This is self-reported: remember the 1000 Indians that were watching people shop at the Amazon Go stores?
Amazon told us it was all "sensors" because it fit their company narrative to do so.
I am not saying that Amazon doesn't have 750k robots and hasn't laid off 100k people... but they usually have some seasonality, plus, the quoted number is from 2021, the height of at-home shopping.
"The world's second-largest private employer employs 1.5 million people. While that's a lot, it's a decrease of over 100,000 employees from the 1.6 million workers it had in 2021"
The 1000 Indians story was completely incorrect though. They has 1000 people in India tagging the videos after the fact for improvements to the machine learning algorithms. They weren't watching in real time.
Debunked where? The only response is from Amazon (in the same article), who are disputing the number but still admitting that the team did intervene when needed.
From TFA:
"While Amazon insists these reports are “erroneous,” it doesn’t deny that humans aren’t involved with the process at all. Instead, Amazon says its workers are tasked with annotating AI-generated and real shopping data to improve the Just Walk Out system — not run the whole thing. “This is no different than any other AI system that places a high value on accuracy, where human reviewers are common,” Dilip Kumar, the vice president of AWS Applications, writes in the post."
Don't see how this refutes any claim. If anything, Amazon is confirming they have human intelligence involved, they're just leaving the quantity/portion ambiguous, as behooves them. Do you have Amazon equity?
I'm certainly not defending Amazon, I'm defending AI. What they build was impressive and worked for the most part (I saw the behind the scenes). It was just really expensive because you needed high resolution full coverage of the entire store.
I just find it disingenuous for people to pull out this example as "why AI doesn't actually work" when it's not an example of it.
For what it’s worth, I purchased from one of these Amazon stores every day for over a year. After leaving the store, it would typically take 2+ hours before my purchases were registered in the Amazon app. I agree that there probably was a large degree of automation but based on how the store worked, there didn’t need to be an army of “real time” reviewers for humans to be involved in each individual purchase… it’s totally plausible based on what I experienced that Amazon did have an army of people involved in the standard purchase path.
I think it was from a video I watched a day or two ago, describing how with those uber and lyft and other order delivery services that consumers apparently are placing those orders wtih the services, picking up the items themselves and walking out, and then when the actual delivery person arrives, order is gone, reported as missing, the consumer also reports similarly, even though they picked up the orders themselves, that this is apparently happening more widespread commonly that still is vulnerable with business infrastructures to not prevent the exploit from taking place, I am curious how this may affect industry standards with regards to this concept because I never witnessed it but I wouldn't be surprised if it is not even that widely realized about the issue yet thereby enabling more theft and loss and damages and whatnot.
How many times can one's order be 'stolen' like this before a pattern emerges and the account is banned? This is also fraud and because it requires using the internet, it's going to lead to felony charges.
If memory serves me well, I think some guys in Florida found out how many times they could do it and got busted for fraud against the delivery companies.
System caught the anomaly, humans investigated, law enforcement made the arrests.
My theory is that purchase tracking is done automatically via AI, while humans are on the loop to check the feed of people taking each item and verifying that the amounts are correct before charging.
It makes sense. Even if your entire AI system is 99.9% accurate, you want a human to detect and remove that one-in-a-thousand mistake.
I have a question: was the customer billed based on whatever the machine learning algorithm determined, or based on what humans decided after the fact? In other words, was the humans' job to just provide training data for future versions of the machine learning algorithm, or to manually fix the algorithm's errors and prevent over-billing?
IIRC, they could correct your bill after the fact.
I used to joke with others that shoplifted items should be retroactively free - as long as you self report after the fact. (Extra useful training data...)
Having not read the original story, this makes more sense.
"1000 Indians that were watching people shop at the Amazon Go stores" ...
Conjured up picture in my head of an Indian man inside an Amazon Go store sitting behind a registerless counter. What his purpose would be, I don't know.
Funny story. The amazon go stores do have a register at the counter now.
But as far as I can tell, they rotate positions. There's never really anyone standing at that counter. If you walk up to the counter someone will come from the sandwich counter or restocking the shelves or whatever.
> remember the 1000 Indians that were watching people shop at the Amazon Go stores?
> Amazon told us it was all "sensors" because it fit their company narrative to do so.
I think you, like many others, fell into the trap of just looking at the headline. The 1k people were labeling training data, not watching you shop like a puppet master..
Per the articles I read, they did more than label training data. They also did other things to facilitate the checkout process. They didn't get into what those things were however.
> "Associates may also validate a small minority of shopping visits where our computer vision technology cannot determine with complete confidence an individual's purchases," the spokesperson said.
The whole discussion around that headline has always been so confusing. It's honestly unclear. What I'd like to know is, what % of actual costumer checkouts required a human being vs what % was 100% automated.
Some of the headlines and negative discussions make it sound like 100% of the checkouts required human approval. If that is true, then that's clearly terrible. If on the other hand it was anywhere below 5-10%, then I think that's a fair target for training the model in more difficult edge cases.
Something like Waymo also has humans constantly monitoring and helping, but I don't see much discussion about that being "controlled by humans".
> Something like Waymo also has humans constantly monitoring and helping, but I don't see much discussion about that being "controlled by humans".
Constantly doesn't really make sense, riders can call for support and get a near instant response, but it's pretty obvious from some of the mistakes that humans aren't watching every ride in real time. There's video from inside a Waymo which is stuck in a lot repeatedly looping, discovering the exit it wants to use is blocked and then just driving to another exit before deciding hey, perhaps the exit I want to use is open (nope, it isn't) and returning again, and again, and again for example.
All the humans involved can see what's wrong here, but the Waymo driver software doesn't get it and so I think eventually a human (Waymo employee) has to drive the car out.
And yeah, there's definitely in San Francisco for example some people who believe it's all a Mechanical Turk again, even though that doesn't make any sense. In their heads, driving is another of those "uniquely human" abilities and so the only way to have AI cars would be fully general AI, and the real explanation has to be off-shore remote driving.
Would it even be feasible to provide a taxi service with remote human drivers? I get the remote driver occasionally dealing with slow movement, but I think full time driving at speed would be impossible due to latency issues. It’s not a remote piloted drone that is in the air and so has much less of a chance of hitting something by accident.
Otherwise you think someone would have tried offering taxis driven by people in India or the Philippines already.
I believe there's a startup saying they'll deliver cars to you, so, cars with no people inside them driven remotely, once there's a human inside the human is driving. I don't recall the name.
Waymo itself is clear that it has no remote driving capability whatsoever. The Waymo driver is in your vehicle, so any technical issue (e.g. phone network drops out, satellite fails, whatever) won't affect the vehicle's self-driving. It may give up and pull over and let you out, but since it's local any situation where the driver is disabled would be similar impact to a human driver - if a block of concrete drops off an overbridge and smashes the Waymo driver, that could easily happen to your Uber driver the same.
Hence their choice to employ humans to attend in person and drive a Waymo which gets into trouble. In fact the humans sometimes have had to "chase" a Waymo car that, unless it has been specifically told to stop, assumes it should try to complete the journey once it can figure out how. Waymo's remote support can reach into the driver's model and tweak it e.g. to label a stray traffic cone as just trash, not actually off limits - but they aren't driving the car.
"may also" still doesn't suggest that the remote observers were doing all the work and the computer vision systems didn't work and it was all a farce though. spot checking that the system works isn't the same thing as doing the job of the whole system
No one is suggesting that. The question is how automated was the system? According to the article 70% of all transactions were manually checked out because the algorithm didn't have enough confidence in its prediction. Amazon has disputed the number, but not the fact that the manual intervention was happening.
I wondered that, too, but I did see that Amazon revenue has continued to grow since 2021. Of course, it’s possible AWS and other divisions are compensating for a decline in physical goods sold. Unfortunately, I don’t have time to find revenue numbers for specific divisions.
That's from inflation. Look at CVS revenue growth and that's with a declining customer volume. Fewer customers buying fewer items, but fairly massive revenue growth.
They've also laid off a lot of office workers, who were not replaced with robots... so we can't correlate the robot number and all of the changes in head count.
It's clear that technological development creates a shift in jobs, i.e some jobs are lost, but new ones are created as a result. Whether the total #jobs increases or decreases is debatable.
The issue that I see addressed less often is that the new jobs require by definition, a higher skill set. (You wouldn't displace 100 manual labor workers with a machine that requires 105 workers to maintain). So by definition, the average intelligence requirement for jobs increases over time (though never stated directly). This means that as time and technology progress, a growing percentage of people will have no jobs that they are capable of doing. [o]
What's the proper social response to that, I don't know.
[o] If and when AGI comes along, that will be all of us.
The chain of logic is falsified by the Whitney cotton gin: it was a labour saving device, which saved enough labour to make cotton much more profitable, which led to the growth of the cotton plantations in southern USA, which led to increased slavery, and those slavers actively prevented their slaves from learning to read.
That said, I would also agree with the conclusion that "a growing percentage of people will have no jobs that they are capable of doing", but for different reasons.
I expect the abilities of AI to expand over time.
IQ is a poor measure, but suitable as a shorthand especially for a comment like this.
Imagine a general purpose AI that runs as fast as a human on 100 watt hardware; first one will be an idiot. Let's say IQ 50: only 0.1% of humans are dumber than this, nobody was employing them anyway. Version 2, say IQ 85: now about 16% are beaten by the AI, this absolutely matters, they're unemployable forever through no fault of their own, give them a basic income of some kind. Version 3, IQ 100, now half the world can't get work. Version 4, IQ 115, now it's 84% who can't get work, etc.
Reality is a lot messier than that, so nobody needs to bother picking holes in the specific details such as "that's a lot of electricity" or "AI isn't a robot" or "comparative advantage": this is a comment, not a research paper.
The assumption here is that an AI with IQ 100 could do anything a human with an IQ of 100 could do, only cheaper. But that's just averages. Really it would do half the things better and half worse, and then people have jobs doing the things that it does worse.
That would continue until it doesn't do anything worse, which may or may not ever happen, but if it did and we're all still alive then the result would be post-scarcity and nobody would need a job.
You don't have to create anything. "There are no bad products, only bad prices." The demand for labor approaches infinity as its cost goes down.
Suppose that robots and AIs can't grow and distribute food, build housing, provide healthcare or produce and maintain robots. Then there will be necessary jobs for people.
Suppose that they can. Then all of those things will be near-free because they can be mass produced with no labor cost and you'll only have to make a trivial amount of money to have food, shelter and medicine. The low cost of living causes a living wage to be so easy to achieve that even extremely low value work pays more than that, creating jobs for anyone who wants one because so much work is viable at that price.
> Suppose that they can. Then all of those things will be near-free because they can be mass produced with no labor cost
I take issue with this statement. The people who have the robots will be able to set whatever price they like. They could make things super cheap but that is not the inclination of the person in power.
As we've seen in so many "collusion" cases - some deliberately negotiated but many simply naturally settled into - the price of things is determined by "what the market will bear", which I see more cynically as "what the market can extract".
> The people who have the robots will be able to set whatever price they like.
If no one has a monopoly on robots, they will not, because customers will take the lowest price and all of the competing suppliers with robots won't want to lose market share.
If someone has a monopoly on robots, that is very bad, because then they can charge up to the cost of doing the work another way. But the other way is by having humans do it, which is the status quo. So they can't make it worse than the status quo or people would just go back to doing it that way -- the point of buying things made by robots was supposed to be that it would cost less, so why would you do it if it didn't?
> As we've seen in so many "collusion" cases - some deliberately negotiated but many simply naturally settled into - the price of things is determined by "what the market will bear", which I see more cynically as "what the market can extract".
Collusion is caused by market concentration. When an industry has only two or three companies, it's easy for them to collude with each other. When there are dozens or thousands, it isn't. So the most important thing in the world is to make sure there are at least dozens of companies in every industry.
The other major problem of the same nature is regulatory capture. This is the method used by landlords. They don't have a secret chamber where they agree to fix prices in dark of night, they just control the zoning board and prevent new housing from being built.
