The great decoupling of labor and capital | Hacker News Reader

The great decoupling of labor and capital

mbi-deepdives.com

85 pointswalterbell64 comments

Charts into the future with an exponential curve. Right.

Look, I could have believed linear at least to some asymptote. And then you plot the exponential to the asymptote after the fact, because curves are nicer.

But at this point, exponential growth into the future that far out is 'busboy giving investment advice' exuberance.

Plot what % of GPP and PPP this represents, and then plot the overhang of debt, and the cycle time to a market correction, and tell me you still think an exponential growth story is going to happen "up and to the right" in this.

And the headcount charts are “incremental”, i.e. first differences rather than absolute values – hardly like-for-like.

What do you mean by 'overhang of debt'?

Debt and equity are equally valid ways to finance projects. See https://en.wikipedia.org/wiki/Modigliani%E2%80%93Miller_theo... for the spherical cow version, but this holds approximately under real world conditions, too.

In general, I share you skepticism and agree that exponential extrapolation should be done only very carefully at best.

Your link itself says that in presence of taxes and tax deductible interest payments, i.e every single jurisdiction you care about, your firm can be significantly better off with debt. Saying debt and equity are equally valid suggests that you should be fine with a 100% equity financed company, which is completely false.

> Saying debt and equity are equally valid suggests that you should be fine with a 100% equity financed company, which is completely false.

I assume you mean 100% debt financed?

In any case, that's also mostly fine and happens in practice. As soon as that dips above 100%, conceptually your creditors turn into shareholders.

No, I meant equity. If you are not using significant (20%+ of capital at a minimum)debt in your capital structure, you are almost guaranteed to be making a financial mistake.

It depends on your industry and stage of your company and jurisdiction. Many software startups run with 100% equity, and I don't think they are all idiots.

A pet peeve of mine is people complaining about too much leverage, but then not doing anything about the tax systems making debt financing cheaper compared to equity.

P.S. Some companies even have more than 100% equity. In a sense: when their market capitalisation exceeds their enterprise value.

"This time it's different"

I dunno, people really are living in tent cities and Hondas in Canada. The gains were permanent.

Linear growth is nonsensical. Growth is driven ultimately by consumption. Consumption is driven by population. The population either grows or declines in an exponential fashion. That's how this works.

Consumption is not driven by population alone, it’s driven by population with a spending power. And from what I’m seeing, that’s shrinking. The article talks about how tech companies are growing revenues with fewer and fewer employees. An important question to ask there is: is that a good thing?

Then the shrinking is exponential as well. That's the point. Population grows or shrinks according to an exponential curve. If the exponent is one, then things are stable. If it's less than one, then things will experience exponential decline, which is different than linear decline since it never reaches 0. If it's more than one, then things experience exponential growth.

Most problems in life come down to misunderstanding exponents.

Of course linear growth is nonsensical. But if you are plotting into the future its a damn sight less nonsensical than exponential unless you have very strong foundations, and for the 1-2-3 year projection it's pretty much ok. Anyone who thinks they know trends 5+ years out is a loony.

And as I said, thats ignoring market cycles, the debt overhang in the economy, all kinds of problems.

Linear and exponential growth look the same in short term horizons depending on the exponent. However, exponential growth is strictly superior as some things can grow very fast if the exponent is high enough. We saw that with covid. The exponential assumption is correct. For the situations you are discussing, where the exponent is low, the exponential pattern correctly captures that.

Basically true, but there are other potential sources of growth :

- using technology to unlock cheaper energy - using technology to automate boring manual labor - using technology to extend healthy lifespan

Given the demographics collapse and ageing population in most 'developed' countries, we need to look at these other ways of generating economic growth.

Have you looked at the history of the population of various cities, countries or even the globe recently?

It doesn't matter. Even the degrowth due to declining population is exponential in nature. That was the point, population grows exponentially or decreases exponentially. There is no such thing as linear population growth or decline.

Good luck fitting an exponential to the table of population over time in eg https://en.wikipedia.org/wiki/Demographics_of_San_Francisco

Of course, you can do it piecewise, but you can also fit a piecewise linear curve (or even a step function) to any graph.

Or fit an exponential to the graph in https://en.wikipedia.org/wiki/World_population

Just eyeballing the global data from 1950 to 2020 it looks like linear growth would be a much better fit than exponential.

