Hacker News Reader

Comment by jordwest | Hacker News Reader

These discussions are always focused on the "robots". What about the spreadsheets, the computers, algorithms, industrial control systems, Python scripts, voice recognition systems, data pipelines, software powered gig economy, algorithmic traders? These types of automation have been in place for decades, and all of them cut costs for the owners of them and reduce the number of workers needed. Many of them can be operated with minimal training.

The benefits of those efficiency improvements are already going to the owners of the capital, not to workers, and this has been going on for decades. If in 1980 you were able to replace 4 clerks handwriting transactions in an accounting ledger with one clerk typing them into a spreadsheet but pay that one clerk marginally more than a handwriting clerk, who benefits most in that equation?

We're always focused on the some day maybe robots, but automation has been changing society and (I'd argue) increasing inequality for decades, and we've been completely blind to it because we're always talking about "the robots that are coming to take our jobs". We already need to find ways to distribute the benefits of these efficiency improvements more equally. In fact, we needed to years ago.

I wouldn't be surprised if this explains a lot of the productivity-pay gap [1]

[1] https://www.epi.org/productivity-pay-gap/

View on HN

I’ve owned a CNC machine shop for 40 years. As we’ve added automation, so have our competitors. All gains from automation have been competed away - there are no gains going to the owners of the capital, as we are all competing. Meanwhile, the real wages for the employees have been falling for those 40 years.

Instead, I’d argue that returns are going to IP holders.

Well, and to the people buying your products, presumably

Who may also have margin competed away.

And to the overall economy - more stuff with less resources is growth. But that's all going to capital, look at the s&p500 vs wages over the same time period.

OP owns a CNC shop which is almost tautologically capital

Saying a CNC shop is comparable to a an S&P500 company because they're both capital is like saying a retail clerk is comparable to an NFL player because they're both labour. While true, it's not really relevant to this discussion.

I agree the parent comment should be more specific, but his point stands: gains are going to a few companies that automate everyone else's processes or benefit from the huge economies of scale that automation provides.

Haven’t many of the gains gone to the customers who are buying CNC parts? That competition and reduction in labor cost per part means that some people can now afford CNC parts who couldn’t have afforded manually crafted parts previously.

> Haven’t many of the gains gone to the customers who are buying CNC parts?

Potentially yes, as long as the competition doesn't destroy the manufacturing landscape, e.g. by emergence of lowest-common-denominator or monopoly suppliers.

Or to your customers who are getting cheaper products?

On the macro scale, this results in deflation, which the central bank attempts to cancel out by implementing interest rate cuts, which pulls money into stock market valuations and mortgages, so ultimately the beneficiary of CNC mills are real estate and stock owners.

Gosh, that's bleak.

I've recently been thinking a lot about policy changes that might help with housing affordability, and the same thing keeps happening: every time I think I've come up with something clever, I eventually realize that I've just accidentally made things even better for people that started off with a lot of capital.

It feels like aside from explicit redistribution of wealth (raise minimum wage, tax high income earners more, raise high-end property taxes), you can't change _anything_ without making the rich richer and the poor poorer.

My preferred policy would be to have central banks adjust rates to maintain inflation at 2% average, and more gradually have governments adjust UBI to maintain a central bank interest rate average at 5%.

(You can certainly have the input to one feedback loop be the output of another).

Then structural deflationary forces would manifest as UBI increases instead of interest rate decreases.

I think the positive real interest rates would ensure the UBI isn't all captured by rent increases, although it's hard to be certain unless it's tested.

The 2% figure is arbitrary though. The correct thing to do is just stop printing money entirely. The idea of a 'deflationary spiral' is one of the many not very well founded bits of economics that's really more like psychology, but inflation is always and everywhere a redistribution of wealth away from savers and people at the bottom, towards those who print money at the top. There's a deep lack of justice in that.

It's arbitrary but I got those numbers from economist John Taylor, who knows more than I do. Maybe zero inflation would be better.

The "savers" don't stop saving just because the money supply stays the same. They will actually save more until there is mass unemployment and they are structurally unable to save because of the aforementioned unemployment.

We got at least two world wars out of "deflationary spirals" and every single gold standard that was attempted failed during deflation.

Also, liquidity preference is a systemic redistribution from the bottom 90% to the top 10%.

The people that benefit from inflation are the savers and bankers in the top 10% because they can live off the interest or capital yields which must be higher than interest.

Bankers and savers benefit because governments, people and companies pay debts first before they pay anyone else. If the cantillon effect exists it exists in the boring form of those who own the most get paid the most. The proximity argument is nonsense because there are banks in every city. You can "print" money yourself by getting a mortgage to buy a house and many people do.

"They will actually save more until there is mass unemployment"

This is pop psychology and there is ample counter-evidence for it.

World War 2 was not caused by monetary policy! A hypothesised, partial cause that Hitler certainly exploited was crippling debts. But that wasn't solvable by economists. Do you think the allied powers that defeated Germany in WW1 would have sat back and tolerated Germany inflating away its war debts?

Land value tax

Yes you can, and it is pretty easy. Build.

Every single, new building puts downward pressure on the average price. The moment supply outstrips demand, prices goes down, housing market stops being reasonable asset, money outflows causing further collapse of the market and houses are dirt cheap.

I think that evidence from China's real estate bubble suggests that no matter how much of a supply excess you build, it might not bring down prices as long as the macro environment favours real estate as an investment, because demand for perceived-as-safe and fast-appreciating investments is nearly infinite. (The excess supply can probably make the eventual bubble pop more dramatic, though, but it seems like it won't be triggered simply by a supply excess).

At the very least, a supply-solution proponent would need to explain how China's property supply glut didn't bring prices down until the government cracked down on credit.

You can introduce negative interest rates to pay off debt and reduce the money supply. Do the same to land via a land value tax.

A lot of equity derives its value from real estate so the obvious escape into stocks won't work in practice.

Negative interest rates increase the money supply though, and explode the valuations of real estate and other assets.

On the macro scale more things are going on, and deflation isn't really considered to happen that way.

Why do you say this? Most economists I've read do consider technology, automation and demographics as the primary underlying deflationary forces, beyond the to-and-fro of business and financial cycles. I could come up with a lot of citations if you'd like.