These things have nothing to do with automation. If you have market concentration and regulatory capture, they must be annihilated regardless of any robots or people will end up screwed and poor.
Thanks for your insightful comment: "If you have market concentration and regulatory capture, they must be annihilated regardless of any robots or people will end up screwed and poor."
The power structures will never allow it. The conditions for a society of leisure have theoretically existed for some time now. We will simply end up with a a planetary ruling class that lives opulently while the other 99% live in abject poverty.
I think your view of our democratic institutions is a bit too rosy. I'm pessimistic that they'd withstand the social upheaval that might occur with a smart-enough AGI. Even now it seems like many people prefer authoritarian rulers -- or at least they think they'd prefer that, as long as the ruler is a part of their political tribe. They'll be in for a painful surprise later, of course.
And that’s not even factoring in the automation of highly targeted yet dynamic political content (not just ads but the consumed content itself) in order to charge/persuade the target to vote for the paying party
"The power structures" want people to need to work so they have to work for them. Better to keep you occupied with the rat race than have you spending time advocating for political reform.
Which is why work expands to fill all available time. They want you to have a job, because what they don't want is what you might do if your time was your own. For some subset of "they" that represents the most malicious pricks.
The thing that happens if they win is that everybody still has a job even if they're not doing anything useful. Which in a lot of ways is what's happening already.
You can live a life of leisure, as envisioned say 100 years ago right now* if you want.
Of course you'd need to be ok with 1920's level of housing (tiny farm house, no sewer or running water), education (stop at 8th grade) and healthcare (no effective antibiotics or cancer drugs).
You could replicate that sort of life today with a plot of cheap land ($3,000 per acre) and $5,000 per year (save $100,000 USD).
Frugality opens up a lot of lifestyle options. Although even given the $7000 a year, this person seems to have a lot of capital in a sense of goods, education, health, and relationships.
That said, there are a few nuances here.
Self-education is now cheap through the Internet (e.g. watching YouTube videos on how to do math or how to fix things). However compulsory education laws make that problematical for children (although there are homeschooling regulations in the USA that can be navigated, homeschooling is illegal in some parts of Europe). The credentialing arms race also means a lot of corporate-type or professional jobs are closed to people who skip college -- even as there are still other opportunities including subsistence production or small-scale entrepreneurship.
Many common antibiotics are cheap due to mass production. But antibiotics can have complicating side effects from wiping out healthy gut microbiota -- which may perhaps include cancer and depression because part of the immune system and part of neurotransmitter production are involved with the gut. Some other cures for things based around herbs are increasingly forgotten and also require access to a large enough area to roam in to find the herbs.
Most (not all) cancer can be avoided through a whole foods diet and active lifestyle. Such a diet and lifestyle is generally cheaper than the mainstream (especially if you have your own garden). See Dr. Joel Fuhrman's G-BOMBS approach as one example:
https://www.drfuhrman.com/blog/237/g-bombs-the-anti-cancer-f...
But, since it is not all, some people in such communities will, as you suggest, die of things treatable in the mainstream. Also, since the people joining such communities presumably have already been eating ultraprocessed food for most of their lives, they are at higher cancer risk than if they had grown up that way (same with other chronic health risks like diabetes, obesity, heart disease, immune dysfunction, and more that are mainly byproducts of Western lifestyle and Western medicine). So, there are these additional background risks for someone going from the mainstream to the alternative -- but the alternatives may struggle under the weight of treating an influx of chronically sick people. (Of course, many mainstream cancer treatments only prolong life for at best a few months at a great cost in suffering and money, but that is a different issue.)
For good or bad, zoning regulations and pressure from neighbors restricts much of what people can do with their land. And cheap land tends to have issues (biting insects, lack of good water, distance from jobs, distance from markets, poor soil, swampy flooding, fire dangers, distance from other people, and so on). So such "cheap" land may actually be quite expensive in health costs and travel costs and labor costs and accepting various increased risks. The book "Life After the City" explores some of these issues.
Decades ago I read an article (in Westchester Magazine?) on someone who learned "primitive" skills for living in the woods. When asked why he did not go and live by himself in the woods using his skills, he replied that it takes a village to live well in the wilderness. So, an overall issue here is that if you want to live well cheaply (which includes some healthy social interactions for most people), you ideally need to be part of a community with related values.
But communities can have problematical dynamics -- especially when surrounded by another culture that is wealthier and more exciting in various ways. That is why, say, the Pilgrims left the Netherlands. The Pilgrims were tolerated in Holland after the left England where there were discriminated against. But they saw their children and other community members starting to adopt Dutch ways, and the older members also encountered other issues fitting into Dutch society. So some of them decided to go to what was then the remote wilderness in North America -- where they could enforce their restrictive norms on their children and neighbors without being surrounded by enticing "Supernormal Stimuli" and "Pleasure Trap" alternatives (both quoted items being names of books on those topics).
https://dutchreview.com/culture/history/the-pilgrims-in-leid...
Of course, many of the Pilgrims died, and the ones that did not survived in part due to the compassion of the Native Americans (as well as plundered Native graves and so on). And their original collectivist vision of land ownership fell apart in the face of massive hard work and starvation and freeloading in that context. But that is another story.
"Who Were The Pilgrims? This Is The Story You Didn’t Learn In School"
https://allthatsinteresting.com/pilgrims
So, if you really want to be as happy as possible with such an alternative lifestyle, a big challenge is finding a lot of other people who want to live like in the 1920s, and actually willing to do that full-time, and all go to the same place, and somehow can afford to buy the land the community needs. And that is all a big challenge, especially since many (not all) alternative communities tend to fall apart over issues of equity or exploitative social/sexual relationships and so on. Some examples:
https://en.wikipedia.org/wiki/List_of_American_utopian_commu...
While not identical in beliefs, some people have quipped the Pilgrims were essentially the Taliban of their day. And in a way, it makes sense, because to leave behind your home as a community takes some common set of core beliefs and strong social bonds which are often associated with extreme religious sects.
To an extent, the Amish are somewhat like this as far as being tight knit religious-based communities which are apart from mainstream US society. But they still emphasize hard work and related material affluence -- and also happily selling goods and services to the lazy "English" all around them. So the Amish are not quite a community of leisure -- even if many people may find happiness in that life.
So, given that, the Amish are far from what Marshal Sahlins describes in "The Original Affluent Society" of hunter/gatherers who have lots of leisure time since most of the food they need does not take that long to acquire and is done often in a way people think is fun or at least engaging.
https://en.wikipedia.org/wiki/Original_affluent_society
Unfortunately, hunter/gatherers are living in a productive landscape were generally displaced by militaristic bureaucracies wielding mass-produced "Guns, Germs, and Steel" (like how the first century of the US Army was mainly about being used to displace Native populations):
https://en.wikipedia.org/wiki/Guns,_Germs,_and_Steel
Which circles back around to why cheap land to live frugally in community is generally not desirable land.
I do think modern technological advances do make new alternatives possible. "OSCOMAK" is a project I came up with decades ago to help support communities of any size to develop whatever infrastructure they desire (but admittedly it is still more an idea than a realization):
https://www.kurtz-fernhout.com/oscomak/
"The Oscomak project is an attempt to create a core of communities more in control of their technological destiny and its social implications. No single design for a community or technology will please everyone, or even many people. Nor would a single design be likely to survive. So this project endeavors to gather information and to develop tools and processes that all fit together conceptually like Tinkertoys or Legos. The result will be a library of possibilities that individuals in a community can use to achieve any degree of self-sufficiency and self-replication within any size community, from one person to a billion people. Within every community people will interact with these possibilities by using them and extending them to design a community economy and physical layout that suits their needs and ideas."
And like with the Pilgrim/Taliban, :-) I can imagine such tools most useful for a dedicated community trying to "live off the land" in the desert, ocean, Antarctica, or outer space (Moon, Mars, Asteroids, etc.). Those are all "cheap land" (with no or low taxes) in the sense of being generally far away and generally not pleasant places right now due to lack of one thing or another --- including unfortunately other people to form a community with.
See also the book "Retrotopia" by John Michael Greer for more ideas, including how to have "zones" of infrastructure and taxation at different levels that people can choose from.
An alternative to increase leisure though extreme frugality and living apart from society is to upgrade the mainstream society we have, such as with a basic income; improved subsistence with gardening robots, 3d printers, and solar panels; better collaborative decisions making in democratic government; and a stronger gift economy (like sharing information essentially for free via the web like HackerNews for example makes possible).
I'm one of the weirdos who can happily live on around €10k/year.
But I'm well aware that I can only afford to live so cheaply because the economies of scale of everyone else spending so much. My laptop could only be built because a million other people would buy one too; the factory could only be built because it also serviced several other computer manufacturers; the mines for the raw materials can only stay open because their stuff isn't only used for computers.
If everyone was like me, the roads wouldn't get paved.
> Self-education is now cheap through the Internet (e.g. watching YouTube videos on how to do math or how to fix things). However compulsory education laws make that problematical for children (although there are homeschooling regulations in the USA that can be navigated, homeschooling is illegal in some parts of Europe). The credentialing arms race also means a lot of corporate-type or professional jobs are closed to people who skip college -- even as there are still other opportunities including subsistence production or small-scale entrepreneurship.
I've just had a very visceral demonstration of how bad this free education is.
I reached a 2500 day streak in Duolingo after repeatedly completing the entire skill tree and watching the gold disappear as the tree was grown, have also been using two other apps daily and intensively for years, have been watching and listening to free podcasts and YouTube videos about the German language most days… and after 5.5 years living in Berlin still only managed to score an A2 on an official language test.
Great point on what would happen to mass-produced items like laptops if everyone moved to remote cabins in the woods to live cheaply.
And yes, a drawback to educating yourself via videos or reading (or even by taking some tests) is that it is relatively passive. A hands-on aspect is essential to a lot of skills, like learning to write by writing, learning to use machine tools by machining, learning to do chemistry by mixing chemicals, and learning to use a language by speaking it in important contexts. And setting up such contexts is one value of good educational settings (either in or out of schools).
As a parent who home-schooled from K-12, definitely the lack of easy access to some settings where various skills could be practiced was a limitation. We tried to make up for it as best we could in other ways.
Some learning can happen virtually to an extent, like virtual chemistry sets (even as they don't convey the visceral feel of working with glassware and liquids and such).
There is also an project at RPI for people to learn Chinese in a virtual Chinese setting with characters who speak Chinese:
https://research.rpi.edu/about/signature-research-thrusts/me...
"Students learn to speak Chinese through the Mandarin Project’s virtual reality, synthetic intelligent characters and gamification."
It is also harder to learn a language when the people you are talking to want to learn or practice yours and so are motivated to speak your language to you. So as a consolation prize, maybe you have helped some Berliners become better English speakers? :-)
Age may also have to do something with language learning -- but not as people expect. When people are young, they are often less embarrassed by making mistakes, and are full of enthusiasm for something new, and they way they are taught is usually point and name and speak and question. Adults tend to try to avoid mistakes, may have other more interesting or important things to do, and also may end up in boring textual drill-and-practice rote learning situations. Those are some of the reasons it is harder to learn languages (especially a first foreign language) as an adult.
A related funny XKCD on simulation and learning (where Kerbal Space Program makes a seemingly difficult learning challenge fun when done in private and also has no significant consequences for mistakes so you can play around with ideas like a kid again):
"Orbital Mechanics"
https://xkcd.com/1356/
Some people are just also better than others at languages for whatever reasons.
Also, sometimes immersion in a context for a length of time just makes the difference. One of my German teachers in High School talked about studying abroad in Germany when younger and speaking only German where they were, and they said they felt a real turning point was when they had their first dream in German. I'm guessing maybe you must be in a situation where you can still speak English or whatever your native language is at work and at home, and so speaking German just is not that important to your daily life?
Anyway, wishing you some dreams "auf Deutsch" if you want them. :-)
> The conditions for a society of leisure have theoretically existed for some time now.
