement is showing how most people can do this in a matter of 20 years instead of 40, with those who are highly-compensated doing it in 5-10 years.

20 years of drudgery is much better than 40-60, so I'd say we're at least making progress. The hard part of realizing that goal is the much more difficult skill than the math: establishing what is "enough" for a good life.

I'd like to see numbers here. There are huge industries that have been built around entertainment and recreation forever: sports, hunting, movies, television, books, music, etc. I expect, as a ratio, we're collectively spending far less on e.g. baseballs and far more on video games. It'd be interesting to see the net real change in society's total non-essential expenditures. I would be quite surprised if it was all that large especially if we do not consider ultra-high end recreation to account for the fact that there are far more upper class and rich individuals now a days, so their recreational spending could make it look like all of society is spending that much more - when that's certainly not your hypothesis.

And Facebook complicates the matter even more. They do not make money by "building things for the entertainment of others". They make money by building one of the most effective datamining and human profiling systems in existence, enabling companies to pay money access the product of this by precisely targeting individuals to compel them to either buy things or be influenced by their brand. It doesn't really answer the question of where money is being spent. Those companies are giving Facebook money because you give those companies your money. What is the distribution and gross effect of that spending?

There can be a substantial difference between what the company and the employee "do."

I walk by a twitter building every day. Probably about 500 people, maybe more. What are they doing in there? They sell ads, manage ad accounts, moderation, data protection audits... stuff that David Graeber wold call BS jobs. Now... I'm not in the same camp as him entirely. I think it's rash to just write off 70% of modern, white collar work as "paperwork" that could just go away.... but I'm not entirely not in his camp too.

Either way, that building is full of "marketing and social media" graduates doing stuff that doesn't need to be done for twitter to be twitter. Twitter was twitter before they had all those people.

You could call what twitter does creative work. That resonates with me. They invented a communication method with all sorts of cultural impact and value. That's creative. But, extending that to the median employee... I think we might divert here.

I think you have your finger on an interesting point, even if I don't know how to articulate it. A generation ago, a lot of the ecoomy could be measured in stuff with value that one could reason about. Cars, tothbrushes, haircuts, podiatrists.... You can count and weigh and value that stuff. Media was always more abscure. The tv viewer mightn't be the customer. The purpose of a paper might be do get a classifieds insert into your hands.

Once you throw in social/user-generate services make the picture more messy. Once you go to Google/FB territory, it's all very obscure. Half of what they do doesn't have any direct economic benefit to them, it's more a way of keeping them in a massively powerful position.

I'm not sure creative is the right word, but here's somethign going on. Ever increasing (already majority) sections of the economy are producing very ethereal things, with obscure business models...

  Twitter was twitter before they had all those people.

Sure, but it can't scale without them. How exactly is Twitter supposed to pay the bills? People use this platform because, for the most part, the basic premise is entirely free. How else can they fund the infrastructure/R&D costs?

scaling...it did. That is it had a ton of users using essentially the same product with a much smaller team, at one point.

Paying the bills.., IDK. Presumably they've determined that these people are useful in this regard. Otherwise, why have them? But from a first principles reasoning, it seems possible to sell add without thousands of ad sellers.

Selling advertising online could be automated for the most part, unless Twitter really does offer personalised offering to ads buyers—which I'm not sure is the case.

At a certain spend level, advertisers want to have a single human that they can call when they need something/something goes wrong ("a single throat to strangle").

Providing this dedicated connection is also an opportunity to upsell them on more ad products.

I did this work at Google for a couple years at the start of my career. They are an exceptionally analytics driven company, and would not employ these kinds of people if there wasn't good ROI

.. because they believe that they'll get more money from the efforts of these people than they cost? If they can sell ads that people wouldn't buy without a person on the other side, they can be worth the cost.

Twitter has people to sell adds to companies that use people to decide to buy ads. If automation makes it easier to use fewer people to buy ads, then twitter and other ad companies would likely sell ads with fewer people in the mix.

As an aside it's odd to think of twitter and google etc as a kind of modern successor to the MadMen ad house. I guess if you were being pedantic, they're more like the modern version of what the network TV commercial industry was in the 60's.

The economic machine used to fulfill our basic needs, to keep us varm and fed. Now, that problem is solved, we have infinite sugar and cheap synthetic clothes. So it's on to massaging our psychological needs. And the market forces who gave us cheap soda gives us the same "solution" in FB, Twitter and friends.

A giant perverted maslov hierarchy for our society if you will.

> stuff that David Graeber wold call BS jobs. Now... I'm not in the same camp as him entirely. I think it's rash to just write off 70% of modern, white collar work as "paperwork" that could just go away.... but I'm not entirely not in his camp too. (...) I'm not sure creative is the right word, but here's somethign going on. Ever increasing (already majority) sections of the economy are producing very ethereal things, with obscure business models...

What I don't think gets mentioned enough in these discussions is Jevon's Paradox[0], which can be summarised as:

> In economics, the Jevons paradox (/ˈdʒɛvənz/; sometimes the Jevons effect) occurs when technological progress increases the efficiency with which a resource is used (reducing the amount necessary for any one use), but the rate of consumption of that resource rises because of increasing demand.

Or more cynically: when increased efficiency creates greater demand than that increase in efficiency. I probably don't need to convince anyone here that this happens in real life all the time - we're all familiar with Wirth's Law[1], right?

So I think the same applies to automation and work. If anyone hasn't seen "Paperwork Explosion" by Jim Henson (you know, the guy who made The Muppets and the original voice of Kermit, and Ernie), this is a really good time to do so:

https://www.youtube.com/watch?v=_IZw2CoYztk

Look at what IBM promised: freedom from administrative woes! We can go back to do real productive work! Essentially exactly what we're hearing now. Now compare it to what actually happened. At the last two universities I've worked I have seen a lot of “the administration department is expanding to meet the needs of the expanding administration department” in action. At the expense of the quality of education and science, I might add.

Buurtzorg Nederland[2] got a lot of attention in recent years for breaking that trend with its simple approach to letting nurses self-organise as independent, self-managing teams (so what Agile tried to be I suppose, except for nurses). What is often overlooked is that this is basically just going back to the model that the Netherlands had up until the 1980s, but a bit enhanced with modern information/communication technology. Jos de Blok, who founded Buurtzorg, briefly that in this short TED talk[3], but my parents were both general practitioners until they retired a few years ago confirmed that this happened all throughout healthcare.

With programming, we know that complexity is something that must be actively fought or it sneaks in. We even have tons of special descriptions for how that happens: spaghetti code, big ball of mud, etc. So ask yourself: why would it be different in any other human-created system?

For me, this is part of the explanation for all the obscurity we see everywhere (that, and a vested interest of a number of parties in keeping things obfuscated). And my prediction is that the natural tendency of automation is to backfire, unless we actively fight this.

[0] https://en.wikipedia.org/wiki/Jevons_paradox

[1] https://en.wikipedia.org/wiki/Wirth%27s_law

[2] https://en.wikipedia.org/wiki/Buurtzorg_Nederland

[3] https://www.youtube.com/watch?v=SSoWtXvqsgg

No, we are moving from jobs that are actually productive, to bullshit jobs.

Advertising doesn't produce anything useful in itself, it simply shifts and amplifies demand, and yet, Google, Facebook, and a vast slice of the rest of the economy is dedicated to producing ads.

Even companies which are nominally dedicated to producing things of importance (eg pharmaceuticals) are mostly focussed on marketing.

Then there's the entire middle management strata of the the economy. Mostly a useless but noisy zero-sum jockeying for a rungs on the socio-economic ladder.

The economy is already large enough to produce plenty for everybody, guarantee food, housing and healthcare for all, and with shorter working hours at that.

Unfortunately we're trapped in this bizarre economic system that works us to the bone in pointless jobs we hate, and can't even guarantee us the basics.

