[nhashem]

Most jobs are selling your time in exchange for money. If your time is mostly fungible compared to someone else's time (ie. most low-skill labor), then I'd say all the algebraic equations expressed in the OP are true. He is essentially modeling supply/demand curves of typical scenarios involving "human capital."

In the past, the value of human capital -- the amount you got paid for selling your time for money -- used to be more lucrative. At one point nearly half of Americans were in a labor union, which artificially constrained the supply of human capital so it would be worth more. Conversely, our previous industrial revolutions required a great deal of human capital (building factories, sewers, etc), generating lots of demand that would keep the value of human capital high. Due to political and technological changes, this is basically no longer the case.

If you're a software engineer, then none of these models apply to you because you are high-skill labor (also sometimes called "talent.") You're paid for your time, but your productivity can be orders of magnitude higher than your pay. In the OP's "expendability of labor" graph, the "sweet spot" is enormous. In my nearly ten years as a software engineer, I've seen colleagues laid off/fired for all sorts of reasons, but none of them were due to, "well we did the math, and we decided your production is not worth your salary." I've seen layoffs where the executives literally admitted they would make less revenue/profits due to the lost productivity of the laid off workers not offsetting the savings in salary, but they had to hit a certain "profit percentage" or some other absurd reason that justified the layoffs.

In any event -- I've thought a lot about the macro-society impact of this new model of economy, if we're truly destined for a world where human capital is just not that valuable. My interest in politics and healthcare has largely been based on this -- what should we consider "subsistence living," and what happens if we reach an inflection point where most jobs paying for human capital are below it?

It's possible, of course, that this will all be moot, and our next industrial revolution will have a huge demand for locally-based human capital (e.g. a "renewable energy industrial revolution scenario," where there is a huge demand for laborers that can install solar panels on houses, or something), and the median income for Americans will increase. Or it's possible for a more gradual shift to happen (e.g. the graying of the baby boomers causes a gradual increase in demand for local caretakers of the elderly). But I'm not sure it's a good idea to just assume that will happen.

[zanny]

> where there is a huge demand for laborers that can install solar panels on houses, or something

Here's the kicker - any new industry, at any point in the future, that pops up demanding labor, is much more likely to invest (since they are investing anyway - it is new industry and new markets) in automating the labor out of the equation from the start. Industries are only slow in that transition today because they have established infrastructure around the usage of meat bags on two legs as units of work, but for anything brand new (like semiconductor plants, or the Tesla auto factory) even in the short term it makes more sense to automated the expletive out of any physical work because I'd imagine in even just a year at minimum wage it would pay for itself factoring in the fact you have to build everything up from scratch.

So the solar panel installer in 10 years would probably be a self-driving fork lift truck that carries a bed of industry-standard packed and oriented solar panels, where the vehicle has GPS and will drive to all customer homes and install the panels automatically, with only the need for an electrician to come wire them into the houses electric.

But that is short term. You can easily replace the electrician with some hand held spider robot that will use the blueprints for the house and sensors to find and rewire the electrical into the panel. In a hundred years, you aren't installing new panels, because every new house is factory made (by automated assembly) with the panels preinstalled, and they all have interchangeable parts quality electric systems that dumb non-sensing robots can rewire because they are standardized if necessary.

It is glorious that we are eliminating the need for human capital to make things happen. We are removing people from so many equations you would otherwise be wasting someones time on. We may have some bumps in the road and some dark ages of depravity and extremely concentrated ownership of the means of production and all that, but a thousand years out (assuming we haven't destroyed ourselves yet) any human still alive (albeit with a nebulous definition of human if we start gene splicing and installing cybernetics pervasively) will almost certainly have no need to "labor" at all, because the machines and infrastructure built up over a thousand years will provide for them.