Not if everyone wants to maintain their current standard of living. Imagine working 50% less hours, but also having to live off 50% of your current total compensation. Some here could probably manage that, but most would probably prefer not to.
Now if you want a whole "society of leisure", you'd have to impose that same lifestyle choice across all of society by government fiat. Its easy to see why that hasn't happened.
Not really a realistic view of the western economy. Or economics.
Things like 'compensation' are fluid, fairly arbitrary and largely unrelated to the industrial complex. It's whatever we decide it is, pretty much.
Automation is in full swing, has been for twenty years and is only accelerating. Ignore that at your peril. Everybody will continue to have toasters, computers, cars even when we've automated most of us out of an industrial/manufacturing job.
How do I know? Because that already happened. Instead of tens of thousands of people on assembly lines, we have tens of engineers and managers overseeing automation. If it hasn't happened in some cherry-picked example, it will very soon.
We have to plan something for the majority of us to do, some way to participate in the resulting economy, without just throwing up our hands and saying "It's too hard!"
So much to say on this subject, that doesn't fit in an HN text field. There's a long history of thought on this subject, and the comments here indicate most folks are still on the first page in their thinking.
Perhaps you could link to some information on this "long history of thought on this subject" because to me everything you just wrote sounds like nonsense.
The idea that we can just double people's compensation and thereby double our total economic output (because compensation is "whatever we decide it is") is so wrong headed I don't even know where to begin. Maybe that's a misunderstanding of your position, but I don't know how else to interpret what you just said.
Right now we have a lot of jobs that are extremely low efficiency, if not outright net negatives. These are the sorts of jobs that markets tend to eliminate, because the company that gets rid of them can charge lower prices than the one that doesn't. But that doesn't work when the job exists as a result of regulatory rules, or the company is a monopoly not subject to competitive pressure.
What does work is to improve the efficiency of the regulations. For the ones that are net negative you can just get rid of them. For the ones that are net positive but still poorly constructed, you can reduce their overhead.
This doesn't increase economic output, it reduces waste. Then you can work half as many hours for the same money, or the same number of hours for twice as much, not because anything more is being created but because people are spending half as much time on useless tasks. If they then spend that time doing something productive, output would increase, but that's a personal choice. If you could work 10 hours a week and that was enough to earn a living and own a home, would everybody still want to work 40 hours just so they could also own six cars and two boats? Some people would, which is fine, but some people wouldn't.
Sure, if you can somehow find a way to double the efficiency of labor then you can either double your standard of living or half your number of hours. We've already done that multiple times since the industrial revolution. That's why modern first world society is so rich compared to the past.
And there are certainly significant inefficiencies created due to government regulation, though probably not enough to double our productivity even if you did have the knowledge and political capitol fix all those issues perfectly. It is also worth noting that sometimes regulations, though costly and inefficient, can still be nice things to have. Building codes undoubtedly make housing significantly more expensive, but I'd still rather live in a society with expensive housing where I don't have to worry about the floor collapsing on me than a society with cheap housing where I do. There's a balance there obviously, but my point is sometimes the extra expense can be worth it even if its not "net positive" in a purely economical sense.
> If you could work 10 hours a week and that was enough to earn a living and own a home, would everybody still want to work 40 hours just so they could also own six cars and two boats?
I think you'd be surprised. There are so many things we consider necessities now that would be considered luxuries 100 years ago. I see no reason why things won't continue to move mostly in that direction as technology improves.
I know a lot of people who earn enough that they could match my standard of living working only 10 hours a week. They mostly don't, and instead spend the extra wealth on things like larger houses, fancier cars, exotic vacations, etc.
> And there are certainly significant inefficiencies created due to government regulation, though probably not enough to double our productivity even if you did have the knowledge and political capitol fix all those issues perfectly.
Keep in mind that a lot of these rules are multiplicative. Zoning rules limit the amount of new housing construction and increase construction costs because now you have to e.g. replace a 10 story building with a 20 story building, bulldozing the 10 story building, instead of replacing a single family home with a 10 story building to add the same number of units. Professional licensing apprenticeship requirements limit the supply of licensed tradesmen, increasing construction costs. These multiply together: You have to do more construction and the construction has a higher labor cost.
Then housing costs more, so you have to pay higher salaries for the same cost of living -- including to tradesmen, which makes construction cost even more, multiplying the effect again. But not just tradesmen, also the salaries of compliance bureaucrats needed by any other form of regulation, and the cost of commercial real estate for their offices.
Double is, if anything, an underestimate. These costs are quadratic.
> Building codes undoubtedly make housing significantly more expensive, but I'd still rather live in a society with expensive housing where I don't have to worry about the floor collapsing on me than a society with cheap housing where I do.
The building codes from decades ago were sufficient to prevent buildings from collapsing. Since then they've been accumulating cruft. Many of these individual requirements each add hundreds to thousands of dollars to the cost of a new house in exchange for a marginal safety improvement with a negative expected value.
And then you don't even get the safety improvement, because making new construction prohibitively expensive causes people to continue to live in old houses that weren't subject to the new requirements anyway. All you do is make housing more scarce.
> There's a balance there obviously, but my point is sometimes the extra expense can be worth it even if its not "net positive" in a purely economical sense.
"Net positive" is the measure of if it's worth it. You have a measure that can prevent a 1 in 1000 chance of $50,000 in damage but it costs $1000. You're spending an average of $1,000,000 to prevent $50,000 in damage. It's not worth it.
> There are so many things we consider necessities now that would be considered luxuries 100 years ago. I see no reason why things won't continue to move mostly in that direction as technology improves.
Most of these things are things that didn't previously exist, like cellphones or computers. Now you need one because it has replaced certain ways of interacting with people and institutions and the old ways are no longer available.
But let's suppose the definition of necessities expands over time. It used to be food and shelter, then we added medicine and transportation, then we added internet. Maybe tomorrow we add a personal robot or something else. But what if all of those things together cost $10,000/year? Would everyone choose to work full time if the surplus was solely to purchase luxury goods?
> I know a lot of people who earn enough that they could match my standard of living working only 10 hours a week. They mostly don't, and instead spend the extra wealth on things like larger houses, fancier cars, exotic vacations, etc.
There are also people who sell their startup and then choose to retire in their 20s or 30s. Different people make different choices.
It seems to me like we're mostly in agreement. I agree regulations can have compounding negative effects, though whether that's enough to account for a potential 2x improvement in overall economic output is something I would still dispute without any hard numbers to back up that claim.
I also think you've misunderstood my point about regulations. What I'm saying is that sometimes regulations can serve as a sort of "luxury good", where even if they're a net negative economically (like in your example of the uneconomical safety improvements) the quality of life benefits can still make them worth the cost in a society that's wealthy enough to absorb those costs.
I'm not saying that's always the case. Sometimes the costs clearly aren't worth it even factoring in subjective benefits. Just that its another thing to consider.
I also agree that there's probably a certain point where people would start to value their time more than what we today would consider "luxury goods". Just that that point is probably a lot further out than you might expect, because what's considered a luxury is relative to the cultural standards of the society you live in. Some people do chose to retire in their 30s, but I know a lot more people who didn't do that even though they probably could have.
Everybody loves to calculate worker productivity, like it's the 1800's and we're hand-making buggy whips. No, its a large industrial machine turning out goods like clockwork.
Paying people is arbitrary. We pay somebody to spend time fixing a conveyor belt, and pay them nearly nothing. Some guy sits in an office doing nothing, but he owns the factory so he gets all the profit from that factory forever, yet has little or nothing to do with it's operation.
It's arbitrary, a direct result of some choices we made about our economic system. Those choices have quit working for most of us - a tiny percent (way less than 1%) have managed to jigger the rules so they skim off of every transaction. The end result of that kind of feedback loop is, they have all the money and we continue to work hourly for peanuts.
We can just decide to change it. Doesn't require magic or even much imagination.
That's a very mid-1800s position to take, given you're complaining about "calculating productivity like it's the 1800s" — owners getting rich from reinvesting the profits of labour is one of the big things that Marx was cross about.
> We can just decide to change it. Doesn't require magic or even much imagination.
Many have tried. What's the quote?
“For every complex problem there is an answer that is clear, simple and wrong.” ― H. L. Mencken
Some attempts have been more successful, some less successful, but the governments which think economics is easy have done the worst, regardless of if they were Communist or Capitalist.
To support your point that we we make choices among alternative, see this document I put together circa 2010 with about 50 alternatives -- some good, some obviously bad -- we can use to address the socio-economic changes needed to maintain healthy human communities in the face of increased automation:
https://pdfernhout.net/beyond-a-jobless-recovery-knol.html
"This article explores the issue of a "Jobless Recovery" mainly from a heterodox economic perspective. It emphasizes the implications of ideas by Marshall Brain and others that improvements in robotics, automation, design, and voluntary social networks are fundamentally changing the structure of the economic landscape. It outlines towards the end four major alternatives to mainstream economic practice (a basic income, a gift economy, stronger local subsistence economies, and resource-based planning). These alternatives could be used in combination to address what, even as far back as 1964, has been described as a breaking "income-through-jobs link". This link between jobs and income is breaking because of the declining value of most paid human labor relative to capital investments in automation and better design. Or, as is now the case, the value of paid human labor like at some newspapers or universities is also declining relative to the output of voluntary social networks such as for digital content production (like represented by this document). It is suggested that we will need to fundamentally reevaluate our economic theories and practices to adjust to these new realities emerging from exponential trends in technology and society."
It covers a bit of the history which you alluded to earlier in a previous post, such as the "Triple Revolution Memorandum" from 1964. But the root of all this thinking go way back -- whether to the original communal Christians in the face of the Roman Empire, various Utopian communities (including ones inspired by Charles Fourier) to Marxism, Socialism, Luddites, resistance to "Enclosure Acts", Henry George, Elizabeth Magie (her educational cautionary Landlord's Game ripped off in part as dystopian Monopoly), Bucky Fuller, Bob Black ("The Abolition of work"), Ursula K. Le Guinn, James P. Hogan, Paolo Soleri / Arcosanti, Marshal Sahlins, Amory & Hunter Lovins, John and Nancy Jack Todd of The New Alchemy Institute, Martin Ford, and many more. Not all the alternative ideas worked out or even got started for all sorts of reasons, but there are a lot of alternatives are out there.
See especially:
"The Dictionary of Alternatives: Utopianism and Organization"
https://www.bloomsbury.com/us/dictionary-of-alternatives-978...
"'There is no alternative to free market liberalism and managerialism', is the orthodoxy of the twenty-first century. All too often, ordinary people across the world are being told that the problem of organization is already solved, or that it is being solved somewhere else, or that it need not concern them because they have no choices. This dictionary provides those who disagree with the evidence. Using hundreds of entries and cross-references, it proves that there are many alternatives to the way that we currently organize ourselves. These alternatives could be expressed as fictional utopias, they could be excavated from the past, or they could be described in terms of the contemporary politics of anti-corporate protest, environmentalism, feminism and localism. Part reference work, part source book, and part polemic, this dictionary provides a rich understanding of the ways in which fiction, history and today's politics provide different ways of thinking about how we can and should organize for the coming century."
Indeed, I was in a way surprised (and yet also hopeful) in reading all the comments here in that many people are slowly rediscovering all these ideas for themselves as the issue grows more pressing.
Essentially, I have been dealing with "AI/Robotics Anxiety" for about forty years, and collected many ideas for coping with it along the way. I'm working in a broader document more specific to that which I may put up on my website at some point.