This bullshit job argument has never made much sense to me, but I see the argument a lot. If companies care about maximizing productivity and profits, and given America's hyper-competitive economy and super-fickle stock market where the difference between meeting the earnings estimate versus missing it by a penny can make the difference between a stock being up 10% or down 10%, why would a company waste money on jobs that are bullshit. Obviously, they (the company) derive value from these jobs or else such jobs would not exist (or at least not for long).

Companies are far from all-knowing and perfectly rational agents. They are made of people.

Also, growing from a lean company into a too-big-to-fail behemoth (which happens all the time) is very valuable to CEOs & co. It what makes you graduate from billionaire to billionaire with political power.

Spoken like somebody who has never actually worked in a large company.

Picture a stereotypical manager who does nothing but attend pointless meetings all day. (S)he attends reliably, and is good at appearing busy and authoritative and otherwise managerial.

This manager will not be fired, as (s)he's outwardly doing work duties and not doing anything specifically incompetent or negligent.

Other management-layer folk like this person as they confirm to middle class white collar norms socially.

Nevertheless, this person is useless. Every company is full of people like this.

The parent argues that it's "a useless but noisy zero-sum jockeying for a rungs on the socio-economic ladder"

And you argue "If companies care about maximizing productivity and profits, and given America's hyper-competitive economy and super-fickle stock market where the difference between meeting the earnings estimate versus missing it by a penny can make the difference between a stock being up 10% or down 10%, why would a company waste money on jobs that are bullshit"

This is the same argument, just with a different spin. In the instance of Facebook, it was able to provide it's core utility with some software cobbled together by one guy in his dorm room over a couple weeks. So what are the other 25,000 employees for? It's not to make Facebook more useful, it's to prevent some other college kid from building a social media network that displaces it. It's all just a silly game that isn't very fun to play. One could argue that Facebook's core utility is now in delivering ads, but nobody likes looking at ads, and companies would rather not spend on ads if they didn't need to to compete, so again, it's all just a silly game. Very little of it is doing anything that is either useful or wanted.

The jobs are useful to the company, they're not useful to society.

Yet large companies often have hugely bureaucratic and wasteful practices. When I was contracting I would go from one business with an IT department of 100 to another employing 1,000. They were broadly similar organisations, yet there was plenty of variance in their efficiency.

Adding complexity often makes it harder for others to compete. So companies do that(via regulation, or technology, or contracts, or ad campaigns,etc...). And than complexity breeds complexity.

To be fair, the stock market is bullshit. Its legalized professional gambling where some participants have way more information than others. The fact the vast majority of business exists to try to "win" the stock market game than provide real productive returns to the economy reaffirms the bullshit economics involved.

IE, vast amounts of effort from major globalcorps is dedicated towards tax avoidance. You spend tremendous amounts of resources to avoid taxes, which is only beneficial to the shareholders in the short term, because the long term economic damage of tax fraud and avoidance cripples economies and stunts growth. But the stockholders need not care since they already made their dividends off all the smuggled money as the country burns around them.

I think you're a bit too cynical. The financial system isn't optimal, and the stock market is less important now than it used to be (because it's easier to raise private capital), but it's not gambling. The rate of return is positive over pretty much any long-term horizon, and it allows companies to raise money when necessary. I agree that many companies engage in unproductive, but short-term profit maximizing behavior though.

the whole point of capitalism is to align incentives s.t. when businesses act selfishly to "win" in the markets, they are producing real returns to the economy

Companies are awful at culling bullshit jobs when the economy is good. That's why they never hire back as many as they laid off during recessions.

It is a little counter intuitive, but it's real.

Check this article out: http://evonomics.com/why-capitalism-creates-pointless-jobs-d...

I wonder if farmers looked at the rise of manufacturing jobs the same way - a move away from jobs that are producing the food that's required for society to function, to jobs that are producing entertaining novelties. Perhaps Facebook/Snap/... will one day be seen as just as vital to the proper functioning of society as manufacturing is today.

Problem is that most of the people in the new industries make no money.

Whether you look at the app economy or you look at the influencer economy few people make most of the money.

>It feels like we're moving away from jobs that are purely productive, to jobs that are more creative.

Yes. Those + unwanted human surplus (most of us in the future) or some UBI scheme.

While I believe industrialization has certainly improved in the last 50 years, it feels that there has been a trend towards displacement of the vital work/product generation towards under-powered regions of the world. Potentially only very fortunate humans may now pursue careers in a more ethereal environment that is both profitable and emotionally fulfilling.

> Now humans can start focusing on spending their day building things for the entertainment of other humans, and also choosing careers based on what they feel is fun/satisfying for themselves.

In this sphere, there are winners and losers, and the amount of losers is vastly understated and the amount of winners is vastly overblown.

Take youtube, for example. Most are not clearing more than a few hundreds per month but there isn't a week where BusinessInsider isn't writing about some youtube celebrity.

So what? If you have a place to live and food, then satisfaction is next on the list. Is the youtube life not satisfying?

For the vast majority of content creators, one cannot live off youtube wages alone.

I'm somewhat surprised by the article's conclusion, but will accept it, given that Scott is pretty thorough with his sourcing and analysis. Even though he says we're getting the worst outcome (something is happening, but it's not apocalyptic, so people won't notice in time to coordinate at scale), it shifted my internal doomsday countdown ~20 years to the future.

One thing I wonder though, and it's a general concern for such analyses - just how reliable are the official statistics? Given the unreliability at every level (which I sometimes personally observed) - from people reporting bogus data, data entry people faking it to make their job easier, to bogus statistical analysis and biased reporting at the top - just how much faith we can put in all of this? I get that it's the only thing we have, but I wonder if someone tried to analyze those issues.

I think the fact that the data shows gradual trends over long terms with reasonable changes at big inflection points such as recessions gives some confidence that it is capturing some real world signals.

If I may ask, what was your internal doomsday countdown before reading the article?

About 10-15 years before progress makes some members of my family unemployed.

Depending on how representative your family is of the overall population, that could mean that the doomsday scenario for most people is already here.

A really surprising thing here is that their analysis is also based on an somewhat unrealistically negative datum. They're using the labor force participation rate. The labor force participation rate counts people who are happy to not work, which is not a problem - and probably to be expected as the total number of upper and rich class individuals increases over time as a share of society. A much more useful measurement is the U6 unemployment rate. The U6 counts people who are involuntary part time (want full time work but can't find it) as well as those who want to work but who are no longer actively searching as part of the number.

And the really surprising thing is that even the U6 is reaching record lows [1]. If the current trends with unemployment continue we'll have the lowest 'real' unemployment rate since the U6 started being recorded, 24 years ago. The one 'zinger' I'd add here is that even the U6 does not account for underemployment. Many of the jobs created in recent years have been McJobs and if somebody has e.g. a STEM degree and is serving coffee, then they should probably count as unemployed. Perhaps it's time for a U7. On the other hand, I do not think this would meaningfully change the numbers.

I still do think automation is going to be disruptive since I simply don't see how transportation automation, which is an enormous industry that offers relatively good wages, can end up otherwise. But I've also become less cynical since we've undoubtedly been sending productivity up, and even with the outsourcing of desirable jobs or bringing in workers to fill those roles domestically for lower wages, our economy seems to be doing phenomenally. And that's something I certainly would not have predicted.

[1] - https://fred.stlouisfed.org/series/u6rate

It’s really hard to measure underemployment. I mean, I have what most would consider a pretty decent tech job, however many of my peers are Senior Director of this and VP of that, and with my background and experience I know I could be a CTO rather than what I am. Am I underemployed? I feel underemployed, but who’s to say I am or am not?

A commonly used measure of underemployment is 'years of education not needed for this job'. In this measure, if you'd be qualified to be a CTO but are not, this doesn't count as underemployment, neither does the difference between a junior engineer and a senior one, nor between a mcdonalds franchise manager and mcdonalds CEO.

However, it does count going to a different/lower industry. A STEM grad in McDonalds; a teacher in a cleaning job, etc; where you can clearly point towards some 4 years spent on something that the current job doesn't need.

The idea that rising disability is a cause of declining labor force participation is interesting, and fits in with anecdotal data from what family members in medicine have told me. They basically phrase it as: If you're uneducated and poor, why would you work a low paying job with little hope of advancement when you could go on disability and collect a similar amount of money without working?