Specific goods getting cheaper sure, but I've only ever heard of demand issues e.g. in the great depression, being able to pull down averages of many different prices.

It's not just specific goods; technology and automation reduces the cost of almost everything in the consumer price index, from food to transportation to energy (although probably not shelter). I think if you search for "technology deflationary" you'll get lots of views on this, but not many in opposition to it.

That said, deflation only ends up happening in specific goods categories like TVs and computers, which are at a particularly sharp point in the technology curve, such that they deflate even faster than average. That's again because the average CPI growth is governed by a feedback loop, where the central bank adjusts interest rates to target 2% inflation, so as, say, food gets cheaper to produce, monetary stimulus is gradually added to the system to combat the falling costs and encourage more consumption. This is done to avoid the kind of deflationary cycle that you're talking about. (Central banks want to avoid deflation so that debts don't become onerous to repay, which can create a depression).

But the result is that while technology is deflationary, it does not create deflation, because that's maintained by a feedback loop; instead it manifests in gradually declining interest rates (and thus increasing valuations and real estate prices, as I mentioned).

That's what it's like to be in a feedback loop. Your computer may act as a heat source for your home but as long as the thermostat is maintaining temperature at a setpoint, the computer won't change that average temperature -- heat it produces will be offset by reduced furnace activity.

(Of course there are also long-term structurally inflationary forces as well. Environmental damage and resource scarcity, like declining fish stocks for example, are structurally inflationary forces. Demographics can go either way. Short term things like wars, debt crises, earthquakes, or pandemics can obviously have a huge impact, and irresponsible monetary policy can throw things right out the window.)

I had started to reply, but I didn't realize inflation was lower in the 19th century. I will have to mull on that.

> Central banks want to avoid deflation so that debts don't become onerous to repay, which can create a depression.

Agreed. In other words I didn't realize the 19th century was so bad!

I don't personally think low inflation, or even deflation, are so bad, but that idea is at the heart of central bank policy frameworks.

I would prefer other ways than inflation to deal with debt accumulation and miserly unproductively money hoarding, such as demurrage. But the 19th century did not do those, so I thus find myself agreeing with the Central Bank.

Also, since its hard to imagine technology process with out debt in a market economy, we have have an interesting interaction here.

I'm not quite sure in what sense you're using the 19th century as a reference here. As an example of a time of rapid technological change? As a time before central bank inflation targeting?

He means the 20th century. Governments thought they could solve the debt problem with austerity but people simply responded to the lower interest rates by literally keeping their money under the mattress.

This is still true today. Interest is at zero, people keep their money in bank accounts of banks who are so heavily overleveraged vs their demand deposits that they are unable to take the risk of lending.

Money is stuck and has to be replenished by more borrowing i.e. QE asset purchases give banks more reserve assets and let them pretend their balance sheet is clean.

How complex is your automation?

Productivity will be up considerably and variance will be down. But for the most part you now have machine operators hired rather than machinists and those are to different professions entirely, which rightly have different real wages.

Someone who is capable of cleaning a machine and loading materials and tools as well as unloading product is valued different than someone who actually knows how to make any kind of object that is within reason to be fabricated with lathes, mills and other shoptools, which is what a master machinist is. Those people are rare, they always were rare and historically they were valued as much in the past as they are today. Using them to operate a CNC machine is a complete waste of their skills. Typically those people are making one-offs that CNC jockeys would not be able to due to the difficulty of setting things up and/or the need to get it right the first time because the part is irreplaceable.

What you have discovered is known as the 'paradox of productivity'. Productivity improvements are essentially running up the down escalator.

The returns go to cheaper items, because there is less labour time going into each item as productivity improves.

If you redenominate in the 'labour hour' currency, then it all starts to make sense. As does the apparent increase in pay of a concert violin player. Ultimately a concert level violin player doing their thing still takes the same amount of time top produce as it always did, society still likes concerts, and they haven't yet become fond of AI violin players. They exchange those hours for products with less hours in them. So they appear to get a wage increase. Classically this is known as the Baumol effect.

This goes beyond 'real wages', to 'real stuff'. People are definitely getting more stuff than they did 40 years ago. My iPhone is testament to that. As is the lack of power cuts.

Underlying everything is essentially an exchange of labour hours.

As to returns to IP holders, those are often the pension fund, who then pay the pensions to pensioners who then spend it on stuff.

Since we have an ageing population there will definitely be an increase in the transfer to the elderly. It can't be any other way. Is that a fair transfer? Well that's the debate.

> As to returns to IP holders, those are often the pension fund, who then pay the pensions to pensioners who then spend it on stuff.

Pensions aren't distributed evenly across the population, it tends to be that the richer you are, the bigger the pension. That means that IP has the effect of making the rich richer, increasing inequality.

>People are definitely getting more stuff than they did 40 years ago. My iPhone is testament to that. As is the lack of power cuts.

Yeah, we can get much more and much cheaper stuff than before thanks to the 'Made in China' boom, but what we can't get now is affordable real-estate. Too bad we can't go live inside our iPhones. /s

Good comment. To be honest I got puzzled by this as well. My explanation is that it is just artificial based on low interest rates and high competition for real estate. So basically the cost of build stuff maybe remained kind of the same but people have more money and can get higher credits so the prices are dictated by this competition instead of real build costs + it is not always possible to build a lot of new houses in nyc for example.

It's simple really, the things that can't be offshored have gotten more expensive.

Real estate, education, healthcare...

Its not that they have necessarily gotten more expensive, its just that there is a global rate of inflation and a local one, which results in higher cost/standard of living countries being crushed because their local inflation rate will exceed the global one until the country is pulled to the average. So its just as valid to say that the global products continue to get cheaper.

And that was entirely predictable, when a country allows the free flow of capital, and goods but not people then it becomes a race to the bottom in sectors which can move to lower cost areas. There is a certain amount of price suppression for some local goods and the demand dries up, but for things which aren't optional the price will continue to inflate until the system breaks. Infant child care is in this position, the workers competing internationally can't afford to pay for the care of their infants because its local labor intensive.

Basically its broken economic policy and its just a matter of time before it explodes. And its probably not fixable because if the cost of labor ever gets equal, then capital is just going to move to the country that say has the smallest carbon tax/whatever.