Ironically, while I won awards for robotics projects before attending Princeton as an undergrad, and my undergrad work at Princeton in AI and cognitive psychology helped inspire WordNet which led to Simpli and the core of Gogole AdSense, it was maybe conversations with Jeff in passing at Princeton about the potential to use robots in commerce and space exploration --- which I doubt he remembers -- that might have had the biggest impact of my career to-date in a robotics sense given Amazon and its emulators. Not to take any credit away from George or Jeff in terms of their specific vision, hard work, improvisations, and persistence in the face of adversity, and also to accept that a place like Princeton can be a Brian-Eno-style "scenius" where ideas bounce around and transform as they are reflected on and refined by different people with different perspectives.
https://austinkleon.com/2017/05/12/scenius/
"There’s a healthier way of thinking about creativity that the musician Brian Eno refers to as “scenius.” Under this model, great ideas are often birthed by a group of creative individuals—artists, curators, thinkers, theorists, and other tastemakers—who make up an “ecology of talent.” If you look back closely at history, many of the people who we think of as lone geniuses were actually part of “a whole scene of people who were supporting each other, looking at each other’s work, copying from each other, stealing ideas, and contributing ideas.” Scenius doesn’t take away from the achievements of those great individuals: it just acknowledges that good work isn’t created in a vacuum, and that creativity is always, in some sense, a collaboration, the result of a mind connected to other minds. What I love about the idea of scenius is that it makes room in the story of creativity for the rest of us: the people who don’t consider ourselves geniuses. Being a valuable part of a scenius is not necessarily about how smart or talented you are, but about what you have to contribute—the ideas you share, the quality of the connections you make, and the conversations you start. If we forget about genius and think more about how we can nurture and contribute to a scenius, we can adjust our own expectations and the expectations of the worlds we want to accept us. We can stop asking what others can do for us, and start asking what we can do for others."
But I am a bit sad about the results so far though, given how robotics, AI, and other automation are turning out to do so much damage in practice to so many people's lives when used from a scarcity-oriented mindset. It was always my intent -- especially having read so much sci-fi like Isaac Asimov robot stories, "The Skills of Xanadu" by Theodore Sturgeon, and also James P. Hogan's "The Two Faces of Tomorrow" and also his "Voyage from Yesteryear" -- that way more positives than negatives would come out of the abundance that robotics, AI, and other automation (including information technology) can provide.
A related video parable I made on such themes circa 2010:
"The Richest Man in the World: A parable about structural unemployment and a basic income"
https://www.youtube.com/watch?v=p14bAe6AzhA
As with my sig (first sent in an email to Marvin Minsky) of "The biggest challenge of the 21st century is the irony of technologies of abundance in the hands of those still thinking in terms of scarcity", I have realized -- as Einstein suggested about nuclear weapons or Lewis Mumford about technology in general or James P. Hogan essentially says through his novels -- it takes change of heart and mindset to realize the benefits of what is possible using automation without otherwise creating a terrible human calamity.
An example of things going wrong is creating problematical working conditions when automation is used to make humans work like micromanaged robots centralizing wealth for a few people. Related US DOL citation for Amazon essentially for not automating enough:
https://www.dol.gov/newsroom/releases/osha/osha20230201-0
As I've reflected on, if you get up in the early morning in the dark and turn a light on in the kitchen, the light may seem so bright you can't look at it even as it even as it makes it possible for you to get your day started. And by the time the sun comes up and lights your entire home and landscape, you may forget the kitchen light is even on because it is no longer noticeable relative to the surrounding broad illumination. Gary Kildall no doubt faced that with CPM. So that is kind of how I see the comments overall on the story. The sun is coming up as millions of people are starting to think about robotics, AI, and other automation and its likely near-term effect on themselves and society, and all the "bright lights" from decades past are just not noticeable from all the ongoing thinking and chatter. And that is maybe a good thing -- even if there are still ideas from then which might be useful now and will either be remembered, reread, or rediscovered.
Why not, though? Advancements in robotics happen all the time. If we end up with a decent AGI in 50 years, I expect the state of robotics to have advanced too, perhaps to the point where it could carry furniture up a stairway to an apartment.
> My bet is that some large fraction of people are currently doing jobs that are far less demanding ("robotic") than what they are capable of doing
Agreed, hence the caveat of 100 watts when at human speed: Humans are not capable of competing against costs of 100W * $0.10/kWh = $0.01/hour even when the only expense that human has is the cheapest available calories.
c/ as far as I can tell, "jobs" have been steadily becoming less skilled over the centuries (because we expect people to switch more frequently between them?); can we accelerate that?
So might a way out be that people do regular 100 IQ people stuff (yoga instructors, community theatre, etc.) and automation does (almost) all the heavy lifting?
Bring on the Centrifugal Bumble-Puppy? If the Morlocks of the future are electronic, they won't want to eat us Eloi anyway? (burning a human for electricity is even more wasteful than giving it spin classes)
TIL there are amazingly many articles online regarding the calorie content of an adult human body; at 125 000 kcal I get USD ~15 per person, making me suspect that they can easily be put to higher NPV use than their carcass value.
(I did not go so far as to search out tanning instructions for making gloves)
No, what they mean is IQ/intelligence, just as they stated. Social class is part of the story, more so in the past than it is now, but it's not the entire story. High intelligence is a way out of upbringing and circumstances at the margins, when you can capture opportunities to escape. Tech jobs are filled with people who didn't come from higher social classes, many who suffered inequities during their childhood even from their social peers, but persevered due to having drive and intelligence.
Their argument is exactly what they stated, no hidden meaning here about social class. Not everything is class warfare.
Wait a minute. Higher skill jobs? For at least a decade, the place jobs shifted to was Amazon. If you lost your job because Amazon put your company out of business, you could get a worse, more humiliating job working at an Amazon warehouse, or as an Amazon driver. It was not higher skill jobs the market shifted to, it was Nomadland jobs. Now that Amazon is switching to robots, where are those people Amazon is putting out of work supposed to go?
> For at least a decade, the place jobs shifted to was Amazon.
It wasn't so much the jobs that shifted there as the customers, and the reason for that is that Amazon had better prices. Then people complained about the jobs they did offer because they're mechanistic and exhausting, so they automate them and then people complain about that. But that too should result in lower prices -- Amazon's retail operation doesn't make any money, it's all going to competition with Walmart, who is doing the same things to lower costs.
But lowering prices creates jobs. People pay less for a dress or a phone case and then spend the money on something else. New jobs are created doing the something else.
Where this becomes a problem is for the things where the prices don't come down, like real estate. You have extra money and now you want to buy a house, but zoning laws inhibit new housing from being built, so instead if people have any money the monthly payment you need goes up or the landlord increases your rent. Or you have to buy something from some monopolist who can raise prices to eat your disposable income. Then the money goes into some corporate holding company that just keeps growing their hoard and never spends it on products and services.
The problem isn't robots, it's certificate of need laws and high tuition.
> But lowering prices creates jobs. People pay less for a dress or a phone case and then spend the money on something else. New jobs are created doing the something else.
Isn't this the principle supposedly behind trickle down economics just by another name?
There are a lot of economic theories that are correct under a particular set of constraints and completely invalid under a different set of constraints.
So for example, if you lower property taxes, theory says that rents should go down. At the existing rents it's now more profitable to own property because the ownership costs have gone down, so more buildings will be constructed as people take advantage of the opportunity, until the increase in supply brings rents down to account for the reduction in operating costs. This theory is invalid if you prohibit new construction because then supply can't increase so rents never go down and the existing landlords just pocket the money.
The problem isn't that theory doesn't work, it's that you have to satisfy its constraints. You have to get rid of monopolies and artificial scarcity or they capture the gains that were supposed to go to ordinary people.
The fact that you can now warehouse with robots lowers the barrier of entry for warehousing, which creates new companies that will host competition against Amazon, offer contract labor for Amazon, and in both ways create jobs that are superior to the previously needed entry level pick and packer. The same laborer can now be promoted at a new company to a higher pay job that requires no greater skill set, simply because there are now more of those better jobs at more companies. We can't all be managers at Amazon. But we can all be managers at 100000 different warehouses that previously didn't exist. That's where they are supposed to go.
That seems wrong. Why would robot warehouses lead to more and better warehouse jobs for humans? Fewer humans would be necessary overall in the existing warehouses, and it's not obvious to me why a lot more warehouses would be created.
Did replacing manufacturing jobs with robot assemblers in—for example—the automotive sector, lead to more auto manufacturers and better jobs for auto workers? I don't believe it did. There may be more manufacturers now, and more jobs, but they aren't high skill or highly-paid jobs, and they aren't staffed by the people who were laid off originally (because they were mostly moved to other countries, where people can work more cheaply than the countries where the jobs were lost).
For capital intense upgrades like robots, why wouldn't the advantage go to a few big players, rather than a ton of small ones?
I also don't understand why the number of new skilled workers in this new world would somehow equal the number of warehouses workers laid off. What's the connection between those two seemingly unrelated phenomena? Why wouldn't it, for example, be a lot of laid off low-skill workers, and a just a few new, high-skilled workers?
Or for that matter, why the next generation of robots wouldn't just replace those higher-skilled warehouse jobs in a few years. And so on.
You aren't going to trust moving 100 million in product strictly to robots for a long time. And that's just a small business. And there's a lot of them.
>why wouldn't the advantage go to a few big players, rather than a ton of small ones?
Because it's still way way cheaper than labor. Maintenance means you're paying for one guy's medical benefits instead of 20 guys, for example. The labor is the cost that's difficult to overcome and gives the bigger players an advantage. When that's stripped away, it's possible to compete against the bigger guys.
Cheaper than existing labor costs, loans can overcome capital entrance, and you can afford to pay on them when there's a smaller operating cost, plus they have some fixed ROI. Existing need for more warehouses, combined, I don't see why we wouldn't see more warehouses. Of course we will.
Look at middle America. Almost every metropolitan in the country is building warehouses in and around their airports. Some indeed are Amazon's and other big suppliers, but the majority are not. They are small storage and shipping outfits. Don't forget who supplies Amazon!
Also, the push for more condensed housing means fewer people per household which means more duplicate junk per person as they won't share with another household, obviously.
Growth means warehousing. There's no way around it. Unless we will manufacture domestically, that demand isn't going anywhere.
The moat is definitely smaller when labor, your biggest cost, is smaller. You can finance a purchase that's much larger if you are able to make the monthly payment because you don't have high labor costs. Purchasing has an ROI, and labor doesn't.
The robots replaced the frontline laborers, yes. But someone now needs to maintain, build, and engineer new robots. Hence they said the laborer's job was replaced with one that requires fewer people at a higher skill (and probably pay) level. Or at least that's what I picked out from their comment.
Please note that while the robot is replacing local jobs, the people engineering, building, and maintaining robots could be ln the other side of the plant also. In a world that's seen a huge amount of consolidation we are likely to see further consolidation.
My understanding was that they're saying it's the act of replacing the worker with a robot that requires more skilled workers to maintain/design/manufacture/etc., not that it merely resulted in a layoff.
> It's clear that technological development creates a shift in jobs, i.e some jobs are lost, but new ones are created as a result. Whether the total #jobs increases or decreases is debatable.
It’s not debatable; we have over 200 years of technological development to look back on and the trend is clear: the total number of jobs has increased at least as fast as human population growth over that time.
In addition, the nations driving the most technological growth domestically have experienced the greatest job growth over that time. With the result that many of them, like the U.S. and UK, have had to develop robust immigration programs.
Even within a single nation, like China, there is temporal correlation between technological development and job creation. As China has leaned into tech over the past few decades, job creation accelerated there.
> So by definition, the average intelligence requirement for jobs increases over time (though never stated directly). This means that as time and technology progress, a growing percentage of people will have no jobs that they are capable of doing.
Again, the evidence shows the opposite correlation: technological development results in more people working, not less.
> It’s not debatable; we have over 200 years of technological development to look back on and the trend is clear: the total number of jobs has increased at least as fast as human population growth over that time.
200 years ago the town [severely mentally disabled person] could chop wood and carry water. What's he doing today?
There are ~9 million people on SSDI (disability) and ~5 million on SSI (considered completely unfit for work, the US version of basic income), and ~50 million retired. Retirement conceptually slowly became a thing around the late late 1800s. Many of these people are in one of these three categories because there is no job that would be a good fit for them, especially SSI, which most Americans don't even know about.
> It’s not debatable; we have over 200 years of technological development to look back on and the trend is clear: the total number of jobs has increased at least as fast as human population growth over that time.