To be fair, many of these people are actually disabled. It's just that previously they could work. Like the article says, their managers cared about them and and had the capacity to let them take lower-impact positions that they were capable of doing.

We don't appreciate how difficult it can be to employ these marginally capable workers. Keeping these people on staff at a small company raises health and operational insurance costs pretty dramatically, so it can be hard to justify keeping them on staff, even if you wanted to.

Disability is also pretty difficult to get, despite how the news makes it out to be easy. My father had multiple heart attacks and cancer, and he's still fighting to get on disability. And when he does, it will take years to come into full-effect.

I agree partly, but from what my family members have told me it seems that some people are gaming the system for their own benefit. They've also told me that people who truly would benefit from being on disability (like your father) often have a hard time getting on it.

(Not in the US.) My partner worked a factory job for many years. They used to have a "specials" group (unofficial name for them), who would be taken on and they would be trained to do just one job, but they would learn it and perform it to a very high standard, day in and day out for many years. They would become very fast at it, and were paid quite well with regard to the average wage in the area.

Unfortunately, they were also the first laid off when automation was installed, and afterwards they were competing for few jobs at the same (regional) low wages with "normal" people.

It's an easy thing to fear. It's another thing to actually demonstrate this is happening in a measurable, widespread fashion.

FWIW most of the people I've met on disability probably deserved to be on there, including people I've met with severe mental illnesses. I'd posit there's more of an empathy gap here than anything.

You need more, less-biased samples!

Well, the proportion of people on disability is certainly increasing as the linked article demonstrates. My anecdotal data all indicates that some of the increase is due to bogus disability claims. The SSC article also makes this claim. It's not an issue I have any particular interest in, so I haven't looked too hard for alternative explanations but given the lack of readily available alternative explanations (anecdotal or not) for the rise in people on disability I think we can say the probability is high that a non-negligible proportion of those on disability do not meet at least the spirit of the criteria for disability benefits.

Most of the patient populations my relatives are referring to are poor, inner-city minorities, who the average HN reader is unlikely to encounter in day-to-day life, so it's entirely possible for the average HN reader to only meet legitimately disabled people while still having a sizable fraction of people on disability collecting disability for bogus reasons.

By the way, I think it's socially undesirable for people who are not really disabled to go on disability, but for many of them it is a rational response to their circumstances.

Warning: the following is conjecture which I have not researched. Because who has the time?

My main question while reading this article was why is the author so stuck on "manufacturing" jobs being automated away? He does address this very briefly in section III but I cannot help but feel his foundational research is highly skewed.

Sure, robotics are growing in popularity and that technology, specifically, will likely impact manufacturing. Comparatively speaking, however, industrial robots are very expensive. Only the larger manufacturers are going to be upgrading plants with multi-million dollar (or very, very recently hundreds of thousands) equipment to replace a few hundred workers each (if that?).

Fun fact: ~90% of the 5.7 million companies in the US had fewer than 20 employees in 2012. 50% of working Americans were employed by companies of fewer than 500 (source of questionable quality: https://www.businessinsider.com.au/us-employment-by-firm-siz...).

Compare all that to search results for "images of 1950s offices". You get photos of rows and rows of desks of identical configurations - people with paper pads, typewriters or tabulation machines. Look at what all these people are doing. They are accounting departments calculating statements and balance sheets by hand. They are typists writing out form letters one at a time. These are the job categories that have been wiped out in their entirety by spreadsheets and word processors that you can get literally for free. The CEO can do the work of that 1950s accounting department by herself in an hour.

Where did those jobs go? I don't know how many there were to start but it looks significant.

Yes, modern offices have the same configurations but it is fundamentally different. The companies I work in can have an entire business unit (or company!) working out of an office that size. One person at one desk is an entire marketing department, the next person is HR (including payroll/accounts/bookkeeping). The only rows of homogeneous roles you see these days are the grunts building the technology.

In conclusion, I get the feeling a lot of people wondering if jobs are disappearing seem to be suffering from a bit of "out of sight, out of mind". Most of the automated jobs are already gone.

Just my thoughts.

> Only the larger manufacturers are going to be upgrading plants [...]

Well, one consequence of this might be that only larger manufacturers survive. The new robots are increasingly general purpose and an army of those could quickly switch tasks based on current needs.

This is why the post-recession employment recovery in the US is complete fabrication. The jobs lost in 2008 were careers, the jobs that replaced them were gig work like Uber. The employment figures today discount the tens of millions of NEETs and the tens of millions of people trying to scrape by on gig jobs that provide no healthcare, retirement, stability, and put them below the poverty line.

Per-capita incomes are messed up because the top 1000 incomes inflate as much as the bottom hundred million deflate.

> will likely impact manufacturing

Robotics have been impacting manufacturing for decades. The death of office jobs you mention was another casualty of automation we already felt. Nothing about the impeding automation apocalypse is coming up ahead because its already happening. Its just the low hanging fruit went first, and the tree will keep being picked clean from bottom to top. The economy warps in response to it, in a fleeting attempt to preserve outdated ideology, by fabricating work and creating bullshit jobs out of bureaucracy and artificial complexity to keep society from falling apart. The problem is every single private enterprise resists giving people bullshit jobs to varying degrees, which creates the discontent between workers who lose income over time and employers who keep reaping larger and larger profits.

The increase in homeless encampments as detailed in this report are also a stark indication of how false the recovery is. Unfortunately most of the homeless get shipped to California, often with one way bus tickets from municipalities that don't want to deal with the problem. https://www.nlchp.org/Tent_City_USA_2017 (Looks like they need to renew their TLS certificate.)

Any company able to obtain financing will instantly replace 10 workers with a $1 million machine.

If the machine is vaguely durable that even works if wages are $10 (~$30K annual cost per worker->$300,000 payroll per year, investment pays off in reasonable time).

That assumes the work stays the same- people can be repurposed to do something else or fired. Investing heavily in tech is a big risk and is something that slows this kind of unemployment significantly

Sure, it's a cartoon example. But I think the broader point, that businesses can afford to invest quite large amounts to reduce labor costs, is not very controversial.

It's really not a cartoon example. I've seen your point but in reverse: small shops that get big contracts go out and buy $500,000 machines instead of hiring 10 people to do the work.

He alludes to that in the "something happening since the 1960s" - this is that thing, the rise of the computer.

What's to stop someone from making a cheaper industrial robot? And someone else to make an even cheaper one? And for someone else to make an open source basic robot and sell a higher level management suite as an add-on?

Or better yet, someone can buy a bunch of these, put them in a building and lease them out by the minute via AWS-like service.

I do agree with your point re 1950's offices.

Only the larger manufacturers are going to be upgrading plants with multi-million dollar (or very, very recently hundreds of thousands) equipment to replace a few hundred workers each (if that?).

Not really.

Bear in mind that when you're talking machinery costing hundreds of thousands of dollars, you're looking at stuff that farmers are buying these days.

It's not only the big companies that can do this. In an economy where a small house in SF costs $1M, a $1M machine is essentially a mortgage payment that you can depreciate.

What do I know, I am not very good at this either and I respect the author's willingness to get other people's perspectives and attempt to reach a conclusion.

But I would like to critically consider the "indicative trends" mentality of the author. Entities with ML algorithms that could replace jobs enmasse do not yet exist, so past or current economic trends shouldn't indicate anything. ML algorithms did not gradually takeover Go, for instance - the transition to AI should neither be so gradual. Maybe if someone did micro-studies on Uber's effect on Taxi drivers, it would server as a proxy but even Uber isn't a self driving car company (yet).

That said, I really enjoyed reading about what is causing unemployment or the loss of jobs. It's an interesting observation that disabilities is rising as the reason for unemployment in the US. But again, attribution of these trends is always difficult.

Anyways, thanks to the author for doing the explorative work.

He touches on the idea of humans having the past laid out before them, with the future to their backs. But then, as you say, fails to get to the realization that paying attention to the present isn't going to give us any predictions for how strong-AI, or even non-routine AI is going to change the job prospects for humans. Robots are expensive, but servers aren't.