[nhashem]

We may have some bumps in the road and some dark ages of depravity and extremely concentrated ownership of the means of production and all that

Well, this is what I'm personally concerned about. I agree with everything you just said. But what is the politically optimal way to cross over that inflection point? Or to rephrase: what kind of rules should we have in government/society so that the transition doesn't effectively cause a revolution and end up derailed? Should we consider an education system oriented around high-skill labor? Should our system of taxation and assistance account for this new concentration of capital? And can you do this without impeding the very technological/economic growth you're accounting for? Too much taxation may stymie private sector advancement, too little taxation may cause an elective or literal revolution, and the wrong adjustments can easily introduce moral hazards (e.g. nobody working at all before our technology is advanced enough to require zero human capital).

[Retric]

I think adding a living wage and removing the minimum wage is the way you get though this bump. If you make human labor cheap enough someone is always going to want to have their lawns mowed by people or whatever. Assuming they can afford to do so. Basically, people with minimal skills end up working not for food, but to have toy's etc. The trick is finding a number where people can have a little disposable income, but working 20+ hours a week even at 3$ an hour gives them several times as much fun money.

I suspect in the US we could do that now with around a flat 40% tax without deductions. Note: Living wage would need to be a national number ~12 to 15k and not what it takes to get by in NYC or whatever.

PS: What makes the numbers work out is you also get rid of social security, unemployment insurance, food stamps, and the lower tax brakes as well as all other tax breaks. Also, people making the average income are revenue neutral in this scheme as they get back the same amount of money as there paying into it minus whatever overhead is involved in running the program.

[zanny]

I think this could have some other benefits, like population centers re-balancing themselves out rather than having the gross concentrations of otherwise unnecessarily dense capital the way we have now - SF is the perfect example. If you had a guaranteed 15k a year, you could not only take more venture risks, but also live pretty much anywhere and work without the dire need for the profit motive to feed yourself and pay absurd $2k a month rent on a 10x10 room because you have to live in SF or NY to get ahead.

[agsamek]

This idea was evaluated by comunism in eastern europe. Most people do not have enough drive to fight beyond minimum. If you additionally give them political power they will be able to bring outlayers down efectively.

[zanny]

Eastern Europe was about state control of the market, though. Getting a government stipend for living expenses doesn't mean you take control of everything.

What it does mean is that jobs like waiting and cleaning that are currently easily replaced by automation, but are not due to the artificial need to maintain the wasting of hours to get food. I don't think the traditional restaurant industry would survive such a transition, for example, because it is much more economical to have a local food generator that knows all the recipes and can ship freshly prepared meals via automated vehicle to consumers than to have them go to a restaurant, get served, get waited on, get cooked for by kitchen chefs (who are usually doing something between heating up frozen food to following an instruction sheet from corporate) when we can eliminate all that work for most of us, and anyone who cares enough to get the waited experience can pay a real wage for it, because labor demand is more reasonable in such a circumstance - people would only work for you if you gave reasonable value for their time, not just because they need to eat and our progressively greater per-worker output and ability to eliminate human capital in many endeavors all together makes most labor unnecessary.

[yummyfajitas]

I don't think your explanation is quite right. The proper distinction to make is between "Garret Jones Workers" (named after the economist who has studied them) and "Productive Workers".

The "Productive workers" are those who generate profitable production with their time - a factory worker producing widgets, a doctor treating patients, etc. 1 hour = 1 surgery = $500 (or whatever). These workers are best viewed as a cost center - if hourly revenue - hourly cost > 0, they are worth hiring, otherwise they are not. They can be low or high skilled - doctors are a great example of a highly skilled productive worker. Software developers working at an agency would be another.

The "Garret Jones Workers" are those who generate organizational capital. As a result of their work, the revenue stream of their company will increase. Hiring them is an investment - you need to look at (NPV(future revenue generated by the organizational capital they produce) - NPV(cost)) to determine if they are profitable. Examples of this include software developers building a product or a doctor improving business practices at his hospital.

[erichocean]

I see where you're going, but I'd argue that the issues you've identified with labor mostly only apply in the first-world.

There's still an enormous amount of traditional "labor" to be done in the third-world, where just getting clean water is itself still an achievement.

The first-world is definitely heading directly towards the problems you've described.