That isn't some immutable law of the universe. 200 years is a short sample size relative to geologic time.
Once we have robots doing the cooking, cleaning, heavy work, etc., what becomes of the Waffle House and Walmart worker? There will be a lower bound capability threshold, and automation will eventually exceed that.
I think a smart comparison would be to look at what job opportunities are available to the intellectually disadvantaged.
Then what happens when that lower bound inches higher?
Ya, when people use this argument ask them "the population of humans always grows right?" Because up till recently that was the consensus unless something drastic or terrible happened. Then in the past few decades we see people having far fewer children then even replacement rate.
Upsetting the labor market is leading us into unpredictable territory, much like at the start of the 1900s and the automobile set off a string of events that lead to two massive world wide wars.
In principle, they could probably just buy their own robots and start their own businesses. Locality is its own quality for SMEs. Whether or not that happens in practice is anyone's guess.
The key point being it's created more jobs so far but we cannot extrapolate the same thing if AGI comes up tomorrow. Like let's say open ai comes up with a new LLM that is capable of replacing a human in let's say software development. What new jobs would it create?
All technological advancement so far has created new jobs because you need someone to actually work on it, like a chip factory or doing devops. As far as I can see an AI is general enough that you don't need much effort to specialize it and with how things are currently going, only a few players have the capability of building and deploying it.
I'm sorry, why would an AGI be interested in programming? My kids are AGI, and they're not interested. I think there's a real moral conundrum when we say "programmer AGI" because, I think, we're implicitly talking about terminating every non-programmer AGI, to meet our labor force whims. Replace "programmer" with intellectual task of your choice.
Presumably the interests of an AGI could be designed into it. It could be made to have programming be the center of its existence.
Even if it couldn't be designed in, then if we accept your analogy with people, we could simply generate AGIs until we found one that was interested in programming, then clone as many copies of it as we needed.
Poor countries have with little physical or financial infrastructure have high unemployment. You'd think there would be more jobs because there is a lot more opportunity to grow, but no, it's the opposite, there few jobs and they are bad jobs. Because there is little opportunity to create actual VALUE, in economic sense.
Technology brings efficiency and brings jobs.
Say entire tech sector, software developers and IT get fully automated - well, now all the VALUABLE services those companies provide are much much cheaper. All the savings are passed on to their customers (B2b and B2c) who will now spend those savings doing things they couldn't afford to before - and THOSE industries are where jobs will move to.
For a more simplistic example, imagine cost of electricity (or some raw materials) dropped 10x, would it lead to fewer jobs or more jobs? Of course more jobs, since you'll be able to do a lot more now.
While I understand your point it seems to only focus on one side of things, a bit like trickle down economy.
Let's say the IT sector is completely automated. What would all those devs do? Now keep automating medicine, legal and everything else and ask what would those people do? What's remaining are probably manual labor jobs for which we don't need so many people.
As with any productivity boost - if you need 100x fewer IT jobs, it doesn't mean you will cut your IT by 100x, you will instead increase just grow your IT department 100x or however much until you hit a growth bottleneck on the business side, then some jobs will shift there.
Just like my raw materials/energy cost example before, if you something becomes cheap - you don't consume less of it, you try to do more with it. Maybe, so much more that you can now do things which were completely impossible/unaffordable before!
It's very possible we will have MORE IT jobs due to new opportunities and efficiencies.
A decent software developer costs $40/hr. Let's say you can make your small iPhone app idea in 100 hours, that's $4,000. If developer with 100x as productive, now that costs only $40. That doesn't only mean more people can afford to build their ideas, but you could build BIGGER ideas, on a $4,000 budget you could now build $400,000 worth of an application! Think of all the opportunities for everyone!
> The issue that I see addressed less often is that the new jobs require by definition, a higher skill set. (You wouldn't displace 100 manual labor workers with a machine that requires 105 workers to maintain). So by definition, the average intelligence requirement for jobs increases over time (though never stated directly). This means that as time and technology progress, a growing percentage of people will have no jobs that they are capable of doing.
You might look up what economics has to say; this issue is well-addressed there. Some fundamentals:
The comment above assumes a static marketplace - the same technology, needs, etc. - and one that addresses the entirety of economic demand, rather than a dynamic market where those things change and resources are scarce (thus when resources become available, they are applied to other unfilled needs).
For example, which skills are in demand changes but there is still growing demand: If you look at the jobs performed 100 years ago, you'll see that most of them are no long needed. Yet not only are most people employed today, we have ~3-4x as many people - most of the jobs disappeared, yet, 3-4x people have jobs.
And yes, a growing economy requires higher-skilled work, but that's good because that work comes with higher pay.
>yes, a growing economy requires higher-skilled work, but that's good because that work comes with higher pay.
The problem, of course, is that wages stagnated from '99 to 2014 and the job participation rate has been decreasing since 2000 while cost of living and general production increased; so no, in a dynamic market work does not necessarily come with higher pay. It actually wouldn't really make sense for all new work to come with higher pay; if you have changes in supply (which is what we are really talking about) that come with associated lower labor costs, the people that used to provide higher cost labor for the initial supply level will have to accept significantly reduced salaries in their industry.
That's what happened with the industrial revolution. Wages overall increased because people entered the workforce for the first time as skill requirements went down, but the average wage of previously employed people went way down as artisan and highly skilled labor was outcompeted by factory work.
> The issue that I see addressed less often is that the new jobs require by definition, a higher skill set.
I don't think that's necessarily true.
A taxi driver of 20 years ago had to memorize roads and routes, and become familiar with traffic patterns at different times of the day. (Many taxi drivers sucked for various reasons, but that's neither here nor there.) Today, anyone can be an adequate taxi driver, without really any knowledge of the areas they work in. Slap a smartphone on your dashboard and you're good to go. That's a lower skill job.
> You wouldn't displace 100 manual labor workers with a machine that requires 105 workers to maintain
Sure, and that's part of the problem. You lay off those 100 manual workers and hire 5 workers to maintain the machinery. It's not at all clear that those 100 manual workers are going to then find higher-skill jobs, especially if they need training to get them.
> What's the proper social response to [hypothetical AGI displacing more and more jobs], I don't know.
Easy: universal basic income, plus free higher education and vocational training programs for people who do want to work in a field that still needs humans in the loop. We need to drop this inhumane view that people who can't (or even simply don't want to) work somehow don't at least deserve a basic standard of living. And I think we'll find that if we remove financial, housing, and food insecurity, more people will actually want to find meaningful ways to participate in society. Some people won't, and maybe that's a shame, but for the sake of all of us, they should still be housed and fed, comfortably enough.
Jobs are a byproduct of capital owners needing labor to sell things and services with a profit margin, so they can buy the good stuff. I’d argue that once the capital owners possess the technology that brings the cost of labor near zero, they will have no need for an economy at all, not to mention other people - unless it’s something more akin to a zoo.
A quick google says that 53% of Americans own publicly traded stock, but that means that 47% do not. There will be some fraction of that 47% who have their own businesses, but I’m guessing that these would be people at the bottom end of the business scale. Would it really be fair to call the guys driving a truck through the alleys collecting crap metal capital owners who would live in a utopia?
The same article that cites the 53% number also says that the top 10% of income earners own 70% of the stock market. That doesn’t really sound like a recipe for utopia.
> 34:24 "Markets are efficient because of active managers setting the prices of securities, firms like Citadel, firms like Fidel.....lity (Fidelity) [...] trying to drive the value of companies towards where we think they should be valued." - Kenneth Cordelle Griffin, Citadel Securities, November 2023
I thought Amazon was a public company, not private, and since supply and demand is dead and real price discovery is gone, that unless it becomes a private company it likely will not survive, unless there are enough retail investors to DRS directly register shares in their name so that DTCC (Depository Trust and Clearing Corporation) and Cede & Co do not hold all the assets for retail in street name to then bankrupt and collect all the assets leaving retail investors holding worthless securities, but I don't see that Amazon is private? Maybe I missed something that happened and it's a private company now?
> thought Amazon was a public company, not private
Parks, libraries and military bases are publicly owned. Amazon is a private company whose shares are publicly traded.
> since supply and demand is dead
This is wrong.
> so that DTCC (Depository Trust and Clearing Corporation) and Cede & Co do not hold all the assets for retail in street name to then bankrupt and collect all the assets
This is nonsense. (For starters, not how bankruptcy works.)
> Maybe I missed something
Mankiw’s Principles of Economics is a well-regarded introductory text [1].
And for most Americans who own publicly traded stock, the income that is derived from it is supplemental, not primary.
You're not really a full-on capitalist unless you can stop working and still have sufficient income derived from economic rent collected from somebody else who is working to sustain you in perpetuity.
I noticed NVIDIA skyrocket from billions to trillions market cap, so wouldn't NVIDIA be able to help deploy some a.i. simulatedly autonomous homo borg genesis to engage in scripted automated consumer lives to replace traditional homosapiens/humans consumers and takeover what otherwise was economy and market for human people but now they are replaced by redefining human and people to be displaced by the automated scripted a.i. controlled ones? This is already happening, right? Or did I just give them idea to make this a reality? lol, wait, why am I getting raided by the FBI right now? What's going on? HALP!
Just like the Turing test turned out to be the wrong problem, AGI is the wrong goal. As long as AI can stoichastic parrot its way to success, who cares if it's AGI? My washing machine broke and I want my robot to fix it. it doesn't need AGI to diagnose and fix the problem, it just needs to have seen that in the training data.
The real shift comes when the robots are dexterous enough to fix each other. that's when no new jobs will be created to fix them because they can do it themselves.
Dexterity isn't the issue, it's the lack of training content. Also, I would be happy with a real-time feed of a robot telling ME, a dexterous being, what I should do to repair something. I don't have access to such technology.
> It's clear that technological development creates a shift in jobs, i.e some jobs are lost, but new ones are created as a result.
That was true in the past, but as technology gets better that won't be the case. Yes new jobs will be created, but there'll be fewer and fewer.
Technology will allow for more generalized approaches that can be quickly adapted to new solutions. So new jobs will also be replaced quicker and quicker.
There's a recent NBER paper that predicts wages will increase in the beginning of the AI revolution, even as AI displaces some jobs. Eventually there will be mass unemployment, but only when AI dominates humans (cost & capability) at almost everything.
It's time we take on bigger and harder problems to solve. Living on another planet is a hard problem but only one company is working on that. We need multiple such companies tackling bigger problems. Solving climate change, dealing with plastic are other bigger problems.
I beleive there's no shortage of jobs. What if we start cleaning earth or reverse effects of human civilization on earth to make it more sustainable. The amount of people needed for that job are huge but we can't pay them at all because how our economies are structured. We need tectonic shift in how the world works today. Machines are taking human jobs, good. Now humans are free to do the work which machines can't do.
> Living on another planet is a hard problem but only one company is working on that.
There’s a lot of unknowns in that and ROI can take very very long time. Not many people can or are willing to take that risk. If that one company is successful then you’ll see that space flooded with new companies.
Our generation is highly advanced but we have become very short sighted. We have made a mess of our world because of it.
I remember a story about Oxford. When it was built they planted entire forest of Oak trees so that in 500 years when Oxford will need repairing they have ample amount of wood available. In that age people were capable of thinking 500 years ahead. And we with all our advancement can't even think beyond ROI. Living on other planet should not be seen as choice but something that is necessary.
All of that money is working as part of the economy, paying salaries and such.
It's not at all obvious that it's less efficient in an economic sense (although obvious not ideal from social sense)
Does NASA have any planetary colonization program?
They do plan the Artemis mission, but AFAIK that is about establishing a tiny scientific base on the Moon, probably with regular exchange of the crew. I don't think they proclaimed an ambition to settle massive amounts of people there.
For me "living on another planet" is only really SpaceX's goal. Build a semi-independent nation on Mars, with a million or more people necessary.
That is very different from a scientific base project.
SpaceX aren't just proclaiming things. They are building a skyscraper-sized fully reusable rocket with completely new engines to get away from the Earth cheaply.