Regarding the horse example, there US horse population is 9 million. The demand for horses never went away despite automation and has remained steady. If one tried to extrapolate the early 20th century horse population trend to the future, there should presently be no horses alive.

Although the labor force participation rate is falling, the total labor force size has not (although it has flattened in recent years).

Rather than the total elimination of jobs, a more probable concern is the bifurcation of the labor force by IQ, with those with an IQ below a certain threshold barely getting by, stuck with low-paying jobs that barely pay ends meet. Due to economic and technological factors, this IQ threshold will likely gradually keep rising. http://greyenlightenment.com/two-nations-within-a-nation/

This is why the decline of manufacturing jobs has not hurt the overall labor market size, because such semi-skilled, medium-IQ jobs are being replaced by service sector jobs. However, the pay for these jobs is not as good.

Often what happens is the new technology will coexist with the older one, as the bank teller example shows. Other times the implementation is so slow that it is not a concern. I remember in 2007 McDonald's, Jack in the Box, and other fast-food chains began to install self-order machines, but the adoption was really slow, the machines were often broken, and the machines didn't replace the role of the cooks and cashiers. People needed help using the machines, necessitating staff. Same for those food store self-checkout machines. Rather than the machines eliminating the cashiers, they coexist. Some people like machines, but others don't. Machines break and someone has to supervise the machines in case someone needs help or tries to fool the machine by imputing the incorrect weight or item.

Overall, I am not too worried but I can see some cause for concern. Regarding post-scarcity and UBI, entitlement spending has been on an unstoppable rise for the past 50 years. We're already headed in that direction. SNAP, medicaid, disability, and section 8 housing is pretty close to a UBI and post-scarcity. The future is one where the government assumes a greater paternal role in helping those who are unable to keep up with an increasingly competitive economy.

Right: like with horses[1], the problem isn't that humans will die out entirely, but that only a tiny subset of them will "earn a keep" beyond subsistence -- i.e. be able to sell labor whose market value is more than the amount needed to keep them alive.

The only horses that stayed around were the ones that could perform a novelty (racing/entertainment) or low-demand-niche role (short trips on rugged terrain).

[1] A really, really underappreciated and overdismissed analogy.

One of the problems is that most jobs (95%) created from 2005 up until now are temporary jobs[1]. That is in itself not necessarily a problem as it could mean that people just become freelancers. But for the US it opens up for two major issues.

1) Freelancers have to pay their healthcare themselves. 2) Most jobs created are forced freelance contracts i.e. even if you work at an accounting company they will try and put you on a freelance contract.

My personal belief is that while globalization is still growing we will not see the underlying changes until it' too late. That's why I think all governments should assume that most of their population will be without a job sooner or later and plan accordingly.

[1] https://qz.com/851066/almost-all-the-10-million-jobs-created...

I wonder if that creates more volatility in the economy, business is much more likely to reduce or end a freelancers hours than a full time employee in a downturn.

[warning: unfinished thoughts]

On the far end of the "technological unemployment" worry spectrum are those who really believe that AI techniques will yield robots that can do essentially all human work better/faster/cheaper than a human could.

When discussing that possibility, I wonder if our current statistics and vocabulary are really sufficient? After all when you talk about employment, wages, productivity, etc we're thinking about governments and corporations. We cannot forget that governments and corporations are human power structures and one of the biggest forces pushing them forward is the desire of the rich and powerful to become richer and more powerful.

If AI really begins to threaten the supremacy of its human masters, don't you think those same masters will suppress the AI and try to keep the status quo? After all a CEO with a billion dollars in the bank doesn't come to work just to collect his/her paycheck, they come to work to be the boss.

Modern humans have been shown to be strongly, and possibly entirely, motivated by relative wealth and power. The millionaire among billionaires feels poor, the wage earner among the unemployed feels rich.

So maybe, however inefficient, we will continue to employ each other only because it feels good for the employer to retain power over the employee? Maybe we will reject a super-efficient robot-produced future because we don't know how to form a society unless it's founded on a human power structure. Maybe we'll find things for each other to do just so that we can make each other do them.

I agree that if you enumerate all of my current material desires it's possible that a future robot will be able to fulfill them. But I cannot anticipate the societal evolution of desires and how much power over other people plays into that.

I don't think many managers and executives enjoy their jobs because of the power they have in the organization. I think they enjoy their jobs by being proficient at their work and getting rewarded for it, like the rest of us.

Did he really show a graph of increasing employment in women, decreasing in men, and then say "Let's ignore the women. Why on earth could male employment be going down?"

Erm.... maybe because... women were taking their jobs?

I guess someone in comments mentioned that it is something more like more capable women replacing less capable men. It is more reasonable in itself. But as I see it leads more inequality as more capable people get together, raise families and generate better income and lifestyle for them. At the same time situation for less capable people further deteriorates.

I think there would be studies about the impact of increasing professional double income couples on rising income inequalities in society at large.

Because in the period of interest (2005+) the female job participation rate was declining as well; female employment reached the peak at around 2000.

>The report linked above is from Ball State University, and argues that, while manufacturing has thrived, automation has reduced the need for workers so much that lots of them have been laid off:

>Had we kept 2000-levels of productivity and applied them to 2010-levels of production, we would have required 20.9 million manufacturing workers. Instead, we employed only 12.1 million.

>The report gets summarized in a few places as “13% of job loss is due to trade, 87% is due to increasing productivity/automation”, which seems like a fair summary of some of its claims.

It's a fair summary of Michael J. Hicks and Srikant Devara's claims, but if you take a closer look at the study's model, they conflate cost savings from offshoring with cost savings from automation, which would naturally make job losses from automation seem relatively larger.

They also lobby for a reduction in corporation tax on page 7.

Great article that seems to leave the question unresolved.

If technological unemployment is not a large factor; but inequality within advanced economies is rising, then what other factors are at play? Are the changes necessary to deal with those causes the same changes that would be needed to deal with an all-knowing AI in 2150?

inequality is rising because some people are simply more skilled than others.

More data supports the hypothesis that inequality is rising because some people have more money than others, and currently it's easier to earn significant returns from capital than from labor. Of course, not everyone agrees, but Piketty et al have some evidence for the hypothesis. Is there anyone with data-driven support for the claim that changes in overall inequality are primarily driven by divergence in high-skill vs. low-skill labor incomes? To be clear, I obviously know there is a difference in high-skill vs. low-skill labor incomes, but it's not clear its magnitude is enough to account for the changes seen in macro-scale indicators like the Gini index.

and that money, where did it come from? look at fortune 500, what percentage inherited it? what percentage inherited it from someone further down the ancestral line than their parents? rich people without skill will squander their fortunes.

Interesting question.

I'm not sure, but there have been a lot of economists who disagree with Piketty and written about where he goes wrong. I assume that this implicitly proves the other direction, that capital growth isn't larger than wage growth, or whatever, so it might be a good idea to look there.

or the gains for being lucky and fortuitous are more and more extreme, and the cost of being left out is more bleak.

The Economist just ran a profile on Chris Hughes, Mark Zuckerberg’s second-year roommate, who, by virtue of being his roommate, lucked into a huge net worth from the early stage creation of Facebook.

https://www.economist.com/news/books-and-arts/21737016-and-d...

OK - would you agree that this presents us with an ethical problem?

How about a societal/practical problem? Is an unequal society sustainable?

What actions can be taken to mitigate the ethical and practical problems presented by inequality?

unequal society was sustainable exactly until now. there is no evidence whatsoever that it won't still be.

If you consider the unnecessary death and suffering of other people, and major environment impacts, to be "sustainable" then yes, unequal society has been entirely sustainable.

In practice the inequality has lead to massive unrest (eg civil war), disease, and famine time and again.

You can't honestly look back at history and consider that inequality has been anything other than a cause of immeasurable misery.

from my point of view all those deaths and massive unrests resulted from various groups trying to do away with inequality, usually by taking what's not theirs.