You might also want to look at logistic growth, see eg https://sites.math.duke.edu/education/ccp/materials/diffeq/l...

Are these numbers full time employees only or total FTEs? Because it mentions Walmart: "Walmart’s full-time employees number remained relatively constant for the last 10 years".

Would revenue / person-hour show a different trend? Because there are a lot of part-time and contract workers out there.

Excellent point— contractor and consultant headcount’s have ballooned at many firms. Lots of technical debt being taken on to make short term profits.

The profit increases are from price gouging and anticompetitive behaviors; these companies don't need to provide value to increase their profits anymore.

The number of iphones sold per year is about constant over the last decade. The amount of materials and work-hours in each iphone is probably constant. So if there are no alternatives, "number go up".

The phenomenon is probably a consequent of monetary inflation and lock-in.

Yep this is what I thought as well. All this is probably linked to money oversupply and generalized inflation. It is particularly noticeable on the non-essential items, like tech stuff.

And revenue is a weird indicator to take; profit would make more sense. We know tech gets very good effects of scale but it seems there are limits even for software.

What if we implement UBI in a way that:

1. Companies not participating in UBI: they have to hire people, even for nothing, but all the staff have to be paid.

2. The others: they are focused on making money without hiring people ("efficient money-making machines"), so they are forced to participate in UBI, and therefore distribute their wealth towards society.

So long with "efficient money-making machines".

Unfun thought experiment: In an AGI scenario, what lever of power do the UBI recipients have to force what you've described?

Historically, humanity's biological monopoly on the fundamental resource of "general intelligence" has always been that lever. Looking at the world today, it's pretty clear that democracies are just a temporary balance between the general contempt of the powerful towards commoners, and the fact that the powerful begrudgingly need our economic utility, which is ultimately based on our general intellect. Even callous dictatorships had to exercise partial restraint on violence and murder, to retain a pool of general intellect.

AGI will be a multiple-whammy here:

- most people will become economically worse-than-useless, they will be a total liability on the rich and powerful

- those same people are unlikely to be given access to any levers of power or influence, because they no longer have anything of value to provide to the rich and powerful

- AGI in combination with robotic platforms, after a certain threshold, will permit for insurmountable policing. The "rise up against the robots" cliches we see in film simply become impossible after a certain point.

I desperately hope we end up wielding AI to usher in a post-scarcity utopia, but looking at the types of people who own the world...we'll get what we're allowed to get.

The people in power (whether a government/king, or company) will have the incentive to shrink the population. In the past, the economy grew with the number of people, but now only the costs would grow. There is basically a non-zero chance that our future contains laws limiting who is allowed to have kids or how many. And that's not even the worst action they could take.

Excellent comment. From a pure resource allocation perspective, walking the current path is like looking down the barrel so to speak. I guess our fate is already in the hands of the decision makers. Hopefully theres enough conscience up there. That said, I would expect that if such a position were taken, it would be accomplished via a mass sterilisation campaign rather than direct extermination. More justifiable amongst other decision makers. Less moral burden.

The lever is that whenever somebody tries to give you a scarcity-only token like USD or BTC you say "that's no good here, come back with something that proves to me that you didn't make it by making life worse for the people around you". We (the masses) still control consensus on what counts as valuable, and if were going to break from scarcity as the basis for that (and we should) then it has to be a proper break.

Play by rules that have something to do with our collective values or we're not going to acknowledge your stores of "value". "I got this from my grandpa who was a successful swindler" needs to stop being a source of money that you can expect the children of the swindled to accept. We need to drop the idea that scarcity implicitly creates meritocracy and instead deal in verifications of merit.

If they're really ready to go it alone in their bunkers with robots for friends then I guess we won't be seeing each other anymore, but I kinda doubt that'll work for them in the long run.

> somebody tries to give you a scarcity-only token

Any alternative tokens you could recommend, other than barter?

No, that's a protocol that were going to have to build.

I think CirclesUBI gets it partly right, re: different dynamics between parties based on their connectedness on a trust graph. But I think we'll want this on a CRDT, not a blockchain, and I think we'll want more nuance than their treat-others'-tokens-as-equal-in-value-to-your-own approach. But it's a decent start.

In scenario 1, why would I ever start a business?

This is interesting thinking but I can quit understand how the incentives play.