The first crewed mission to Mars is very likely going to be a join SpaceX-NASA mission... to set up a crewed research station. Basically the same thing as Artemis Base Camp, except further away.
Yes, SpaceX wants to ship millions of people to Mars, but realistically they are several decades away from even starting that. You want to get a small permanent settlement on Mars (like a crewed research station), and get some experience with operating it, before you start sending heaps of people. And probably the initial focus will be just on growing that crewed research station (from 20 people to 200 people to 2000 people). And only then will you have enough information to really plan making it significantly bigger. I doubt we'll get there until some time in the second half of this century.
In the medium term, I think the Moon is a more realistic target for the private sector. The total cost per a person-year of a crewed lunar base is a lot less than a crewed Martian base, so you can pay for a much bigger lunar settlement for the same budget. You can also sell lunar surface tourism to the ultra-rich, and hope that economies of scale will gradually drive the price down; your average billionaire can spare a couple of weeks for a trip to the Moon, not the almost 3 years a Mars round trip would take. How about Hollywood filming on-location on the Moon? Reality TV shows? Professional sports competitions? All a lot more feasible given the much lower travel time (about 3 days) and light distance (a bit over 1 light second). I doubt any of these would be massive revenue sources (at least at first), but they'll be economically feasible long before their Martian equivalents become so.
I don't think anyone can say what will 'very likely' happen, but NASA often hires many contractors and also does much work internally (e.g., at JPL) for a major project. Also, there may be other national space programs partnering in the mission.
My point is - I doubt the first crewed mission to Mars will be a purely private SpaceX-led venture. Almost certainly be done in cooperation with NASA, quite possibly some partner agencies such as ESA, maybe some other firms involved too (e.g. Axiom). And that mission is essentially just going to be “Artemis Base Camp on Mars”, a long way off Musk/SpaceX’s dream of sending a million people there.
Not saying the dream won’t eventually come true, but it if it does, it will take many decades. Come 2100, I think there will be a lot less than 1 million people on Mars.
"SpaceX evaporates the moment the juicy billion dollar public contracts dry up."
So, not in the foreseeable future, given the new Space Race with China.
Ofc the government is an important customer for any launch provider (not just SpaceX), given that a lot of current space activity is military in nature. That just comes with the territory.
Edgar Rice Borroughs died before the first man-made rocket reached the orbit. Writing about Mars isn't the same as building a ship that can reach it.
I am not a native speaker of English, so I am not sure how much deprecation does the term "pipe dream" contain. But I don't think that manned flight to Mars is wild and unattainable fantasy. I would say that it is on the same level as manned flight to the orbit by 1945 - not yet here, but technically and economically feasible in a decade or two.
Well, I reacted to this sentence, which is a bit ambiguous. But I don't think there is a big gap between manned flight and colonization in case of Mars. Mars is very far, which means that from the very beginning, the crews will have to maximize their independence from the Earth.
You can supply Moon bases by cargo from the Earth easily, but people on Mars, even if only dozens during the initial period, will have to optimize everything for self-sufficiency as much as possible.
And if you prioritize self-sufficiency, you need more people, more robots, more factories = in a word, colonization comes naturally from that.
"I don't think there is a big gap between manned flight and colonization in case of Mars"
That's because you think the problem is getting there. Staying there is the real problem, and it's one that hasn't had any real advances. It's why things like Artemis are important and colonization of Mars is still a fantasy.
I tried to explain my opinion in my previous comment.
In the case of Mars, the gap between getting there and staying there is much, much smaller than in the case of Earth orbit or even the Moon. It is just too far for Earth to be of any help. Whatever goes wrong, the astronauts will have to solve on their own. Which means that they must be prepared for a very wide spectrum of adverse events.
This is a good fundament for a very resilient settlement - out of need.
Modern people residing on Earth don't generally have to be self sufficient for months. Even in isolated places like the South Pole during winter, windows of opportunity for transport to civilization open here and now.
On the other hand, the Sentinelese are self sufficient. Yeah, on a primitive level, but they manage it, because they simply have to. None of us here on HN would be able to live like the Sentinelese.
If someone really dumps a hundred people on a distant planet, they have to prepare the mission to be self-sufficient for at least a year or two.
It is painfully obvious to me that we haven't even managed self sustained colonization of the most extreme environments of Earth. Mars isn't just going to happen by accident because someone flies there.
Additionally, SpaceX is not going to manage it alone when a manned flight does happen. It will be NASA's Artemis program.
Prior to the advent of GPS, taxi drivers in many cities were tested on their orientation skills and had to memorize a lot of streets. I would be surprised if no one ever failed this part of the exam.
When intelligence becomes truly abundant and cheap, then human intelligence won't be valued by the market. You only get paid well because your intelligence is scarce, but that won't be the case in a future with ASI.
It's hard to predict what will be valued, maybe personality traits that lend themselves to jobs like social media influencer, politician or actor, because people may still value real humans in these jobs.
I don’t agree with this take. A more accurate and precise way to describe the effects of automation is that it creates more human-centered jobs, some of which may require higher IQ but many that do not. Your average massage therapist does not have a higher IQ than your average Amazon line factory worker, for instance.
>The issue that I see addressed less often is that the new jobs require by definition, a higher skill set
No they don't. If AI reached the point where it was capable and willing to do all white collar work, there'd be no more need for humans to do that kind of intellectual work. What would still be needed is service jobs that rely on the "human touch", and trades that AI lacked the dexterity to do. We're already seeing that now, with AI posing a greater threat to programmers' jobs than to plumbers or electricians' jobs.
But exactly as the person you replied to said, they increased the average intelligence needed to do the new jobs. That leaves so many marginal people unemployable. They could have maintained horses but working in cars is harder. I like this observation.
I'm not convinced that animal husbandry is less skilled than working in cars. Different skill, and as I've never done it I can't be certain, but horses are wet and messy biology with brains that are terrified of anything they've never seen before. Production line work I did do as a summer holiday job during my A-levels aged 17 or 18, it wasn't skilled work but also that was HVAC production line not cars.
The specialization has certainly taken off - people are much more specialized in their jobs now whereas “farmer” was really a jack-of-all-trades with passing capabilities in many different skill sets.
There is a thing to say about "unnatural-ness". Handling horses up to a point had to be more intuitive and more approachable, we've been around horses and other mammals since forever. Around spreadsheets? 40 years, max.
They actually resulted in a decrease in overall skills required. It takes a lot of skill to use a loom and make a napkin, the same is not needed for factory work yet you can make 100 napkins at the same time.
Similarly, we had the rise of the service industry in the US - manufacturing required a lot of skilled labor; retail and wait-staff do not require the same skill.
As a software developer, I've personally eliminated many jobs. Software was eliminated entire classes of jobs. Almost all investment in technology by businesses is about cost reduction and the number one cost is labour.
I think we're past the point where technology is making new jobs -- all that low hanging fruit has been gone for decades now. Growth now is all about optimization.
That's undeniably true, but not the point: obviously new technology eliminates entire classes of jobs (not much call for telephone switchboard operators these days), but ultimately if more jobs are created than are lost, we're fine.
I personally do wonder, though, if many of the new jobs that are created are worse jobs. For example, I have not yet taken any work in the "gig economy", but most of it seems pretty miserable, with shit wages.
> but ultimately if more jobs are created than are lost
For people who believe more jobs are created, it seems like they rely in chaos theory or something. They can't see how or where these new jobs will be created or how automation leads to it. We automated physical labour so we increased intellectual labour. Now are automating intellectual labour too so what's being increased now?
amount of people who can do stuff that is considered "intellectual"?
It is not different from happened with guns - with introduction of guns, military has grown bigger not smaller - in comparison to knight era.
Now you will have more people who will be able to produce art, write music, songs, develop games, write stories and so on. People will be able to produce scenes without months or years of learning how to use photoshop, after effects and co. We now have more painters than in da vinci era and we have more musicians than in bach era.
Same with software engineering - ability to do stuff easier will produce more people doing that. Not less. We do not code to the metal much anymore - most of the software engineers do not use assembly anymore and higher level languages simplified stuff that required hard learning in the past - yet the amount of developers is more than ever.
Software development is a bit of an exception because we have, probably by a whole order of magnitude, too few developers for the world.
However, what about all other things that are intellectual? Accountants, word processors, assistants, etc? We need fewer of those people than ever before.
As for the effect on art, the greater the unskilled people who can produce art the less it's worth. AI will eventually drive the value of art to zero. Even now 99.99999% of skilled musicians are unable to make a living doing it.
You do realize that most governments run massive jobs programs to ensure this number of available jobs stays high, right? In the US we give massive tax breaks in exchange for hiring numbers.
Computers may have very well reduced job numbers but we're running a contrived system at this point.
The embryos I just had implanted in my surrogate are the result of eugenics. I hand picked the egg donor based on SAT. When the embryos were created, I got their 23andme reports and validated that the ones implanted had the appropriate SNPs for intelligence by cross referencing snpedia.
Almost everyone does step 1, but very few people do step 2.
Select eggs/sperm and make embryos at fertility clinic.
Ship them all off to Genomic Prediction.
After you get their simplified test results, call and ask for the full report that is compatible with 23andme.
Search snpedia for keywords like intelligence, retardation, educational attainment and cross reference their existence/non-existence with the 23andme text file.
Have fertility doctor implant the embryo of your choosing.
If you have/are a female, the price is dramatically cheaper. Otherwise its about $200k (all post-tax). Embryo creation $45k, surrogacy agency fee $30k, surrogate $70k, baby delivery $25k, embryo testing $6k.
It's difficult to hold the paradox evident in most of the commentary here: both the desire to free people from the unhealthy repetitive tedium they are bound to participate in daily, and the overarching need for "jobs".
I'm much happier watching robots take over dangerous, mundane tasks, than seeing them edge into creative, intellectual, and decision-making work. Wasn't that the dream? That robots do all our work and we just chill, enjoy a life of leisure and fulfilling activity?
Looking back in history across all the -isms, the surplus created by those labouring has rarely gone to benefit them (unless we look as far back as hunter-gatherers). I don't know that advancing technology is the credible threat here; I think the misalignment is deeper than systemic even, as it persists across lifespans of civilisations.
> I don't know that advancing technology is the credible threat here
The thing most people don't realize is that the robots in question aren't really even taking jobs that existed prior to the ecommerce model, and the work the robots are doing used to be handled by the consumer.
Details:
1-The robots are taking over the job of moving goods from the storage shelf to the packer.
2-This is pure "walking" time when humans do it, 10 miles per day in a typical DC, not value add.
3-Before ecommerce, with brick and mortar, the goods were picked in bulk and shipped to stores and placed on the racks, and the consumer was the person that walked to get the item and bring it to the packer (cashier). Some of the walking was done in the DC, but not nearly to the same extent.
4-There are multiple different ways to solve this problem of humans walking all over the DC to pick one+two piece orders in a way that is efficient. "Robots" in the form of moving shelves (Amazon), or conveyors+sorters with bulk pick then sort by order, or etc.
This is insightful; thanks for pointing it out: "the robots in question aren't really even taking jobs that existed prior to the ecommerce model, and the work the robots are doing used to be handled by the consumer."
> I'm much happier watching robots take over dangerous, mundane tasks, than seeing them edge into creative, intellectual, and decision-making work.
I used to think this way, and to an extent I still do, but the reality is that not everybody wants to do intellectual work or decision making as a day job. Some people prefer to do labour, or to have their work relatively unchanging. I know smart and intellectually capable people who have said they like the fact that they don't need to think that much in their day jobs.
(Those preferences are of course a spectrum, as are jobs.)
For lots of us, automating away the mundane work is a dream, but at the same time the development puts pressure on those who don't find it desirable. One person's dream is stressful to another.
> Wasn't that the dream? That robots do all our work and we just chill, enjoy a life of leisure and fulfilling activity?
Unfortunately, I don't see that ever happening.
As jobs get lost to automation, unemployment will increase, and eventually our economy will collapse.