Basically a prereq for free trade should be a common government, and free movement of people.

It's not that deep.

During the pandemic the price of used cars started to rise. Did we attribute that to people with more money who can then get higher credit on the now more expensive cars?

No, we attributed it to a lack of supply of cars, and worked to free up the blockages so that the market would return to its correct state where cars depreciated.

It's the same problem with housing. We need to build an awful lot more, and we need to move the work to where we build the houses.

Once we cross the rubicon and house prices start to depreciate, as they should because they wear out, then we'll see a phase shift in the market and vastly more supply. It'll be like the latent heat of condensation.

Housing prices vary inversely with interest rates. Monthly payments are and always have been between 1/4 and 1/3 of income. This is actually up right now due to the bubble which cant last.

More people are paying more for rent than they used to.

In 2015, of all renter households:

* 38% were rent burdened (defined as spending 30% or more of pretax income on rent), an increase of about 19% from 2001.

* 17% were severely rent burdened (spending 50% or more), an increase of 42% from 2001 (the 38% rent burdened includes the 17% severely burdened).

For households headed by someone 65+, 50% are rent burdened, ~20% severely rent burdened.

https://www.pewtrusts.org/-/media/assets/2018/04/rent-burden...

Interest rates are lowest in the most developed countries.

The only exception is the US because it has a structural deficit.

It's actually the opposite. Countries have tried to artificially raise interest rates above liquidiry preference via inflation targeting and stimulus for the last 40 years.

>but what we can't get now is affordable real-estate

Median housing in the US has been around inflation adjusted $100/sq ft for decades.

Houses today are massively larger than in the past, making the costs look higher. And what you get today is also a lot better - more efficient, cheaper to heat and cool, safer, better electrical, water, etc.

burn down much faster... https://idahonews.com/news/local/how-fast-can-new-houses-bur...

So my options are 3k a month for a walk in closet in SF or 3k a month for a mansion in Reno

Maybe we need more housing options

Maybe people need to move where housing is cheap if they want cheap housing.

>Median housing in the US has been around inflation adjusted $100/sq ft for decades.

I was talking about the European real-estate market. I don't know the US market but I heard it's much more affordable than here, at least outside of SF/NY/etc.

European costs vary by well over a factor of 10 across countries, especially urban to rural.

I don't see evidence across all of Europe any data for much difference than in the US. Dense places cost more, sparse places are amazingly cheap. Poorer countries, like poor states, have even cheaper housing.

You can't import land from china, that is a problem that your local community is responsible for. Better get used to being screwed by your fellow towns people.

Interestingly, this is basically the marxist analysis of the economy, in terms of surplus value.

The rosy side of the equation happens when we look at the price of useful products that can be automated, like fridges or computers, where labourers are much richer in such products than they were 100 years ago. However, the ugly side is visible when looking at the price of products that can't be automated, like land, where labourers are much poorer then they were 100 years ago. The difference is often made up of useless products that could be automated, like cheap fashion or fast food, which marketing keeps pushing as hard as possible.

Not quite. It avoids the thorny question of "what is value" that Marxists disappear up their own fundament with.

Here the Labour hour is valued only in terms of its exchange rate to the denomination that the area pays tax in.

Yep. Nearly everyone but equity owners is either hurting or not going anywhere while delivering more for the same or less. Graph after graph Robert Reich has put out there shows the dismalness of the situation. Another infographic video about the distribution of wealth in reality vs. perceived reality vs. desired: https://youtu.be/QPKKQnijnsM

Somehow there is also a systematic economic double-standard that works for some but not others: "free markets" of dog-eat-dog inputs (i.e., suppliers), but less free markets on outputs (i.e., monopolies, mega corporations, patents, proprietary-ness, closed-source). Unregulated greed fixes all problems by the "invisible hand", right?

Most of the benefits of that sort of automation are actually going to customers. Aggregated across all such automation in an economy, that means we all benefit.

I wonder if there have been any serious studies to determine whether there is truth behind Marx’s claims of an automation arms race leading to a declining rate of profit.

Profit must go down to zero because of structural reasons.

Profit isn't a reward, it is a reallocation of capital to a new sector. At some point growth stops and there are no new sectors meaning there is nothing to reallocate existing capital to.

> These types of automation have been in place for decades, and all of them cut costs for the owners of them and reduce the number of workers needed.

They reduce the number of workers needed to produce the same amount of output. What is slightly counterintuitive is that having done so, they may actually increase the demand for workers, because they now can produce far more output in the same amount of time.

Modern languages and tools have made programming much more productive, and now there are lots more programmers and they get paid more.

Theoretically yes, but in practice the given industry as a whole often does not need much more output even if the per unit cost of the output declines. Like, you don't need significantly more accounting just because you can use spreadsheets now.

And not all automation increases worker productivity, lots of it just eliminates workers from certain jobs altogether, e.g. self driving cars / trucks.

This can indeed cut both ways. Instead of needing huge numbers of manual workers to harvest the wheat, one guy can sit in an air-conditioned combine harvester cab for a few hours listening to tunes, and we get the same food. Humans can (and do) eat slightly more food given it's cheap and readily available, but there's a limit. So, the percentage of workers engaged in agricultural work fell.

On the other hand, turns out that once a shirt is cheap humans will buy not just one extra so there's one to wear and one to wash, but a dozen, or hundreds. And they'll throw it away when it's a little worn not when it falls apart. So while industrialisation of fabric and clothing production made clothes cheap, the result wasn't that fewer people were employed in that work but that we just all bought more clothes.

> one guy can sit in an air-conditioned combine harvester cab for a few hours listening to tunes,

I hope he's not listening to tunes. A change in the sound it's making is often the first warning that your $500,000 piece of machinery is about to self destruct.

I see it all the time, from combines to snowplows.

> A change in the sound it's making is often the first warning that your $500,000 piece of machinery is about to self destruct.

Whose machine?

Anything that's got risk of becoming expensive will give you tactile or olfactory feedback.

More likely some warning light will illuminate on the dash first.

I’m not very familiar with accountancy, but there seems to have been a lot of growth in internet-based tax prep for individuals who would previously have done their own. I couldn’t find historical data, but forecasts I found all agreed the number of accountants needed would rise over the coming years.