I mean, when your business is big enough, you'll naturally start to hire more people. Small businesses don't require a large staff and typically don't generate a substantial amount of revenue. The size of a company (i.e., its revenue) generally needs to be in proportion to the number of staff it has, with some flexibility, but not too much.

So you disagree with the article then, these things have not been decoupled?

Compare https://slatestarcodex.com/2018/05/16/basic-income-not-basic...

The capital [in these companies] might not be as decoupled as portrayed but rather funneled from the labor which is not in their employment. This is more glaring in ad tech. For the manufacturers in the list, you have to consider the cheap labor they outsource too, among many other factors. You have to look at global/overall labor statistics vs capital. There might be a trend there but the article doesn’t expand or do any investigation.

It is surreal. It goes like: revenue growth > headcount growth in select companies, therefore capital alchemy on global scale + something something AGI.

Confounding several things here. Of course revenue per head is higher for platform monopolies especially as they get bigger.

> (...) think pieces on how our lives have become consistently better thanks to these tech companies!

[citation needed]

The data presented only talks about the revenue of the companies, not their effect on the lives of their customers (which is admittedly difficult to measure).

The "I personally don't find short form videos too addicting" sentence is just anecdata and dismisses the negative impacts. In my opinion, being able to instantly access content that "makes [you] smile for free" is not obviously a good thing. Even if attention spans weren't affected, I think the brain is wired for homeostasis (balance), and easy sources of dopamine are dangerous since they affect the main motivation to do real, impactful tasks.

Ultimately, it's unclear whether modern (2010s-now) tech companies have _consistently_ improved lives overall. There have certainly been positive changes but also negative ones. Let me know what you think.

The author’s examples are very prominent tech companies during a period of rapid growth in the industry and a single massive retailer (Walmart) with a big online presence. For comparison , I spot checked Target: revenue increased from $85B in 2020 to about 105B$ in 2024, about a ~20% increase. Headcount went from ~ 360,000 to 440,000 during that period, also a ~ 20% increase.

I think a better title might be the less catchy, ‘The decoupling of headcount and revenue at quasi-monopolistic tech companies that are scaling platform products and services.”

Also revenue and capital are not interchangeable terms. The author appears to be confused over this point.

There is a optical illusion that happens on this page for me. When scrolling down or up around the black and white bar graph, a halo around the text (either above or below the bar graph bars), flashes a green hue. Speed seems to matter. Can anyone confirm else see this? I wonder what this optical illusion is called?

On topic though, digital goods and specialized robotics/automatic manufacturing made the first part of this graph possible, now we are on the cusp of generalizable robotics and AI for an even steeper curve. Even if LLMs stay as dumb as they are now, there are incredible gains to be made from them.

Isn’t that how economies of scale is supposed to work?

No, it supposedly makes things more efficient and thus cheaper.

We will need fewer people doing some of the jobs that are being done today, and more people doing other types of jobs. It’s hard to know exactly where this nets out over time, but in the next few years, we expect that this will reduce our total corporate workforce as we get efficiency gains from using AI extensively across the company.

I can’t think of many major resources left to extract. Attention seems to be the last novel one. So what’s the pivot from here? If people are increasingly digitally pacified, what genuinely new needs or industries emerge next? As a consumer, what would I still spend money on if my material needs are largely met, and my emotional needs can be managed, or at least soothed, through digital experiences? (For the record I don't see it as black and white as this- especially the ability for indefinite digital pacification - it's more food for thought) I don't believe the system as we know it is equipped to handle wealth inequality in an exhaustively exploited resource landscape. Especially not one where labour as a resource is being threatened to disproportionately lose value against assets.

Things do tend to balance out over time, but it feels like we’re heading for a real crisis before that equilibrium returns. Unless we pivot away from “productivity” and “efficiency” as our ultimate economic north stars, life for the working class could become increasingly unstable. In many ways, it already has, especially in advanced economies where housing prices have long outpaced wage growth. Not to mention the youths of the world are entering adulthood with far fewer opportunities than previous generations (even college educated). To me it looks like a potential storm of unrest brewing, that could be genuinely historically paradigm shifting.

