Farmers. In 1850, >70% of people in the U.S. were farmers. Today <2% are. (Source: http://www.nytimes.com/1988/07/20/us/farm-population-lowest-...)

Yet we don't have 70% unemployment.

That's true, but that should not lull us into complacency. Some day we may run out of productive things for unskilled human labor to do, the same way we ran out of things for horses to do (http://www.cgpgrey.com/blog/humans-need-not-apply). Not everyone can be a software engineer, and even if they could, we probably wouldn't need 7 billion of them.

Yes, but that also means humans are usually more resourceful than what we can imagine in the first place. Look back also in the early 1900s where all major newspapers were proclaiming the end of world and massive famine if we ever reached 1 billion people on Earth, and look where we are now: virtually no famines anywhere (except in war zones and in the worst dictatorships, which is now a minority) and the population is still growing and poverty is decreasing progressively. This would have been unimaginable 100 years ago.

So, let's not fool ourselves with predictions. We are notoriously bad a predicting long term trends.

PS: the analogy with horse is wrong in so many ways (because horses are not humans and cannot adapt like humans) but this would be another topic altogether.

Humans as a species are pretty resourceful, but individual humans often not so much.

The problem is that throughout human history the demand for unskilled labor has exceeded the supply, and that has been the foundation of our social contract: work, and you eat. Don't, and you starve. But some day soon, technology could make that assumption no longer true, at which point we could have big problems unless we radically rethink how our society is organized. That is something we humans have historically not been very good at.

  Look back also in the early 1900s where all major newspapers were proclaiming the end of world and massive famine if we ever reached 1 billion people on Earth"

"It is estimated that the population of the world reached one billion for the first time in 1804." Source: https://en.wikipedia.org/wiki/Population_milestone_babies

>> Not everyone can be a software engineer, and even if they could, we probably wouldn't need 7 billion of them.

If we all put our backs into it, we can make sure that we need at least that many.

We just have to write really bad documentation and really sloppy code, which lot of people believe we are doing already.

With a bit of effort maintaining the software being written today will become such a monumental task that no Soft Eng will ever be out of work for millenia to come.

Have you heard the anecdote about the fox with fleas? It lowered itself into a river, very slowly, giving the fleas enough time to move ahead of the water. In some versions, it was holding a leaf in its mouth. In others, the fleas just gathered on the tip of its nose. Either way, when eventually all but a tiny fraction of the fox is under water, it ducks under completely and the fleas all float off downstream.

When the fox's shoulders went under water, I bet some of the fleas said "ten minutes ago 70% of us lived below the fox's shoulders and yet they're all still on the fox, so there's nothing to worry about."

There are a number of assumptions, key among them are the fox is definite extent and if it is, we are near it's nose already.

is going up the nose an option?

Only if the given flea is driven to be 'disruptive' ;-)

No, we developed lots of non-essential busy work to cover for it and sold it as "services".

So how does that work? The government/rich people/"capital" conditioned people to consume services whether they need them or not, in order to create these jobs? If so can you give some examples of this? Or is it that government/industry created regulations and practices that created these jobs (I could think of law as being like this, but not many people work in law, less than 1% of non-farm employment in fact)?

My take is that we use the labor that was being used in farms for new productive purposes including health, education (as consumption, not investment) and entertainment. It's undeniable that life expectancy is increasing so we are getting something out of this increased efficiency.

Services which generate way more revenue/profit and better waves for employees than individual farm workers. I dont think we have lost anything as a whole, on the contrary.

I'm not comparing it to farm work, but to what it could be.

You don't seem to realize that allowing labor to move from critical/essential work to nonessential work is a good thing. That's how the system is supposed to work. What would you call sommeliers, chefs who create Michelin-rated restaurants, movie stars/writers/producers, surfing instructors, etc if not nonessential?

Automation makes things cheaper. Look at food, it was the largest expense for most families for a very long time. Now for most families in the US it's not even close. We have taken the jobs that used to be in that area and gotten rid of them, and instead moved them into service industries like the ones I referenced above. Making things cheaper means more disposable income to be spent on luxury/nonessential items, which creates a market which didn't really exist, which in turn creates jobs. This is not hard stuff here.

Yes price reductions lead to new demand, hence new jobs.

But what if for almost every human supplied product/service there's a good enough machine substitue, or even just a human substitue that demands very little labor ? couldn't we reach a stable state of high unemployment ?