Even if your McDonalds is automated, they're still going to charge you $15 for that meal. In the fast food world, labor is typically ~20% of the cost, so even if they reduce prices because labor is no longer a factor (they won't, after all, because they already know you'll pay $15 for your meal), it'd only reduce the price to $12.
I mean, it's already happened over and over again throughout human history. Hell, in the US ~100 years ago, 90%+ of the population was farming. 'Automation took our jobs!' and now ~1% are farming.
Work doesn't go away; it just shifts. Unemployment/collapse is a pipe dreams, the human components will simply move when large market shifts come in. Yea, they're disruptive - what will an older career truck driver move to once no one is driving? But the next generation coming up will find new niches that are still needed or desired.
A life of leisure is the mirage-like carrot that the PTB dangle in front of our noses. Simultaneously, they intensify advertising to keep us on the hedonic treadmill, and thus trapped in a life of debt peonage.
I hate government authoritarianism, but in my mind the only answer is to force robot labor tied to individual humans instead of corporations.
Some kind of coded link so one human operator must be living and healthy and somewhat more than homeless to control the edge cases of the robot messing things up.
As the companies get bigger and more monopoly they at least have to support more humans at least somewhat.
As Amazon really using that humanoid robot in production? Or is that just a demo?
The first big breakthrough was when Amazon bought Kiva, which makes those little mobile platforms that move racks of goods around. Those are mechanically simple and cost-effective, and have been very successful. Amazon is now making about 1000 units a day of the current model.[1] This is about 10% of the human birth rate in the US.
Moving standardized totes around automatically is decades old. Picking things out of bins has been difficult, and gets better as computer vision gets better. Amazon is still struggling with that.
> The significant investment in robotics showcases Amazon’s commitment to innovation in its supply chain and highlights the company’s belief in the synergistic potential of human-robot collaboration. Despite the massive scale of automation, Amazon emphasizes that deploying robots has led to the creation of new skilled job categories at the company, reflecting a broader industry trend toward the integration of advanced technologies with human workforces.
An article about robots that's written by a robot. It's robots all the way down.
Military drones have to be covered in a Farraday cage layer to prevent jamming, so that would double as nuclear EMP protection. Finding a place to charge afterwards might be tricky though...
That’s too hard to do. What is “a robot” and “a job?” Replacing workers is easy enough to calculate that with, but what if I start a brand new factory that only ever used robots? The solution is wealth tax and UBI.
Get rid of regular income tax, then do a progressive wealth tax, which is enough to pay for a progressive UBI. This way, it’s no problem if you lay off all your workers to replace them with robots. In fact that’s great! Because now these people have a UBI, which frees them up to go find (or make) other work. Meanwhile the factory owner’s taxes should be going up, since his wealth is going up since he doesn’t have to pay all the workers, unless he’s still spending that money to further invest in things, which is also great.
The biggest problem really is people just hoarding wealth. If they are spending or investing it, then that’s not a problem.
Very few people hoard wealth. You imagine wealthy people like Scrooge McDuck with a room full of gold coins. It's almost always invested in stocks which helps companies grow or in bonds which helps finance public works.
Pretty sure a diverse portfolio will feature gold. It might not be in gold coins in the owners basement though.
Investing in stocks where most companies are doing their best to suppress worker compensation or out-right replace them with robots doesn't feel like it's doing much to help the average person either.
Public investment is on a downward trend since 2008 when governments had to save the bankers. What government spending there still is often has to filter through private service companies that are given government contracts. This usually results in good dividends and bad service.
You need to start with no/little increase in taxes from corporations and see how that works. Previously unthought of loopholes will keep the status quo of low corporate tax payments.
> The biggest problem really is people just hoarding wealth
That is a pretty substantial problem and doesn’t seem close to the biggest one.
About 20 years ago I was a programmer at a manufacturing company. I worked on a project to automate part of materials requirements planning. The woman who was in charge of it was spending about 20 hours per week wrangling Excel spreadsheets to figure out what materials they needed. It was all based on data we had: inventory, orders, etc. A few weeks later, and with a fair amount of her help, I'd automated the whole thing. She was happy, it was her least favorite part of her job and she desperately needed time for other things. I was happy, I was showing that I was worth what they were paying me! (Salaried at the equivalent of $12.50 per hour.) BUT: That project took half a job out of the economy. A decent one at that, a nice cushy desk job.
Automation is definitely having an impact on labor demand, with software driving out some jobs much the way cars and trucks drove horses out of our economy. I think it points to an increasing bifurcation, with jobs being high skill / high pay or low skill / low pay, without much in the middle.
Of course here in the U.S. this is also being driven by our monetary policy, our tax policy, our trade policy, our labor policy, I could go on. I'm in favor of UBI, but that seems about as likely as the use of any of those levers to reduce income inequality and spread the wealth our society generates in a way that supports most of those contributing to it.
But, some already operate at their peak. Then, you automate that and they can only shift to unemployment. The conversation shouldn't be about replacing the most skilled workers. The conversation should be about what happens to the average worker. That has an average IQ level. That can't learn something to compete with the automation. These people have families. These people will suffer.
> The conversation should be about what happens to the average worker. That has an average IQ level. That can't learn something to compete with the automation.
There are many tropes to unpack, including that IQ is somehow fixed and people's fate, and that even the "average" person "can't learn something". Also, it of course glorifies and cements the power of the wealthy who think they are smart and very special and deserving of everything.
Also, history has shown that people overwhelmingly that people can learn higher skilled jobs en masse. Just look at universal literacy - did anyone say, 'the average person can't learn to read'.
The US has long believed that anyone can achieve and succeed with hard work. That's been my experience, but that the limit is almost always in opportunity.
That's a good point but things do change. Malthus was correct about humanities past but won't about the future (from his point in time).
Even aside from that, Learning takes time for everyone. At a certain point we're automating new tasks faster than humans can learn them.
Scaling works against humans too. Say it takes 1 year to train a human in something, and 1 year for a robot/AI. well it might take 1 year for the first AI, but copying software is easy. Training the next person takes another year. Even if it's parallelized and you save some time, the cost of training the marginal additional person is was larger.
A lot of that subconciously assumes the doomsday outcome - machines and automation will sweep people aside, and then the reasoning becomes circular.
> At a certain point we're automating new tasks faster than humans can learn them.
That assumes the pace of automation is increasing, but similar concerns have been around for a long time, going back to the industrial revolution. Read Dickens or HG Wells (though a specific cite doesn't come to mind), or look at the 1927 silent film, Metropolis.
But right now businesses can't find enough employees.
> Say it takes 1 year to train a human in something, and 1 year for a robot/AI. well it might take 1 year for the first AI, but copying software is easy. Training the next person takes another year. Even if it's parallelized and you save some time, the cost of training the marginal additional person is was larger.
That's how automation works. Then the people go on to the higher skilled jobs that the machines can't do, including designing, manufacturing, operating, and servicing the machines. Cars made the entire horse industry redundant; calculators and computers put lots of human calculators out of work.
Yet today, with a much larger population, employers can't find enough workers.
If things like that didn't happen, then productivity wouldn't increase and we would be able to afford more shelter, food, healthcare, education, etc.
> Has anyone claimed technological changes will impact every group and subgroup on a literally equal basis?
No, but too much inequality is not sustainable. Since in a democracy everyone has an equal vote, there will be a force that balances things out over time one way or another.
We already are at the point where the ultra rich just buy politicians. Hell, we were ever since oil production was mainstreamed and the oil tycoons were able to tell the US to go to wars for oil.
I would disagree with the inference that menial tasks are automated and interesting work isn’t. I can clearly see a future where the pleasure is automated away leaving the stressful uncomfortable tasks to the human.
With better oss tools, we've been able to achieve more with less people but for a long time there was more demand for skilled developer than what is available in the pool, hence the decent salaries coming from the market squeezing that pool. If robot takes over and the pool of available labor expand instead of shrink, the effect is the exact opposite, pressure on salaries from the remaining participants, worse conditions, ...
Assuming the world is moving towards more automation and high skilled(not sure this is the right term. But places where you need to be more adaptable and think more) jobs. Also assuming for a moment todays adults are the transitionary generation.
How do we prepare the next generation for what is to come? It feels like schools are still stuck in the industrial age. How do we teach them not math and science but the actual act of learning/adapting. Our parents could not predict what the world would look like today, neither can we. How do we educate the next generation on the foundational skills rather than specific skills which they can learn on their own depending on what the situation calls for?
This is an idea I am thinking about, so I would love to hear other opinions and thoughts.
>How do we teach them not math and science but the actual act of learning/adapting.
This is literally what they're trying to do. People complain about the way things like math (concepts instead of rote memorization of facts) and spelling (stems and etymology instead of rote memorization) are taught. People complain that kids aren't learning cursive and other things like that.
But schools are trying to teach critical thinking and how to reason your way through problems, rather than blanket "facts".
I, personally think teaching kids critical thinking is great. Here in Eastern Europe they are only taught useless facts and not to think too much on their own, remnant of the past I guess...
Eh you require a lot of factual knowledge to know how to think. I mean humanity has been pretty smart for the last 10k years or so, but without understanding of scientific laws and a massive amount of world knowledge breakthroughs were very slow compared to the last 300 or so years. We need these facts + philosophy + critical thinking, but it seems like no one knows how to teach this.
> That project took half a job out of the economy. A decent one at that, a nice cushy desk job.
It's effectively a form of the broken window fallacy [0]. You don't actually have a closed system when you say a "job was removed". The company one way or another got more efficient - more widgets were manufactured per $ spent. It might be tiny but this saving was then passed on to every downstream purchaser of the widgets who, with that left over money .... could employ someone else to do something. By the time you add all the net effects back to regain a "closed system", economic theory says you actually have more jobs.
But the key here really is that this doesn't account for temporal effects, and especially when there is rapid disruptive change, you can absolutely have a huge dislocation and mismatch of labor supply / demand in the economy.
> It might be tiny but this saving was then passed on to every downstream purchaser of the widgets who, with that left over money .... could employ someone else to do something.
So a company used technology to automate someone out of a job but some downstream company isn't also going to do that with their new money?
The reason technology lead to more jobs is because it created exponential growth. The lesson of the last 200 years is that we can grow indefinitely. But that is already falling apart. The growth that is happening now is optimization. And it's happening at every level of society and it's not good for workers.
This sounds a lot like one of the last projects I worked
On, where some significant operations were to be assigned to “AI” which was meant to mean Artificial intelligence for marketing slicks but everyone knew it would be “Actually, Indians” for the foreseeable future.
Automation can eliminate lots of jobs without replacement jobs.
But other times will also created new higher level jobs.
Especially when automation impacts a companies total cost for a high-demand independently consumed service or product. So that decrease costs can translate into decreased pricing for an upwardly elastic demand. And the revenue growth results offsets the need for fewer workers per revenue.
—
But most jobs within a company are only part of a product or service. With other parts not as automatable.
In this case, even automation that produces higher level jobs is likely to produce far fewer of them. Since savings in part of a value chain increase profits but impact total costs, and therefore pricing, at a much lower percentage. So increased demand is less likely to offset job reduction.
This is also true for any product or service from one company whose demand is anchored to any complimentary services or products from other companies.
—
Worse still, automation taking over only part of a high employee count, low skill job is very bad for employees. The number of workers is likely reduced, and as the job becomes simpler, their replacability goes up. Their value, and therefore pay and respect, go down.
They start to get managed ruthlessly, like just another part of the overall automation.
I expect the vast majority of Amazon’s headcount, across each of its logistical divisions, fall into that latter category.
This seems a bit backwards. Most demand for automating existing jobs is driven by the specific goal to reduce workers and labour costs, so of course there will be fewer jobs afterwards than before - otherwise the automation didn't fulfill its goal from the companies' POV.
Automation is to reduce cost per revenue, by reducing jobs per revenue.
Which is not quite the same as an absolute desire to eliminate jobs.