True, but...

Lenin (for example) didn't walk into a prosperous, free, egalitarian country, and persuade everyone to rise up against the existing system. The inequality, poverty, and misery gave rise to people willing to listen to Lenin, and to try what he was selling.

I'm not excusing Lenin. I'm not apologizing for him. But without the existing inequality, he would have gotten no traction.

he also wouldn't have gotten any traction if people had no concept of envy. yet eliminating envy is just as stupid as eliminating natural differences between individuals.

> OK - would you agree that this presents us with an ethical problem?

My instinct is that inequality is not inherently unethical if it's a fair reflection of people's skills and/or efforts. If people are so poor that their needs are going unmet that's an ethical problem, but those people are not made any worse off by the existence of wealthy people.

> How about a societal/practical problem? Is an unequal society sustainable?

An unequal society where people still feel like they're contributing something seems to be sustainable. I think we're hitting an issue where it becomes clear that some people have essentially no ability to contribute to society, and that might be unsustainable, but I don't think we can or should turn back the technological clock to make them able to contribute. At which point we're down to handouts or make-work jobs, and while I'm all for supporting the worst off, that doesn't address the fundamental problem.

>those people are not made any worse off by the existence of wealthy people. //

How does that work? If there are wealthy people using more than their share of resources and energy, staying wealthy on the back of other's labour (rent-seeking, capitalism), then surely the presence of those wealthy people makes resources/energy more scarce and the need to pay the excess in order to maintain the wealthy means all goods/services/resources are more costly.

If one person owns the farmland and charges rent without giving back value then all the workers must work harder than if the farm is owned cooperatively.

Most of the cost of most things - especially most things consumed by wealthy people - rests much more in labour than in physical resources; poor people are net sellers of labour so wealthy people inflating the price of labour helps them more than it hurts them. Land is an exception (being in limited supply) and I do think a land value (Georgist) tax would be morally desirable, with the proceeds distributed equally to everyone.

That could only be said with some certainty if we had equal opportunity. Most western societies are far off the mark in that regard.

i agree. free market capitalism should replace current corporatism. but that's currently impossible - there will always be lobbying and special treatments while the goverment is able to interfere with the market.

"Some people are more skilled than others, therefore inequality exists" - sure. But why would that predict _rising_ inequality?

because people continue to live. those skilled in making money make more money, those not - lose more money.

Paul Graham says that technology is an amplifier. And technology has been giving increasing leverage over time.

what research can you cite supporting this hypothesis?

Rothbard, Hayek, Hoppe, et al.

So, one issue I had with this is that a big part of the argument is that "when I see shocks to employment, I don't see shocks to LFPR", but we do definitely see shocks in unemployment rates, and I don't think it's crazy to think that these shocks to employment may transfer to LFPR over long periods of time as different people face different levels of success finding adequate work.

I think the strongest argument we can make is that whatever is going on with LFPR has been going on for 50 years, and it does seem to indicate a more difficult job market where people are finding that the jobs available are not better than the alternative.

He may be misreading stagnant productivity growth.

Manufacturing productivity is measured in dollars[2], and prices of manufactured goods have been going down[1] (largely from China competition):

[1] https://ourworldindata.org/grapher/price-changes-in-consumer...

[2] https://www.bls.gov/lpc/faqs.htm#P03

We have three sectors, agriculture, industry and service.

Technology has replaced people in agriculture, at the same time industry jobs exploded. Most people were shifted.

Technology has replace people in industry, but there were services to provide - lower paying - jobs. Most people were shifted.

Now technology replaces jobs in services. There is no "next" sector.

Employment is only a result of an individual wanting someone else to do something for them, because they don't have the time or skill to do it themselves. In exchange for someone doing this thing for them, they compensate them with something of value, normally payment in the form of currency. When you consider employment from this angle, I find it extremely hard to believe that a day will come where the majority of humans on this planet won't want help from other humans to accomplish their goals as they go about their daily lives. Yes, employment in the form of large corporations giving out biweekly paychecks may not exist anymore, but employment as a way to help your fellow human will never cease to exist.

I don't know, eventually all unskilled labor will be replaced by machines, and in order to give someone else value you'll need skill.

Sure. This kind of "employment" - or, rather, trade - is fundamental to humans as a socializing species. This can't go away without fundamentally changing what it means to be a human. But this is not the kind of "employment" people usually talk about.

What can, and should go away, is the need for employment to survive. I think we can also imagine the times in which employment isn't necessary in order not to suffer (crappy food, cold in winter, lack of healthcare), and to keep upward mobility open. The optimistic view of eliminating jobs is all about eliminating being forced to work to live a happy life, not about eliminating the entire possibility of work and trade.

> I find it extremely hard to believe that a day will come where the majority of humans on this planet won't want help from other humans to accomplish their goals as they go about their daily lives.

If robots are better at helping people accomplish those goals then why would people ask other humans for help rather than robots? I already choose to book my holidays with machine rather than human assistance, pay for my groceries via machine rather than human.... For the moment I pay humans to do some things (cleaning, food preparation, art production, ...), but it seems entirely conceivable that machines could get better at all of those as well.

>I already choose to book my holidays with machine rather than human assistance, pay for my groceries via machine rather than human

It's funny you should mention this. My last job was at an OTA (online travel agent). They started out with a website to book holidays and an intent to streamline the crap out of the process catering to individuals like yourself. They hired a skeleton call center staff to deal with unforeseen problems with an intent to rely upon them as little as possible.

Over time they realized that a lot of people were calling up to do bits and pieces of the booking over the phone that they could have done online.

Initially they tried to steer them back to the website but they realized that they were mostly dropping out of the sales funnel.

It turned out a lot of people actually wanted to talk to a human when spending $1,500 on a holiday. Shocker.

So, they changed tack and ramped up call center staffing, advertised a free phone number on the every page and let people essentially decide themselves where in the sales funnel they wanted to talk to a human - the whole way through or, like you, not at all.

This company had a really smart approach in other areas too about when to use automation and when not to.

>For the moment I pay humans to do some things like cleaning

You say that as if there has been a multitude of advances in home cleaning technology since the 1950s. I'm skeptical of the notion that Roombas took anybody's job or that I'm about to get a toilet cleaning robot any time soon.

> It turned out a lot of people actually wanted to talk to a human when spending $1,500 on a holiday. Shocker.

I'm surprised, and I wonder what the age breakdown of those people looks like.

> I'm skeptical of the notion that Roombas took anybody's job

It must have on the margin, surely? I was on the fence about getting a cleaner for a long time; it wouldn't've taken much to push me into not doing so (and many of my friends in similar circumstances, some of whom have Roombas, don't).

> or that I'm about to get a toilet cleaning robot any time soon.

Toilet cleaning will probably be one of the last things to go as the worst-case scenario there is pretty bad, but I read some interesting blog posts a year or two back about a robot that can pick up clothes from the floor and run a laundry cycle (with the help of a QR code sticker on the washer). I don't expect a robot to replace a human cleaner immediately, but I do expect the gradual automation of more and more domestic cleaning tasks, which will push more and more people across the line where it doesn't make sense to hire a human cleaner.

>I wonder what the age breakdown

Skewed, but not skewed towards the elderly nearly as much as you clearly think it was.

Not everybody is like us.

>It must have on the margin, surely?

Not necessarily. It might just mean that, on balance, people who were never going to employ a maid in the first place live in slightly less filthy houses.

>Toilet cleaning will probably be one of the last things to go as the worst-case scenario there is pretty bad

I can't actually see any evidence that anybody's even trying. It's not like driverless cars where we're being told every year that they're, like, 5 years away tops. Who is even working on this? As far as I can tell, nobody.

> Not necessarily. It might just mean that, on balance, people who were never going to employ a maid in the first place live in slightly less filthy houses.

But, like, there's no massive discontinuity in how likely people are to employ a cleaner (is there)? I'd expect that distribution to be a bell curve like anything else, with lots of people on the margin (and I'm pretty sure I was one of them), and the introduction of the Roomba shifts where that line is slightly.