We don’t even have self-driving cars yet, how can you know what effect they have?

Well, yes you do, because the spreadsheets make it easier for the regulators to add more regulation.

Think: there used to be a theoretical limit to how much regulation could exist, because the regs had to be known by a person (otherwise how are they enforced?).

Now we have regtech. A whole industry built on automating compliance.

How long before each individual needs a regtech software program (or two!) just to function in society? I'm not talking about tax filing, that's just the start.

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

> What is slightly counterintuitive is that having done so, they may actually increase the demand for workers, because they now can produce far more output in the same amount of time.

They may do that.

Too many people take it for a guarantee that technology by itself must do that, and can never result in a situation that a human hand is needed to steer us away from.

> Modern languages and tools have made programming much more productive, and now there are lots more programmers and they get paid more.

Well not really. Computer industry pay is increasingly becoming bi-modal at least for a decade. The highly paid programmer is just a US phenomenon and that too in few major IT hubs.

There are lot of programers/IT admins who would make pretty decent money in late 90s to 2000s. And they make less than they were 20 years ago and yes all hadoop/big data training did not really make any better paid.

Inflation adjusted IT salaries have gone down in lot of places in the world. Of course 1 percenters of IT in FAANG+ make it sound like IT salaries are going up heavily for everyone.

> The highly paid programmer is just a US phenomenon and that too in few major IT hubs.

That's simply not true. Programmers around the EU are among the best paid professionals in their local markets, even more so for young people (out of college/school/university). The salaries are far away from Silicon Valley levels, but are still very high locally.

Think this depends on market. Sure, salaries in Germany are decent to good, but Italy? Ehhhhh.

According to some sources online, the average salary for a junior developer in Italy is around 24k[0]. Indeed that isn't a lot, however it's not that much below the average salary in most regions [1]. Furthermore, in a per type of job, IT and tech are the highest paid bar "management" and "top management"[2].

0 - https://www.payscale.com/research/IT/Job=Junior_Software_Eng...

1 - https://www.statista.com/statistics/708972/average-annual-no...

2 - https://www.paylab.com/it/salaryinfo

I guess it depends on what you define as 'highly paid', which is what you were originally discussing. More senior developers making 35-40k EUR doesn't sound 'highly paid' to me, even though yes, the local market isn't as strong. High-er, certainly, but it's not far above the GDP per capita for Italy as a whole.

The thing about management is what I've heard Italian devs complain about (and it's a common complaint is tech-weaker countries), that to get decent money you have to transition to management, whereas countries with stronger tech sectors like the US have more companies that let engineers continue to grow as engineers (or architects).

> Programmers around the EU are among the best paid professionals in their local markets, even more so for young people (out of college/school/university). The salaries are far away from Silicon Valley levels, but are still very high locally.

Was my original comment and point. Salaries for tech are very good for the local markets, even if they might be considered low compared to other markets.

You said they're "very high" locally, which doesn't appear to be the case. "Slightly above GDP per capita" just isn't very high. If everyone in one country makes 40k EUR except for one dude who makes 40,001 EUR, he has the highest income in the country, but his income still isn't very high at all.

Of course, part of this may be that the wage spectrum as a whole is compressed compared to the US, but that's an explanation, not a contradiction.

> "Slightly above GDP per capita" just isn't very high.

Yes, it is in any capitalist economy, because income distribution isn't normal. In the US, for instance, per capita GDP is about $64k, but median personal income is a little over $35k, and median pay of a full-time worker income, annualized, is just under $45k. An individual worker making just over the per capita GDP is doing quite well.

In the US, the median software dev comp is more like 2x GDP per capita though. The BLS says 110k for devs + QA testers median salary, and IIRC that doesn’t factor in stock.

I'd say 40k EUR isn't high for someone living in Milano (Italy's main tech hub), for example, it's decent, but not high. The cost of housing is what gets you.

> that to get decent money you have to transition to management

This is the main Problem. Technical excellence is not valued and rewarded enough. The best Engineers get promoted to Mediocre managers because that's the only way to move forward. The result is an economy focused on micromanaging for small gains instead of innovation for big gains.

Yeah, and that's why lots of IT jobs are outsourced to Eastern Europe; why pay 70k€ per year for people doing devops in the Nordics, Germany or the Benelux countries, when you can get a Romanian to do it for 20k€?.

The situation in the EU is the same as in the US, and outside the local hubs, wages are lower.

I'd say it depends on the EU markets. I have the feeling that IT professionals from Eastern Europe (I'm from there) are a lot better paid, comparatively speaking, to IT professionals from many other Western countries (with the focus on "comparatively").

I've heard from a lot of my colleagues that they couldn't afford to have the same lifestyle in an Western capital city with equivalent IT salaries. I know I wouldn't be able to rent in downtown Amsterdam, Paris or Milano with my equivalent IT earnings were I to live in those cities the same way I now do in Bucharest.

I wouldn't be able to rent - or buy - in Amsterdam either even though I was born there and lived there for half my life. The prices there are simply insane. It's ok with me because I wouldn't want to either but between AirBNB, businesses using residential real estate as fancy offices (half the inner circle of the canals is like that now), the crime and the tourists you probably would have to pay me to live in Amsterdam.

Having visited Amsterdam, but mostly the tourist bits, I'm curious: where do reasonably paid tech professionals live in that city? Can you give me some examples of areas to look up?

Amstelveen, de Pijp, Overtoom and surroundings, the nicer parts of East and West.

>I'd say it depends on the EU markets. I have the feeling that IT professionals from Eastern Europe (I'm from there) are a lot better paid, comparatively speaking, to IT professionals from many other Western countries (with the focus on "comparatively").

It's true. Unless you work for a FAANG, or some Fin-Tech or a unicorn flush with cash, dev wages in Western EU are not higher than Eastern EU anymore.

I know devs in Poland and Romania making way more than me and some friends do in Austria/France/Belgium.

It's because of huge amount of outsourcing from US companies moving directly to Eastern EU, pushing the wages higher than what Western EU based companies can offer locally, coupled with the higher taxes in the Western countries, meaning devs in the West taking home lower wages than their counterparts in the East.