I don't really have an overall point, but would love to encourage people that say there are naturally going to be new jobs as a result of all this to provide some speculation as to what they are and how we get to that. I don't doubt the notion that some new jobs will emerge, but to think we can find new opportunities for the displacement of even 20% of well established industries seems too optimistic to me. I think we seriously need to start to grapple with a new way of life. UBI being an obvious first step, as this is somewhat achievable as a incremental change, but that could be too little and potentially not meaningful enough. A system that values more than that which can be priced. Maybe money as a whole evolves completely with a well designed digital currency. Sustainability (not in regards to material resources) and wellbeing over exploitation, but none of this is easy to even begin to implement without a collapse of what is already there. Who knows what the future holds.

> and more people doing other types of jobs

IQ is normally distributed.

Therefore, what?

Operating elevators ourselves is impossible

It's true!

Many of these companies have overlapping products and services. If they were to properly combine into a single entity they could probably shed a good fraction of their employees.

Think about it does AMD need to exist at all? Wouldnt NVIDIA be worth more if they didnt exist?

Indeed, competition is highly inefficient. It would be superior to eliminate it. (Signed, The Bourgeoisie)

There would be a lot of negative externalities to removing all competition... Innovation would slow, prices would rise, quality of service would fall.

This would absolutely not be worth the few extra points of efficiency you might get. Maybe you would get a short-term benefit, but long-term the expected outcome would be much worse.

I suspect this has much less to do with AI than the consolidation of industries, especially tech. There’s almost zero antitrust enforcement left in this country, and it shows in these numbers and in the lack of meaningful technological advancement.

As much as it easier to exponentially grow, it will also be easy for companies to shrink earnings or relative status in the hierarchy. This is especially true for companies that make the bulk of their profit selling software.

From spring 2019 to July of this year, I worked at IT at a Fortune 100 retailer.

The project I worked on was enormously successful in terms of revenue growth. I, and people on my team had a huge nationwide impact, which started when we were about a dozen people (it has grown now to several dozen).

Whereas even a manager of one of the big box stores would only have a limited geographical impact. Whereas my work would always have nation-wide impact (and at some companies programmers would have world-wide impact). I turned on a payment option for my platform one quarter, and very quickly people were using it for one million a month in purchases. Which kept going up.

The book Capitalism without Capital talks about this. Some aspects of it are alluded to in Fred Brooks 1975 book The Mythical Man Month.

To build a car, a lot of effort has to be made in making the car - not just the end result, but the making the glass, tires and so forth. Whereas with programming, I write an app, or a feature for an app, and the end result is duplicated and distributed around the country (or even around the world) for free, or virtually free. I'm not helping make commodities one at a time like someone on an automobile line is. It is something different.

> I turned on a payment option for my platform one quarter, and very quickly people were using it for one million a month in purchases.

How is that making a difference? These are sales that were already happening and you allowed them to do something like "pay with paypal" and then claim the entire sale?

Making impulse purchases easier can capture sales that would never have happened otherwise.

That's why Amazon developed (and unbelievably were able to patent) "one-click ordering."

If you implement different local payment providers that you could see higher conversion rates.

As a result, one would assume price should crash to zero and everything would balance out because even though you're having nationwide impact, it fundamentally doesn't cost as much as making a car to make an app... So why should people pay as much for it?

Presently, it's mostly because laws sustain the author's ability to gatekeep the software, whether or not it's even running on capital they own.

One does wonder how long such laws will last, one way or the other, if their end result is a massive inequality of outcome.

AGI is an asymptote. I love it.

Garbage.

1) There's a paywall for part of it.

2) The author finally figured out that, in software, revenue does not scale with labor. The author has not figured out that this applies just as much to a 1 man startup as to a megacorp.

Most ex walk st analysts don’t understand software.

Technology is inherent deflationary due to scaling laws of increasing productivity. This is awesome because think if all the zero or very cheap apps etcs. The productivity is largely based on lack of regulation which allows innovationto occur at speed. Now if we only could deregulate "meatspace" applications (think nuclear, housing, etc) we might be able to achieve great gains.

an alternative reason is that there little to no property with new technology.

if you could remove private property rights over land, housing too would be much easier. Thats a regulation governments arent gonna touch though

You don't have to go that far though. Why does it take months to get housing permits? Or many years to get nuclear permits? Just making those processes more efficient through technology would greatly help, but we can't. It's not the technology slowing these things down.

Reply on news.ycombinator.com