> I can't actually see any evidence that anybody's even trying. It's not like driverless cars where we're being told every year that they're, like, 5 years away tops. Who is even working on this? As far as I can tell, nobody.

I'm not aware of anyone working on toilets, but I think that's the highest hanging fruit. You already mentioned Roombas, and I mentioned LaundryBot or whatever it's called; I'd expect us to progress gradually from there to things like dusting, window cleaning, oven cleaning and toilets in that order or close to it.

>But, like, there's no massive discontinuity in how likely people are to employ a cleaner (is there)? I'd expect that distribution to be a bell curve like anything else

Even if it was a bell curve (and I see no reason why reality ought to be that neat), there's no reason to suppose it doesn't just lead to slightly cleaner houses rather than fewer maids.

> You already mentioned Roombas, and I mentioned LaundryBot

Roombas are singularly unimpressive as a means of cleaning your house (I know of nobody who actually owns one, and the reviews on amazon are... only really impressiveif you are without stairs and have a pet hair problem). Googling laundrybot gives me some japanese thing invented in 2015 that costs $280,000 and, product-wise, doesn't seem to have gone anywhere since then.

It's almost as if the automation hype isn't matching reality.

I wouldn't be surprised if in 20 years time nothing has really changed in this space either. Roombas came out in 2002 - 16 years ago. Significant potential I'll be dead by the time a robot that can clean my toilet exists and I'm not exactly old.

> there's no reason to suppose it doesn't just lead to slightly cleaner houses rather than fewer maids.

If we assume people's willingness to hire a cleaner is a function of how dirty their house is (and how could it not be?) then the two are equivalent. The only way roombas could not be putting people out of work on the margin is if there's some sort of mysterious discontinuity, and why would we assume such a thing?

>If we assume people's willingness to hire a cleaner is a function of how dirty their house is (and how could it not be?)

Because:

a) lots and lots and lots of people clean their own house and maybe they're the ones buying the vast majority of the roombas, not people with maids.

b) just because your roomba cleans up your pet hair really well upstairs doesn't necessarily mean you will want your cleaner to stop coming on tuesdays.

c) tolerance for dirt to me seems like a textbook example of something that has a high demand elasticity

> lots and lots and lots of people clean their own house and maybe they're the ones buying the vast majority of the roombas, not people with maids

Surely the one substitutes for the other. The thought "I'm spending longer than I want to cleaning, and my house still isn't as clean as I'd like, and I've had more money lately, I can afford to..." could be completed "hire a cleaner" or "buy a roomba".

> tolerance for dirt to me seems like a textbook example of something that has a high demand elasticity

What do you mean? Different people will have their threshold in different places, certainly, and the amount they're willing to pay will vary, but I don't see how that changes anything.

I'm generally struggling to understand what your model is here. Are you saying people make the decision about whether to hire a cleaner when they're born and then never change it?

>Surely the one substitutes for the other.

If a roomba approximated a maid, perhaps. It very much does not.

>I'm generally struggling to understand what your model is here. Are you saying people make the decision about whether to hire a cleaner when they're born and then never change it?

No, I'm saying that:

A) building a neoclassical economic model on this at all is pointless because of the assumptions you'd have to make. You've just tried to make about four assumptions in order to build your model and they weren't just poor assumptions, they were cringeworthy assumptions.

B) the total number of people who actually own roombas or roomba like devices suggests to me that the effect on cleaner employment is likely, in any case, to be < noise.

> If a roomba approximated a maid, perhaps. It very much does not.

It fulfils some of the same functions, some of what some buyers are buying for. A milkshake is not a close approximation to a muffin, but McDonalds found that they did substitute for each other in terms of what people bought for breakfast.

> A) building a neoclassical economic model on this at all is pointless because of the assumptions you'd have to make. You've just tried to make about four assumptions in order to build your model and they weren't just poor assumptions, they were cringeworthy assumptions.

You're essentially saying "give up trying to understand the world", to which I can only respond "no" (compare the opening quote on http://slatestarcodex.com/2015/08/24/probabilities-without-m... ). I think all the assumptions I gave were true and reasonable; if you can actually make a case for why any of them is wrong I will listen, but just dismissing them is not constructive.

>A milkshake is not a close approximation to a muffin

I would say it is. They have similar calorific content, they are similarly portable and they are similarly stuffed full of sugar.

>You're essentially saying "give up trying to understand the world"

...and now you're using a cringeworthy straw man.

I'm saying "if you're going to make a claim like "roombas are killing jobs", it probably shouldn't be based on a model that makes 4 really reaching assumptions, each of which are completely unproven and quite likely wrong.

Oh, but the New York Times and a few economists thinks robot kills jobs, so this fits in with everybody's groupthink, so that's ok, right?

>I think all the assumptions I gave were true and reasonable; if you can actually make a case for why any of them is wrong

This is reminding me of the conversation I had with my vicar who asked me to disprove god's existence if I thought I was so smart.

>just dismissing them is not constructive.

I disagree. I think it's very constructive to dismiss unproven assumptions that are probably wrong, but neatly fit into some kind of groupthink. It's how a lot of science is advanced.

> I would say it is. They have similar calorific content, they are similarly portable and they are similarly stuffed full of sugar.

Exactly - they're superficially very different, but what they provide to the buyer is very similar. Just as with Roombas and cleaners.

> it probably shouldn't be based on a model that makes 4 really reaching assumptions, each of which are completely unproven and quite likely wrong

So put up or shut up: which assumption do you think is wrong, and what do you think the appropriate replacement is?

> This is reminding me of the conversation I had with my vicar who asked me to disprove god's existence if I thought I was so smart.

It's reminding me of the tortoise who denies modus ponens - if you deny the ability to make inferences then where do we even go from there?

> I disagree. I think it's very constructive to dismiss unproven assumptions that are probably wrong, but neatly fit into some kind of groupthink. It's how a lot of science is advanced.

Saying "groupthink" as a meaningless buzzword isn't constructive either. Science advances by questioning assumptions but you have to actually do the work of considering an alternative; science doesn't advance by people introducing a question like "but what if effects didn't follow causes" and then just stopping there.

>what they provide to the buyer is very similar. Just as with Roombas and cleaners.

Roombas do not deep clean carpets, cannot traverse stairs, can't clean bathrooms at all, are blocked by clothing and other heavy objects, closed doors, etc. and cannot clean anything that is not on a floor.

That is not what I would qualify as "similar to a person".

>So put up or shut up

Why? I am not the one trying to advance an agenda of "roombas are clearly putting maids out of business". I'm the skeptic.

I think it's entirely possible that all of your assumptions are wrong and probable that at least half of them are.

>if you deny the ability to make inferences then where do we even go from there?

I might start with questionnaires asking roomba owners about their cleaner hiring practices in order to confirm or deny one of the key assumptions underpinning your thesis. It'd be a good topic for research, but not one I have seen pursued.

>Saying "groupthink" as a meaningless buzzword

It is not a meaningless buzzword. The idea that robots destroy jobs is commonly accepted because it is popular. I could find 20 or so articles in a couple of minutes advancing the idea, whereas the skeptic's view - that it doesn't, is barely represented in the media.

Most of what appears in the media advancing this agenda is unsourced and the little that is sourced comes from things like that junk study from Ball State U.

I mean, does that not sound like groupthink to you? It does to me.

>Science advances by questioning assumptions but you have to actually do the work of considering an alternative

Well, if you're really interested in proving your point - why not fire off some questionnaires to roomba owners - did they have a cleaner before they bought the roomba? did they fire the cleaner after? did they cut the cleaner's hours? This seems like a more productive line of enquiry - reality is usually more accurate than shit models.

Roombas are great if you run them constantly and you have more than one in your house; but even a single one saves a ton of time and effort.

What they "shine" at is keeping up on cleaning away surface dirt and such. Instead of needing to vacuum a couple of times each week, you just run the Roomba(s) a few times a day. It takes away a dull and difficult chore.

Occasionally though, you do have to break out the vac for a deep clean (once a month or so) - and on the odd occasion the carpet cleaner for a good scrub. And you have to keep up on the cleaning/maintenance/emptying of the Roomba as well.