Someone here said it well, but I don't have the original post to quote anymore: {a cleaner in Sweden makes way more than a cleaner in Poland, but a dev in Poland can end up making as much or more than a dev in Sweden}. That's whack but that's the reality of supply and demand caused by the bleed from the US tech boom to certain regions.

> Programmers around the EU are among the best paid professionals in their local markets,

That's not really true. I live in Austria and many more skilled professions earn better than most devs here as the local SW industry sucks, almost no product-based companies and a market overabundant with talent.

I'm sure in hot markets with plenty of product based companies like London, Berlin, Amsterdam or Eastern Europe the situation is better.

> Computer industry pay is increasingly becoming bi-modal at least for a decade.

Source?

I really doubt this. Programming doesn’t have the kind of cartel dynamics mixed with prestige obsession that resulted in dualists labor markets. It doesn’t have licensing based on credentials that have no effect on how well you do your job like K12 teaching, nor is it remotely as prestige obsessed as law firms.

Any evidence?

This only works to an extent. US manufacturing output is higher than before outsourcing manufacturing became a thing, but US manufacturing jobs are far fewer, because the total demand for a whole lot of products can be met with far fewer people.

When there's an inequality in productivity, more productive firma will see an increase in labour, bit that doesn't mean the whole sector does.

In theory people work because they want something from someone else (says law). So in theory there can never be unemployment.

In practice some people want money itself, which is an empirical contradiction of says law. There are also people who hoard jobs because they like them, which is bad for people who want the money to actually spend it.

Blue collar work vs. white collar work is actually the issue.

The assumption is that blue collar work is the stuff that gets automated away. The reality is that blue collar workers are getting more productive instead but are still essential. And we already have vastly less of them than we used to due to the industrial revolution and the wave of automation in the last centuries. Meanwhile a lot of white collar work is melting away or increasingly of the "bullshit job" variety where the added value is dubious. So, a strange side effect of automation is that we've created a lot of bullshit jobs instead of creating mass unemployment.

An even stranger side effect is that a lot of people with college degrees are now doing artisanal stuff instead of committing to a life long of a meaning-less and soul crushing office dwelling existence. So, they're choosing to become blue collar workers. And artisanal here of course means "choosing to do things by hand" instead of maximizing quantity and revenue through industrialized processes. Artisanal is more fun to do and quality of the goods means they are worth more. You can buy bread from a factory, or buy a sour dough bread from a hipster with a college degree. Lack of automation is the whole point there.

What the lockdown last year showed quite convincingly is that a lot of white collar jobs are not really that essential and it doesn't matter that much if productivity there drops by few tens of a percent. The economy suffered a little but mostly that's because of reduced consumption and not of reduced production. And even that is hardly proportional to the productivity loss of masses of people suddenly no longer doing what they were supposed to be doing or doing a lot less of it. A lot of those people never got back to their former level of "productivity" and it continues to not matter economically. Their most important job is spending the money they earn. The labor performed earning that money is secondary.

It's a great argument for basic income: give people money and get them to spend it in order to boost the economy. It's a lot more effective than funneling it to the offshore accounts of the 1 per-centers. They hoard rather than spend their money.

And it frees people up to do something more valuable than stare at a cubicle wall doing something that absolutely and provably does not matter economically. Like bake bread, create art, or do something else that people might spend on.

In many organizations, the lockdown showed productivity went up as white collar people worked from home. Some even worked longer hours of their own accord.

> The benefits of those efficiency improvements are already going to the owners of the capital, not to workers, and this has been going on for decades.

Per your source, the "productivity-pay gap" largely explainable through increases in pay inequality while the change in labor's share of income is a much smaller factor (and it's worth mentioning EPI is probably overstating the size of the gap in the first place[1]).

As an alternative to your hypothesis, my guess is that workers in jobs amenable to automation are reaping huge increases in pay (e.g. engineers) while jobs that are more difficult to automate are falling behind.

> We already need to find ways to distribute the benefits of these efficiency improvements more equally.

I wholeheartedly agree. I propose we find ways to automate low productivity jobs so all workers can enjoy the associated benefits.

[1] https://fee.org/articles/the-myth-of-the-pay-productivity-ga...

What you're describing was outlined pretty clearly in The New Ruthless Economy[1] in 2005. A very interesting read, given when it was publshed.

[1] https://global.oup.com/academic/product/the-new-ruthless-eco...

Interesting, thanks! Will definitely give it a read.

the term "robotics" is beginning to expand in some ways though. "robotic process automation" refers to tools and techniques to automate a lot of the data-entry and human-facing software components of industrial workflows. so while hard robotics tend to dominate the media, it's important to remember that automation is making gains in many areas.

> The benefits of those efficiency improvements are already going to the owners of the capital, not to workers, and this has been going on for decades. If in 1980 you were able to replace 4 clerks handwriting transactions in an accounting ledger with one clerk typing them into a spreadsheet but pay that one clerk marginally more than a handwriting clerk, who benefits most in that equation?

The effect of productivity improvements on the real incomes of workers ultimately depend on how the change affects productivity of the marginal worker. You need to look at more than a single firm to figure this out. If the availability of spreadsheets makes clerical workers so much more productive that overall demand for their services increases more than four-fold (which might be definitely possible), the marginal spreadsheet-tweaking activity will be paid quite a lot, and workers' income will increase.

That doesn’t match real world outcomes though. Quantity demanded goes down. Wages go down.

Mostly because automation isn’t making that person more valuable, it’s just removing it from that persons role. Skilled labor becomes unskilled labor. Supply goes way up. Clerks are a great example.

They used to do a lot for a store and had to know all the products. Now a clerk is just a last mile button presser. They don’t add much value at all. They’re not doing 4x the work. They’re doing 0.1x the work. Trying to say “but the value couldn’t be captured without that guy” is wishful thinking.

> They used to do a lot for a store and had to know all the products. Now a clerk is just a last mile button presser. They don’t add much value at all. They’re not doing 4x the work. They’re doing 0.1x the work. Trying to say “but the value couldn’t be captured without that guy” is wishful thinking.

Not to detract from your overall point, which is absolutely correct, but having been a grocery clerk there's a slightly different phenomenon in play there. Having a clerk who knows the products well is a big help, but you really only need one available at a time. So rather than make acquiring that knowledge an actual part of the job, they rely on scale to have enough workers who learn it through their own initiative.