The result is initially lower employee count, but in cases where the automation significantly impacts total costs, and therefore pricing, of an independently consumed product or service, with upwardly elastic demand: the lower cost & pricing will trigger growth that also creates a job growth effect.
But those are special conditions.
As I point out, and you do too, in most situations efficiencies reduce jobs and improve profits, more than they increase demand that would create more (or mitigate lost) jobs.
Yeah, that's essentially saying, "yep, your slice of the pie will become ever smaller, but that's fine, because what you get in absolute terms will stay the same - the pie just has to keep growing exponentially forever".
It's obvious that this is not a sustainable solution.
(And that's assuming the company fully passes the reduced labour costs on to the consumer and doesn't just pocket it as increased profit margins)
But, now that amazon is saving money on labor costs, where does that extra money go? Into the pockets of shareholders, who may then go out and use that to start other businesses which hire people, or increase their consumption of goods and services which also results in greater employment.
Meaning money will likely go to continued investments in automation, corporate acquisitions, capital investments (high & low tech) like real estate for logistics, etc.
The thing that's missed here is that it's not clear that robot purchase, maintenance cost and depreciation would be cheaper than paying 7.5 people a salary... Not to mention that the robots aren't going to be spending part of their salaries at Amazon as a worker might. Also, robots won't be eligible to vote in elections to support Amazon's political aspirations which give it access to billions of government contracts.
It seems we have gone to the opposite extreme since the days of Henry Ford where he wanted to pay his workers enough for them to be able to afford the cars they made.
That belief about Henry Ford is a myth. It doesn't even pass the smell test, if a company is only selling to their own workers they can never make a profit (note that any company, even a large one like Ford or Amazon, is different from an entire economy).
Ford raised wages and cut hours because his annual turnover was 300%. He ended up making more money through savings in hiring and training costs and workers staying long enough to gain experience.
In a way it was an expression of worker power. Which only applies because workers are humans, robots don't get tired, they don't have feelings, they don't leave for a competitor paying more. The best ones are more expensive than humans right now, but robots get cheaper over time. Ideally we don't want human labor to get cheaper over time.
I don't think it was implied that Ford's entire customer base was his own employees, just that he paid them enough to afford the cars the company manufactured. I agree with the rest of this though.
Ford was a ruthless capitalist and antisemite. He never cared for workers affording his cars, only that they bought them. His union busting gangs were some of the worst.
Ford offering higher pay and better working hours were for employee retention which was important at the time.
I wish I could believe that all this automation would lead to more leisure time for all of humanity, but it certainly won’t happen on its own out of generosity from the robot owners.
All what you say is true, if the wage share stays constant.
The "working poor" phenomenon where you can't pay for necessities like rent, healthcare etc. while working full time is the result of declining wage share.
Wage share has steadily declined in OECD countries since early 1970s. In the US since 1960s. https://fred.stlouisfed.org/series/PRS84006173 Wages make smaller and smaller portion of national income while capital share increases.
Health insurance and other benefits, household composition have hurt wage share. Not that it didn't decline in real terms for the lower skilled workers, just less than the naive chart you put.
Here is good reading on the topic (maybe my favorite paper of the last decade):
Your first link points out that income inequality is still growing, however. Which unambiguously means that those at the top are increasingly taking larger and larger slice of the pie.
> people don't actually want to tradeoff their labor time for leisure time all that well. They value consumption too much.
We don't actually know that. Most people don't have the option.
For all but a select few industries, there is no such thing as "work a part time job, and trade the rest for extra leisure" since almost no part-time job can sustainably cover cost of living. (Heck, most full time jobs don't really cover basic cost of living today).
> There are plenty of 15 hour jobs out there. But you'd need to have a smaller apartment, an older car, and eat more lentils.
The average wage in Michigan is currently $21/hr. ($36k/yr after federal, city, and state taxes). The average basic 1-bed apartment here is about $1300/month or $15k/yr. Minimum income requirement to qualify for an apartment is usually 2.5x to 3x times monthly rent.
Math that out and you can quickly see that the vast majority of people simply aren't going to ever afford basic housing on 15hours a week. Or even 30hours a week.
--
And on the flip side, when it was actually possible to pull this trick off, more people used to use it. In the 60s and 70s, when you could actually live off a part time 20hr a week job, people routinely did so (and then used their free time to do things like enroll in college and pay for it on the side).
You don't see that anymore, because wages have fallen so drastically, and costs risen so drastically. It simply doesn't math out for the vast majority of folks.
There's a youtuber I've seen who I think is mostly legit who spent not much money on land in the middle of nowhere and built himself a one-room cabin and an outhouse on it. He's up-front that he isn't actually living off the grid or even in the cabin full-time, but his point is that if you live a very modest life and are handy, you can live on just a side hustle. Like you said, you're unlikely to meet people like this in the (sub) urban tech bubble.
I concur. They're not looking in the right places. Musicians and artists would know about these sorts of hustles, but not your typical HN bubble dweller.
Agreed. I think this is clearest when you look at benefits, e.g., the many American companies that only offer health insurance benefits to their full-time workers, but you see it in terms of pay, too. Most middle and high-paying jobs are exclusively full-time (and some expect 50, 60, or even more hours a week).
I live in a country where health insurance is individual, not employment based, so maybe that's part of why we offer 80% (most common) but also 60 and 40% positions?
What if you bring the ideas of measuring productivity and wage growth of employees? How have productivity, wage growth, and corporate profit changed over the decades since Keynes? What was that 15 hour work week, how "productive" and paying is it supposed to be?
If things go as before, wage share continues to decline while unemployment does not increase.
Wage share has steadily declined in OECD countries since early 1970s. In the US since 1960s. https://fred.stlouisfed.org/series/PRS84006173
Wages make smaller and smaller portion of national income while capital share increases.
I don't know for aging economics this is somewhat good. But for developing countries it's horrible. Where humans has to search work each day to find food for night
Is humanity really at a point where we can't think of anything more useful than "carry things about" and "put things in a box" for those people to do? If that's really the state we're in then a bunch of robots doing it instead is the least of our problems.
It doesn't seem like we are close to robotics providing housing, food, and healthcare for nominal costs to everyone. Seems like we still need a lot of humans working in those spaces.
In particular, durable goods (and human bodies and even hair) need a lot of maintenance. So far, at least, robots aren't especially gifted at maintenance tasks.
Isn't it also a problem that we've put people who's capabilities are equal to those kind of tasks into the "useless" category? Not everyone is cut out to be a programmer, which by the way, will also eventually be in the "useless" bucket. In reality, "better things to do" usually means just a lot of people in poverty.
Welfare systems are nationalistic. The US will tax companies using robots and redistribute it to Americans while the people assembling widgets in developing countries will be out of a job. I'm still optimistic that the raw increase in economic productivity will somehow sort everyone out eventually, but it could be bumpy.
And once it's cost effective for robots to do it, you might as well move the factories close to where the stuff is needed and transport raw material that's more fungible.
It's a good thing actually given that the worker conditions are pretty terrible there.
But it is only a good thing if the increased productivity translates into better wages, increased vacations and reduced work days and not helping Jeff getting a bigger yacht.
And for that, it must be an active government policy.
Soon Jeff will have ten yachts and his job each day will be choosing which yacht to use once he has automated himself out of work. It will be gruelling.
The last 200 years of automation being followed by new jobs is slightly stronger evidence that further automation will lead to new jobs, but I don't see any way that could remain true once AI is actually human performance when on hardware that runs under some appropriate electrical power threshold. Induction and Turkeys.
> The real problem is when others follow Amazon in scale, leaving humans to survive in a new order.
No, the problem is if humans in the set of those who have to work are able to keep ahead of the creation of new automation.
UBI short-circuits the problem because then nobody would need to work. "Who pays for UBI?" you may well ask… well, Amazon can't ship anything when nobody has any money to buy stuff with.
Separately from that, AI currently needs a lot of examples to learn from, relative to humans, so "business as usual" with no further AI breakthroughs will mean lots of people shifting employment every few years. That scenario doesn't force the world economy to choose between UBI and collapsing due to all the jobs being automated and therefore nobody having any money to buy the products made or delivered by the automation.
Only from the selected few allowed to work as druids on the AI tower, after the letcode trials, while remaining mankind struggles with what is left as jobs from a corporate managed government.
I can't even parse this sentence. If you're trying to deny a part of my previous comment, it would help if you indicated which bit you're trying to replace.
Likewise, but "we pissed everyone off and they refuse to work for us any more so we had to invent and manufacture robots at a huge and expensive scale" is very different from "we made robots for funsies with our big pile of money and you're all redundant now".
I have no idea what @pjmlp is saying, but I myself would caution against using the past as too strong a guide to the future — we don't have pre-existing examples of tools that can out-think their owners.
(The closest historical examples to this description are feudal serfs and slaves, the latter is a much broader category than just the USA's example, and these are also bad examples from which to imagine the future for a variety of different reasons).
The early usage of the term "singularity" with regards to computers and automation was about the unpredictability of the future; even now, I cannot say with any confidence if we'll get a future more like that of The Terminator or more like that of The Culture, let alone smaller-scale and less dramatic dystopias and utopias.
When I worked in an Amazon Warehouse, while a bathroom break was considered time-off-task, it wasn't a big deal to use the bathroom (or take some other break) whenever one wished.
Maybe jobs that require moving around should be split into 2 hr shifts, and then people with sedentary jobs can moonlight for getting in their cardio, instead of paying a gym, or walking pointlessly.
If Amazon would offer a part-time shift option where it was 1 day a week, I'd probably do it, mostly for the excercise, and partly for the extra money.
Isn't this just a thing on delivery trucks? And it's pretty standard industry practice, not just Amazon? And in the scale of worker abuses, this is pretty low on the scale and more pragmatic than anything.
Is there an index that ranks companies by their benefit to society? Something like amount of taxes paid, number of employees, benefits, environmental impact, etc?
Every attempt to algorithmically boil down a complex economy into a single number does more harm than good. Figure out your own priorities and decide for yourself.
Any company? Maybe. It depends on what the company does. Either way, that was just an example of a metric off the top of my head to support the question, not a proposal for a solution.
Value delivered or increase of value per unit of labour is very murky to measure, but something to think about. In some sense Amazon improving efficiency of retail could be a good thing in general improving the purchasing power of average consunmer.
I assume this is being upvoted because HN thinks this is good? At least that's what I hear in AI threads...
AI just allows humans to work more efficiently right? The few fulfilment center workers left are probably now 100x fulfilment workers now that they can manage a fleet of robotic workers instead of human ones. As HN says, these types of innovations historically only create new job opportunities so there's probably loads of new amazing jobs these replaced workers can go do instead.
For those who are worried about the future we're creating for their children, don't worry, because when robotics and AI can do everything humans can do (and better) businesses will still want hire humans, because they still hired humans when the combustion engine was invented. I'm sure it will be the same this time. Don't be so anti-progress!
Amazon's role in society is not to provide menial jobs to keep people busy though. If anything, the automation provides good, innovative jobs in robotics. Those are the sorts of jobs I think everyone should have. Leave fetching and carrying to robots, and educate people to be useful in more meaningful, fulfilling, and impactful ways.
Suggesting that robots are bad because they take jobs from people doing horrible, boring jobs just says that you want people to carry on doing that work. That's not a good look.
The people working in the low skill jobs being eliminated aren't capable of contributing in more meaningful, impactful ways. That's why they work in low skill jobs.
If the robots are the ones altering the position of matter at or near the earth’s surface relatively to other such matter, there are still roles for telling them to do so, and history suggests the latter work is capable of indefinite extension: there are not only those who give orders but those who give advice as to what orders should be given.
Amazon told us it was all "sensors" because it fit their company narrative to do so.
I am not saying that Amazon doesn't have 750k robots and hasn't laid off 100k people... but they usually have some seasonality, plus, the quoted number is from 2021, the height of at-home shopping.
"The world's second-largest private employer employs 1.5 million people. While that's a lot, it's a decrease of over 100,000 employees from the 1.6 million workers it had in 2021"
I think a bit of skepticism is in order, is all.