But overall, at least for my family's household, having our Roombas have been a plus.

re: Toilet cleaning

What's needed there is not so much a robot, but a better toilet or materials that don't get dirty or are self-cleaning.

Or maybe a chemical - what might be interesting would be something like those blue discs you drop in, but coupled with something like Rain-X (for windshields). Make it so nothing can stick to the bowl...

That of course doesn't solve the external surface cleaning process, but it would be a nice start.

I don't know about that. I'm used to thinking in terms of automation and I can easily envision a semi-automated way to clean toilets. You basically need to remove the cleaning fluids, etc. faster than you apply them. A set of conformal brushes, high pressure pump and high flow vacuum should do it.

The question is: is it faster/more cost effective than a human? If I could build a few of these things in my basement, could I get commercial toilet cleaning contracts and make more money than the traditional cleaners?

The answer is probably "not yet," but I haven't looked through any commercial cleaning publications to see what's on the market.

Individuals are not just exchanging labor. What we want is result of combining factors of production together (capital, labor, land or natural resources).

What happens if the things people want today are 70 percent labor and 20 years from now just 40 percent labor (this is called capital deepening). Will the return of capital decrease and how much wage share decreases? Is government spending or wealth transfer increasing in the economy?

This reminds me of the old joke that GDP goes down when a man marries his maid. But this article manages to not only make that error, but ignore how clearly the data suggest that there are a fair number of men who are housekeepers and are going off the labor market for similar reasons.

Come back when the retirement age drops below 50.

At the moment it is going the other way.

The difference I see is that historically automation was used to increase productivity. I don't see that in the current wave of "AI"-based automation. There's an emphasis on speed, accuracy, and working with more unknowns, but there isn't any talk of getting more work done. That makes it scary.

I'm not sure what you mean.

"Speed" is almost directly exactly analogous to "increased productivity" isn't it?

Also accuracy - reducing scrap means more can be produced with less materials in less time.

"working with more unknowns" - doing things that couldn't be done before.

These all seem to be kinds of "getting more work done".

>The difference I see is that historically automation was used to increase productivity.

Is this true? The stocking frame (and the reactionary group of Luddites that it incited) didn't really increase productivity. It just did work humans were doing for less money than a human salary.

I guess I also just don't see much difference. All that will happen will be again a shift away from many kinds of jobs (the easily automatible ones) to different kinds.

I mean, there are all kinds of jobs now that didn't exist 50 years ago (software engineer, social media manager, influencer, etc. etc.)

Yes, all kinds of jobs - but are there actually a substantial number of people doing those jobs compared to the number of jobs that are easy to automate away?

I don't see why it matters. In the end, there won't be massive job loss to automation.

Automation right now can eliminate most farm laborers. So why are there still so many farm laborers in the world? Because

a) human wages in most of the world are cheaper than expensive machinery

b) There isn't a profit motive to automate away jobs that pay so little

c) structure of society is such that many technologies take decades upon decades to trickle out to most people in the world (talking the whole world here, not just Euromerica).

In addition, the service economy won't be automated away for a long time because human contact is valued by lots of people in that field. Service economy scales with the purchasing power of communities and AI is going to create a consolidation of purchasing power wherever it is introduced.

> didn't really increase productivity. It just did work humans were doing for less money than a human salary.

this is by definition "increasing productivity" isn't it?

> All that will happen will be again a shift away from many kinds of jobs (the easily automatible ones) to different kinds.

If that's true, then what will the shift be to? These are interesting questions. And how long will the shift take? Because I think there is a general feeling that current levels of inequality - the "hollowing out" of the middle class - are due to trends in automation.

If that's what increasing productivity is, then AI will surely increase productivity.

I think the shift will probably be towards internet-based jobs. Content creation / software development / content curation (think social media managers) / customer service, the things that it's hard for AI to get right.

It will probably take longer and definitely leave some malcontents but, so did the stocking frame.

What will.the shift be to ?

A few possibilities, at least short term:

Some manufacturing returns to the US with help of robots. Maybe even as an exportible industry.

Maybe it's possible to simplify those complex jobs that more people could do them?

It's possible for the some of the new,highly automated industries to be a net job creator. For example one scenario for self-driving is that it will be partly remotely-operated ,i.e. jobs , instead of people driving themselves.

Things may become cheaper and people will spend the money left on human based experiences/services ? Or even an whole economy based on human interaction because that's the one thing robots couldn't do?

Or like we always did , we'll find some useless thing(like consumerism) that will create jobs and build society around it ?

>Is this true? The stocking frame (and the reactionary group of Luddites that it incited) didn't really increase productivity.

It was largely used to employ cheap labor to produce crap. The luddites produced higher value artisinal goods.

The luddites weren't reactionaries intent on destroying technology any more than strikers are protesting against the concept of 'work', they were just saboteurs trying to improve their negotiating position.

The whole thing about them being anti-technology is historical revisionism. Similarly, French taxi anti-uber strikers in 2013 were literally depicted by Forbes as being "anti smartphone" - the motive for the lie being the same in both cases.

The reason why low wage employment and crap clothes were more economically successful at the time had more to do with the economically hemorrhaging effect of the Napoleonic wars - which led to cheaper, more desperate labor and a demand for cheaper goods. It's a bit like how McDonalds gets a sales boost during recessions while mid range restaurants take a hit.

Okay, I'm not a scholar on European history, so you're probably right. It was just an example. Still there have been automations in the past. We're all still here.

Well actually, we're not all still here!

Those shifts - i.e. the industrial revolution - were enormously disruptive!

Assuming that this is happening; or going to happen, how can we smooth it out a bit?

And of course; that is on the basis that this is not a science fiction AI special; in which case an even more thoughtful approach might be required to minimise the damage and maximise happiness; or whatever variable you prefer.

I worry that attempts to smooth it out can actually cause worse disruption.

Like, let's say we know factory workers will be laid off, so we need to allocate resources to retrain them.

Very likely the resources will not be enough for all the workers. So who gets them and who doesn't will become a political battle, with the left-out group resenting everyone in society and society ignoring their needs because it reminds them that they left them behind.

And then you build a society where one group's triumph is based on another's suppression, and you can see where this is going...

I absolutely agree. I think that the whole notion of automation being responsible for job losses is also lies and misdirection. The man behind the curtain is telling us to avoid paying too much attention to austerity, outsourcing, offshoring and union busting.

In 2011, for instance, amid a period of labor unrest, Hong hai announced to the world with much fanfare that they were going to soon eliminate 1 million jobs with robots. They subsequently created about 400,000 jobs. Nobody noticed.

I agree with everything except one minor detail. Not sure if outsourcing and offshoring are such bad things. I mean, if someone will work for less than you, you either need to do better work, do cheaper work, or find different work.

In the luddites' case that 'person' was willing to work for less using the weaving machinery because they were desperate. The Napoleonic Wars made them lose their livelihoods, homes and families.

After the industrial revolution, the enclosure movement was used to strip peasants of their land. After being dispossessed, they were then desperate enough to accept the paltry wages and disgusting working conditions offered by the factory owners.

Today, a Chinese factory worker may be saving money because they need to care for their infirm grandmother living in a cancer village because the state provides little to no social welfare support for her.

Do you think it's germane to argue on a moral basis that because China provides no welfare for that grandmother that American factory workers ought to take a pay cut?

Personally, I think trade policy needs to be intimately connected to human rights and environmental and labor conditions. No improvement in those sectors, no deal. That isn't the world we live in though. We live in a world where Obama tried to excuse Malaysian slavery in order to try and enact a trade deal so that American workers could be forced to compete with slaves in Malaysia.

>Personally, I think trade policy needs to be intimately connected to human rights and environmental and labor conditions.

It's never ever ever going to happen. Be careful what you wish for. If what you wanted to happen happened, you would be paying over $100 for the same shirt that costs you $6 in Walmart right now.