> Skilled labor becomes unskilled labor.

Which is bad for skilled labor (wrt. that set of skills) but good for unskilled labor.

It's not good for unskilled labor because they get paid even less than unskilled labor as prices rise and wages stagnate.

> who benefits most in that equation?

That depends on how the extra profit is distributed, aka taxation and social welfare. Slaving away 40 hours a week on a bullshit job that can be trivially automated is not a benefit per se.

> Slaving away 40 hours a week on a bullshit job that can be trivially automated is not a benefit per se.

For some workers, that's a huge benefit. If they get ousted they have to find another degenerate position or work for a lower salary. Usually it's both.

I mean that money is the benefit, not the work itself. Our wealth distribution system is not suitable for significant levels of automation at all - that is the root of the problem, not automation itself.

People measure prosperity by employment, not the work they do, which is how we get backwards economics where it is better to destroy the planet.

Most things that are perceived as repetitive and therefore something you can automate aren't likely to be easily automated. The reason is that many repetitive tasks have contingencies that have to be mapped out. Even some of these aren't covered by employee training or corporate policies and so it depends on the judgment of the individual employees which in the end destroys the fantasy of robots everywhere and programs to replace paralegals and accountants.

> What about the spreadsheets, the computers, algorithms, industrial control systems, Python scripts, voice recognition systems, data pipelines, software powered gig economy, algorithmic traders?

At first I thought you were bringing these up to point out that the robots create other jobs, because robots don't just spring from the earth fully formed and ready to run forever. They require an army of engineers to build/operate/maintain them.

> We're always focused on the some day maybe robots, but automation has been changing society and (I'd argue) increasing inequality for decades

Inequality is a political problem. I'd rather we tackled that for what it is: bring on the fully automated luxury communism and free us from our mindless toil that can be trivially performed by mindless robots. We have nothing to lose but our drudgery.

For your first point: the new jobs required is of course much less than the jobs removed, or else there wouldn't be a point to automation in the first place because it's not saving on labor efficiency.

Agreed on your second point. Attack the inequality, not the automation.

The totalitarian state that is necessitated by the communist political system always turns out to be as hellish, selfish, and mismanaged as any late capitalist dystopia—except usually with more murder. So I would say the political problem is very much not solved and post-scarcity Utopianism isn’t going to get us there.

> The totalitarian state that is necessitated by the communist political system always turns out to be as hellish, selfish, and mismanaged as any late capitalist dystopia

Leninism (and it's descendants) have a totalitarian state system (socialist, including non-Leninist Marxist Communist, critics tend to call it “state capitalism”) that replaces private capitalism as a hack to avoid the requirement of developing proletarian class identity in a developed capitalist system present in OG Marxism.

That this turns into a dystopian nightmare that mirrors and even exaggerates the horrors of private capitalism and fails to actually bypass private capitalism is a surprise to no one but Leninists, least of all non-Leninist Communists.

Leninism, Marxism was founded at a time when the labor equation was very different. Back then, generally there was a fixed ratio between labor input and productivity output. So communist ideals were about how to centrally distribute that productivity output while centrally controlling the inputs.

With automation though I think that obsoletes a lot of those ideas. Now there's less and less connection between labor input and productivity output.

We need some new ideas for distributing the fruits of automation - that take into account the productivity improvements thanks to machines, computers and efficiency improvement in general. While at the same time keeping some incentives to make things even more efficient.

I think that can be done within our current system with taxes, socialising things like healthcare and access to food and healthcare (and perhaps something like UBI). No need to completely overhaul the political system, although we could be heading towards that too if things don't change.

Marx was writing precisely at the time that the equation shifted to its modern form. A single spinning jenny operator could do the work of eight or more spinners.

Marx spent the first chapter of the Communist Manifesto fanboying over how capitalism was increasing the productivity of labour to previously unthinkable levels [1].

He explicitly argued that this was what he believed would make communism possible.

So whatever else one thinks about Marx it's clear he did not think in terms of a fixed ratio.

The comparison of Lenin and Marx is also flawed - Marx argued for the dismantling of the state, not a highly centralised state.

In Critique of the Gotha Programme he even pointed to the US as a better model for education (arguing that government-set standards but devolution of control over the actual provision of education was something to aim for) than the SPDs proposed state run education because he disapproved of giving the state control over even that, arguing that it was rather the state which needed to be educated by the public.

And in his work on the Paris Commune he lamented that one of the biggest errors he believed the commune made was to not go further and smash the rest of the state. The focus on centralised planning in the Soviet Union was Stalinist, not even Leninist, and certainly has nothing to do with Marx.

[1] E.g.: "The bourgeoisie, during its rule of scarce one hundred years, has created more massive and more colossal productive forces than have all preceding generations together. Subjection of Nature’s forces to man, machinery, application of chemistry to industry and agriculture, steam-navigation, railways, electric telegraphs, clearing of whole continents for cultivation, canalisation of rivers, whole populations conjured out of the ground — what earlier century had even a presentiment that such productive forces slumbered in the lap of social labour?"

> The comparison of Lenin and Marx is also flawed - Marx argued for the dismantling of the state, not a highly centralised state.

Marx still believed in some sort of state-level synchronisation or command and control - otherwise he knew that you would just have lots of little communes / barracks-communists inefficiently working at cross purposes, e.g. with resource utilisation. But he never spelled out in detail how the central planning function should work - it was in effect an implementation detail. It was in this space of undefined behaviour that Lenin, Stalin and Mao created the totalitarian states that they did.

I think a lot of people forget that there was no actual collaboration between Marx and Lenin and the others. Lenin was only 13 years old when Marx died, ill and in a city on the other side of Europe.

Although there is a mass of academic detail in Marx's works, the communist manifesto itself is very thin in terms of the remedies to the problems it identifies. The 'call to action' part of the manifesto is essentially ten short bullet points, one of which is 'etc etc'.

> I think a lot of people forget that there was no actual collaboration between Marx and Lenin and the others.

Well, I mean, people “forget” because equating Leninism with Marxism served the purposes of both Leninists and Western opponents of Marxism, so it became a common element of propaganda of both the USSR and other authoritarian communist regimes and the capitalist west.

Marx argued the state inherently would cease to serve a purpose and would "wither away" (the term itself is from Engels who argued it was based on Marx writings) once class antagonisms no longer exist, so I don't think there's much basis for the idea he believed in "command and control" other than at most in a transition. Ironically given his authoritarianism, Lenins The State and Revolution is one of the best summaries of Marx' criticism of the state, and even Lenin plaid lip-service to the notion of the withering away of the state to the point of arguing for an eventual society without a state.

The core distinction between Marxists and many forms of Anarchists, as early as during the dispute in the First International between Bukharin and Marx, was their view not on the long term dismantling of the state - where they agreed - but the role of the state in the transition from capitalism to socialism, where the anarchists wanted the state dismantled immediately, while Marx argued the state could serve a role in a transition, itself becoming a tool to execute the changes necessary to change society.

In his works on the Paris Commune he explicitly criticised the Communards for failing to take the next step and destroy the bourgeois state. While that could certainly have been replaced with a proposed federation, a key distinction was that he celebrated the bottom up delegation of authority, which is hardly compatible with "command and control" if any part can withdraw.

As such, libertarian Marxism directly opposed to the very existence of a state became a thing, based in large part on his writings on the Paris Commune.

> Leninism (and it's descendants) have a totalitarian state system (socialist, including non-Leninist Marxist Communist, critics tend to call it “state capitalism”)

_Lenin_ called it "state capitalism." That was the official policy of the USSR under Lenin. State capitalism.

> _Lenin_ called it "state capitalism.”

Lenin called a transitional phase that was intended to take 6 months to establish and another 6 months to phase out into socialism “state capitalism” (it may be the case that “war communism" of 1918-1921 was officially recognized as something of an extension of that phase necessitated by emergency), but state capitalism was never officially the long-term policy or the self-description of the state of the economy past that point; the criticism from non-Leninist Marxists is that what the steady state of what the USSR became was state capitalism, not socialism.

Yes it was supposed to be temporary, but the policy still remained official until after his death.

(Did he really say just 6 months?? That doesn't sound right. But anyway it would have only been a prediction -- which failed.)

"Joseph Stalin abandoned the NEP in 1928 with the Great Break."

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

https://en.wikipedia.org/wiki/Great_Break_(USSR)

The NEP wasn't (in its own theory) state capitalism but a transition from war communism (which I think was, even in its own theory, still state capitalism) but to a mix of “socialized” industry and state-guided private capitalism. (Yes , critics would say that the “socialized” industries were really state capitalist and that the degree of centralized “guidance” made the private sector also state capitalism, but we’re talking about what Lenin and the internal theory of the USSR portrayed it as, where it was a different temporary stepping stone to fully socialized industry.)

> The NEP wasn't (in its own theory) state capitalism

That's not what wikipedia says. I can't argue as I am no expert in such matters myself having read almost none of the primary material (but I did read "The Tax In Kind"... which I was assuming outlined the theory of the NEP).

Wikipedia on NEP:

Lenin considered the NEP as a strategic retreat from socialism. He believed it was capitalism, but justified it by insisting that it was a different type of capitalism, "state capitalism", the last stage of capitalism before socialism evolved.[15]

I don't think you're projecting as far as OP.

F U L L Y is Star Trek years away, not next Wednesday.

> They require an army of engineers to build/operate/maintain them.

Yep agreed, but there's a key difference. Unlike most other jobs, your labor continues to provide the company profits even after you leave. I think that's why software engineers are typically paid a lot and are often paid with a stock option component.

> Inequality is a political problem. I'd rather we tackled that for what it is: bring on the fully automated luxury communism and free us from our mindless toil that can be trivially performed by mindless robots. We have nothing to lose but our drudgery.

I think we could already do this. Obviously many workers would still be required, but they could be much more fairly compensated. Someone else mentioned Graeber's Bullshit Jobs, we're already inventing jobs just to keep people busy and spending. As Adam Curtis said[1], "your real job is shopping".

[1] https://www.facebook.com/watch/?v=1414301661936421

> Unlike most other jobs, your labor continues to provide the company profits even after you leave.

That's certainly the engineer's sales pitch, but let me introduce you to the engineer's fine print, tech debt.

> Unlike most other jobs, your labor continues to provide the company profits even after you leave.

Maybe in terms of pure job function, but anyone who's ever solved a problem at their job in a way that is still used when they leave has continued to provide the company profits.

>If in 1980 you were able to replace 4 clerks handwriting transactions in an accounting ledger with one clerk typing them into a spreadsheet but pay that one clerk marginally more than a handwriting clerk, who benefits most in that equation?

In a market economy competition will squeeze down those newfound profits. The owner of the company benefits, the employee benefits, and all the customers benefit.

The only possible outcome is automation of most if not literally all jobs. At this point, capitalism will necessarily break down because it depends on continuous consumption and there will be no disposable income with which to consume. Hopefully society will evolve into post scarcity at this point instead of a cyberpunk dystopia where gigacorporations create artificial scarcity economies just to maintain the status quo.

I'm hopeful for a solarpunk future.

https://i.redd.it/k51q3zw8hn981.jpg

Frankly I think we're already at the tipping point for that. If I think about it, how many jobs are actually involved in:

  * Providing food
  * Providing shelter
  * Providing healthcare

You could argue just about everything else is really a form of artifical scarcity, none of it is needed for survival, perhaps just for comfort.

That divide became a lot more clear in the past 2 years with the acknowledgement of "essential workers". Without Google, Facebook, or auto insurance, we'd be inconvenienced. Without essential workers society as we know it would collapse. And almost all of those workers are paid less than an entry level Facebook engineer.

So now the question is, are we already post-scarcity? And perhaps we don't know it only because it hasn't been distributed equally.

Are we already post-scarcity?

Thanks to industrialization, in the first and second world yes for basic food, clothing and shelter, as well as other mass-consumed products.

Things that remain (increasingly) scarce in my view - globally - are broadly educated people, political/economic/behavioral freedoms, unspoiled nature and the temporal wealth and whim to access it regularly.