I also want to point out that, yes, in a globally competitive world wages become globally competitive. If a Chinese worker with an infirm grandmother is surviving on a lower wage than an American was making without an infirm grandmother, then that American needs to find better work, do better quality work, or take a pay cut. We would never force consumers to buy products produced only by ethical means or only by human-rights approved conditions (if that were the case, no more chocolate!). Why would we do that to multinational companies?

There aren't other options available. No President is going to preside over the term where all the trade deals were made alongside human rights conditionals and all the goods skyrocketed in price leading to another Great Depression.

> We live in a world where Obama tried to excuse Malaysian slavery in order to try and enact a trade deal so that American workers could be forced to compete with slaves in Malaysia.

You can boohoo about it all you want, but your chocolate, your coffee, your clothes, and your computer parts are all made by or had their materials harvested by slaves, and most likely you would scoff at voluntarily paying twice the price for all those commodities. I'm not saying everyone should be enslaved. But if consumers have the right to purchase such ill-made goods, then companies have the right to offshore their labor forces to such places too.

Where did you pull your $100 figure from? Suppose this accounting of environment and labor results in paying say $50 for a shirt that would be durable, then this might simply be awesome. Fewer, more durable goods that in the long term may be cheaper through automation. The answer to slavery may or may not be automation but certainly isn't "well, though, that's capitalism for you."

>you would be paying over $100

$100 is sheer fantasy. It might double the price of a shirt from $6 to $12.

>There aren't other options available. No President is going to preside over the term where all the trade deals were made alongside human rights conditionals

This is actually how it used to happen and this is partly how TPP was killed in Congress (by attaching an anti-slavery rider).

>You can boohoo about it all you want, but your chocolate, your coffee, your clothes, and your computer parts are all made by or had their materials harvested by slaves, and most likely you would scoff at voluntarily paying twice the price for all those commodities.

No, I'm more than happy to pay a extra for chocolate, coffee, clothes and computer parts if I get to live in a world where labor exploitation, environmental destruction and slavery is made more difficult.

Are you really willing to go on the record and state that your preference is for the opposite?

> It's never ever ever going to happen. Be careful what you wish for. If what you wanted to happen happened, you would be paying over $100 for the same shirt that costs you $6 in Walmart right now.

Honestly, IMO scaling down western consumerism by a lot (for example, by fixing and mending the same small set of shirts for years) would be a small price to pay for the resulting increase global happiness. We more or less lived like that during communism in Poland (chocolate, coffee, oranges were uncommon, meat was not an obvious choice for dinner, we fixed appliances, electronics and cars for decades) and I think people might have actually been happier back then.

The TPP contained provisions for ensuring worker's rights, including the right to unionise (chapter 19), along with environmental standards (chapter 20) and mechanisms to increase transparency and fight corruption (chapter 26). All of these issues will suffer from the US having ceded their influence to China by abandoning years of painstaking effort - and make the pharmaceutical and tobacco industries, who fought the TPP and lost, very happy.

>Is this true? The stocking frame (and the reactionary group of Luddites that it incited) didn't really increase productivity. It just did work humans were doing for less money than a human salary.

Mechanisation of the textile industry led to vast increases in the quantity of production. Prior to the industrial revolution, most peasants would have owned only one set of clothes and replaced them only when they were entirely beyond repair. Cheap cloth created demand for wardrobes and dressers, because a lot of people were now in the peculiar position of owning more clothes than they could wear at once.

The luddites were ultimately wrong because of Jevons paradox - labour efficiencies in the textile industry vastly increased the demand for their product, and with it their demand for labour. Hand-loom weavers didn't lose their livelihoods to the power loom, they just made ten times as much cloth for twice the wage.

Automation and mechanisation in other manufacturing sectors likewise stimulated vastly increased consumer demand, contributing to a virtuous cycle of increasing profit, wages and living standards. Over the course of the 19th and 20th centuries, the efficiencies of mass-manufacturing gave ordinary people secure employment and access to a dizzying array of consumer goods. The more stuff we made, the more stuff we could afford to buy.

I don't see AI/ML following the same pattern. A lot of tech-heavy businesses seem to be startlingly close to zero-sum. Uber has decimated the taxi industry, but it doesn't seem to have vastly increased overall ridership - the most generous figures I've seen suggest perhaps a 40% increase in some markets. Google and Facebook have become multi-billion dollar companies through advertising revenue, but the overall growth of ad spending across all media has barely outpaced inflation, with the newspaper industry being the most obvious casualty. I can't think of many major tech companies that didn't kill an industry.

I'd love to be proven wrong.

>I don't see AI/ML following the same pattern.

AI/ML which can mine through sports footage to create highlight reels and funny clip collections, etc will create tons of consumer goods (videos) which can all be monetized on Youtube. Just one example.

How about Netflix? They use AI/ML quite a bit. It seems to have driven quite a lot more people to watch movies and tv than before.

Amazon uses AI/ML in its Echo, which was quite a high-selling consumer good last holiday season.

> How about Netflix? They use AI/ML quite a bit. It seems to have driven quite a lot more people to watch movies and tv than before.

... and the word 'cable cutters' is being thrown around with a lot more intensity, plus various charts of rapidly declining cable TV viewership.

> Amazon uses AI/ML in its Echo, which was quite a high-selling consumer good last holiday season.

[gestures at the latest brick-and-mortar bookstore news with B+N on the rocks and Borders in its grave ... given AMZ got it's start as a bookseller.]

A pessimistic response to be sure (and one that I'm reluctantly making), but there's coincidence, and then there's a pattern that starts emerging whenever a disruptor unicorn cements its place into an industry.

A 40% increase seems pretty vast to me.

> There's an emphasis on speed, accuracy, and working with more unknowns

How is this not getting more work done? It sounds exactly like more work is getting done to me.

If you produce one report in a day, or a computer does it more accurately in 50ms and you don't need the person, the same amount of work is done.

That doesn't make any sense to me. So, if a tractor digs a hole in 5 minutes that a person could dig in a day, the same work is being done because a person could have done it? If a horse carries a person 10 miles in an hour and a car does it in 10 minutes, the same work is done because the horse did the same work as the car? Does an ATM do no work because a teller could do the same thing, just slower and at higher cost?

By your standard, does the industrial revolution mean that more work is being done?

"Work being done" refers to the output, not the way in which that output is produced.

For example, let's say that the task is to add up a list of 10,000 numbers. It might take a human a while to accomplish this accurately. A computer won't take much time at all. The computer is accomplishing the same work as the human in a much shorter amount of time. The computer could handle a vast number of such tasks per day for a lower price than any human, with better accuracy.

Similarly, imagine hauling cargo across the country. A human could put the packages on their back and walk, or alternatively a human and freight train team could haul them by rail. The human and freight train team accomplishes vastly more work than the human alone, and is able to haul much larger and heavier packages at a much faster speed.

Then you make the computer do a thousand reports in a day, and get a qualitatively new (not necessarily better) ability of working with real-time data. Increased productivity is an enabler for things that previously were not possible or did not make economical sense.

I do not want to be employed. I want to do with my time what I want, and when I want. I do not want to continue living in the age of want, I want to live in the age of plenty. I think it's fair to assume this wish is pretty universal.

So I do not give a rats arse about 'productivity', and neither does anybody else, what we want is to live in paradise, today.

Paradise, for me, is when I can have all of the below as a matter of basic human convenience, 'free' as we call it now in the age of want:

* min 4000 calories of excellent and varied nutritious food whenever I want it

* 1kl of 60℃ water / day to my exclusive disposal, I will generally lie around in it

* 1l of alcoholic beverage, typically high hopped beer, I'm not fussy

* sex on demand with a like minded individual, I'm not fussy, hotness would be nice but I'll be happy with someone less ugly than me

* I want to be able travel to any place in the solar system that humans have built homesteads, on demand, by getting on the nearest bus to Titan

* I want to live as long as I please, in excellent health until life literally bores me to death and I decide to die of some ancient crippling disease as an expression of art, or some other bullshit like that.

Now, if I can have the above, you may own all the AI and tech you want, no worries.

Reply on news.ycombinator.com