This is an important point that gets overlooked. It's likely that we already reached something like "the singularity," but most of the extra productive output didn't go towards finding ways to increase productivity, it went towards things without a positive impact and often a negative impact on society (advertising, junk food, conspicuous consumption, regulatory capture, sales, etc.).

Completely blind to it? Not likely. Marx was writing about it in the 19th century. Computing based automation and management is just the accelerated evolution of the automation techniques from the industrial revolution.

It's complicated.

Is it complicated?

Greater efficiency is good, because it makes society "richer".

If that societal wealth increase doesn't go to the average person, then why should they care about efficiency?

The luddites didn't hate efficiency, or machines, they sabotaged the machines because that was one of the few ways they could hurt the interests of the people who were mistreating them. Specific machines were targetted because of who owned them, and the way they were treating workers.

For wider context this time the royal families of Europe were waging war to prevent democracy spreading. A great deal of this funding came from the UK, yet at the same time their poor people were rioting due to lack of food, and tax increases to fund the war. Folk songs at the time tell tales of people signing up to fight the French because the other option was starvation.

It's convenient that we remember these people as "silly folk who were scared of machines" rather than "downtrodden masses revolting against a cruel dictatorship waging war against democracy".

I'm sympathetic, for sure. I say complicated because it's not a straightforward linear causal relationship or one where we can just say "bad". Automation causes the rate of profit to rise for a period. And for some segment of the working population this can lead to increased prosperity. For others less so, as you point out.

But in the long run, competition between businesses leads to at least a short term equilibrium, where automation and labour exploitation rates are roughly equal, or monopolization has occurred. And then rates of profits can plummet and then it's really bad all around.

The point here is that I don't think it's automation itself that creates the unequal distribution of incomes. It's just the market economy. The market economy creates automation in order to gain competitive advantage.

Curtailing automation isn't going to improve wealth inequality. Might even make it worse.

Equalization of wealth requires the intervention of the working population either through unionization or forcing the state to intervene.

Robots don't matter much to the worker productivity of industrialized nations because manufacturing is mostly outsourced to less industrialized ones. Robotic automation matters to less industrialized nations.

Deep learning will eventually eat 99% of office jobs, including software and hardware development. Self-programming and self-designing systems seem as inevitable as killer robots. These automation technologies can lead to a rapid collapse of standards of living in industrialized nations that most people "won't see coming."

As a consequence of massive efficiencies, regulatory capture, and political influence, the long-term vicious cycle trend is towards most of humanity becoming borderline homeless with a small middle class and a tiny strata of insanely rich royalty.

There is an obvious solution to spiraling inequality "pure" capitalists refuse to consider on "religious" grounds: take past, present, and future absurd gains from the billionaires and redistribute them according to a needs-based formula.

I don't believe machine learning will ever reach the level of sophistication that you're thinking. The reality is that we can't even make 5th generation (goal oriented) programming languages yet. Mostly because we keep chasing the unicorn of generalized AI rather than building the foundations of useful programming standards. Like imagine using optimization the way we do with loops and branching to handle more complex strategy based problems. We don't even have that level of sophistication in our compilers yet. We're in what amounts to as the Neolithic Age of programming.

> yet.

You appear to be denying the progress made over the past 30 years by deep learning, ML frameworks, constraint solvers, and immense computing power.

Prolog has been around for 50 years. I suggest you look at what it can do.

Programming languages exist only for humans to specify and discuss intended computing behavior. There are plenty of goal-solving algorithms and heuristics already. Turing completeness is all that's necessary. "Flying car" programming languages are unnecessary because they offer no additional intrinsic power, only convenience of expressiveness that could be provided by libraries.

Self-programming and -designing systems, if left unconstrained, will develop their own IRs, design, and manufacturing specification protocols that would likely become rapidly incomprehensible. They won't need anthropocentric programming languages after several iterations because it would be an inefficiency. That's what the technological singularity will look like.

"It is difficult to get a man to understand something when his salary depends upon his not understanding it." - Upton Sinclair

> You appear to be denying the progress made over the past 30 years by deep learning, ML frameworks, constraint solvers, and immense computing power.

Most of the progress in the last 30 years was immense computing power, almost all foundations for todays ML are revised old concepts. What you propose is AGI, how you want to achieve that? We don't even know where to start in theory, this is not my opinion but current top names in ML world[1], which was discusses on HN many times.

1. https://venturebeat.com/2018/12/17/geoffrey-hinton-and-demis...

>Turing completeness is all that's necessary.

As a programmer and someone who's earned a computer science degree, I'm going to say you're wrong for many reasons but I'll point out a couple here. First, the problem with modern programming is the problem that all computing has struggled with: how do you define meaning (semantics). We've resolved almost all the issues of representation of syntactic reasoning (the ability to encode and process symbols) but to tell a computer the general meaning of a program, its limits, its problem space, and even potential issues that it will need to form strategies to resolve are the realm of pure research right now. And have been for decades, it's why most research has been focusing on things like machine learning, genetic algorithms, and neural networks because these can be trained on a specific set of problems with the hopes that given sufficient time that we can augment this via hardware improvements (it's why you see Google touting android phones with neural network processors and the like). The problem still remains that we can't just do a TNG science scene where you speak out constraints to a computer and it generates tentative results on those constraints.

Second, we have another problem that still persists to this day: how to get firms to fund such research. In the past, the military industrial complex (which still does to an extent) and monopolies would fund such research to keep ahead but in this day it seems most folks aren't keen to fund what won't produce profit within a few quarters and as such most research has been pared down to limited scopes. I believe this is an issue due to how much pure research in the past seventy years has resulted in the majority of gains seen in the private sector. Thus, the private sector assumed all progress was natural and not inevitable build up from decades, even centuries, of hard work. And without that pure research being allowed to exist and even fail in its pursuits, we've put ourselves at odds with the intergenerational aspect of all scientific research. Meaning, we're most likely only be able to achieve such gains many centuries from now.

>That's what the technological singularity will look like.

There won't be any technological singularity because in the past we never had one. People who think this are fools like Ray Kurzweil. Anyone with an inkling on the subjects of anthropology will know better. The advances of the human species before recorded history were building on the small steps of pre-human ancestors. There wasn't a magic Eureka moment for us and never will be.

Reply on news.ycombinator.com