There are still a bunch of jobs where machines are unlikely to replace humans anytime soon or ever, because the cost of deploying machines everywhere and maintaining them will also reach a plateau. We may be good soon at making brains for machine but they still need mechanical parts and require maintenance. So they have a non negligible cost.

>You don't seem to realize that allowing labor to move from critical/essential work to nonessential work is a good thing. That's how the system is supposed to work. What would you call sommeliers, chefs who create Michelin-rated restaurants, movie stars/writers/producers, surfing instructors, etc if not nonessential?

I call them "creative professionals who do what they do by choice".

The majority of non essential work is not "movie stars and Michelin chefs", it's either automatable burger flipping or bureucratic busy-work created to overcome the lack of a system for distributing social wealth outside of "roll around for it".

To put it in another way, we could maintain our current production and give free massages and pedicures to everyone with far fewer jobs than we have now.

Wow, yeah that's a pretty compelling example. Thanks.

Well cloth / clothing manufacturing is probably the largest decrease I can think of. From the cotton pickers, to the spinsters, to the weavers, and tailors. Much of it has been mechanized.

The difference with robots though is that they can actually 100% out perform humans in the not to distant future. Once a robot has hands as agile and controllable as ours, and once they have energy requirements that are cheaper than ours, I don't see how humans could compete with them in non-cerebral jobs. And there is a good segment of the population that can only do physical work.

I would intuit that the proportion of labourers to the rest has been decreasing throughout human history. I think what has me most curious is how low can it go. Is it possible to have near 0% labourers?

I'm not sure you can blame refrigeration on the absence of milk delivery people. In England, we had milkmen/milkmaids well into the late 1990s, despite refrigeration. Electric milk floats (sometimes carrying orange juice and bread, too) were a common early morning sight and sound.

It's 2016 - and still do. We live in a region with an operating milkman.

The price per pint is roughly 2x what it'd cost in a store, but the milk is fresh, organic, delivered to our door - and very importantly for us - generates very little waste. The empty glass bottles go back with the milkfloat, so we use far less plastic cartons.

I think it is very difficult to understand the relative inefficiencies between different product delivery pipelines. In this case the difference between door-to-door delivery of fresh milk vs. the store distribution model.

You may be 100% correct in this case, but in general, I think this is a very hard thing to discern, especially without a pretty deep analysis.

I'm not saying it is perfect, but I would think that 'price' is highly correlated to efficiency of the product delivery pipeline and efficiency is probably well correlated to 'waste'. The major confounding factors would be understanding what externalities are (or are not) being included into the final price and the distortions introduced administratively by government policies.

So a 2x price factor, to me, is indicative of possible inefficiencies relative to store distributed milk but certainly not definitive.

In this case I would think it fairly easy.

First, the volumes are much smaller. A large store chain can negotiate prices lower because they will be shifting a lot more product. The effect on unit margins adds up on both sides of the equation.

Second, the cost of labour. A grocery store should have a fairly big figure for units sold per employee. (Not just milk, but everything they sell.)

Third, we are obviously paying for service and convenience. The milk is delivered fresh, early in the morning. We just open the door and get the bottles. The fact that the empty bottles are taken away for no extra effort is an additional convenience factor, but one of the big reasons we originally went for a milkman service.

Number four - the milkman company is a business. They need to make a profit too.

Because the company operates with a dairy, I can probably discount the cost of specific cold chain maintenance. They would be maintaining the equipment and cooling in any case due to the products.

We could get groceries delivered from the shops, but that is still less convenient. The delivery slots are either during the day or in the evening, and we need to be at the door. Delivery is an extra charge, and priced so that it makes economical sense to order a lot more in one batch. This actually adds a disincentive to ordering milk: we enjoy fresh milk, and with longer delivery intervals it sure wouldn't be as fresh when we drink it.

And finally: the amount of plastic waste, even with our reduction tactics, is pretty significant. If we bought our milk from the store, we would need at least one more plastic recycling crate to hold all the milk cartons.

A store still has to buy the milk from a dairy. All their buying power isn't going to make their milk cheaper than milk delivered directly by the dairy.

Is it really less wasteful? Cartons seem to be easily recyclable using an hydrapulper, while the milkman distribution requires an extra truck trip, wasting a lot of gas compared to just carrying a couple of milk packages along with the rest of the groceries.

It's not wasting gas of the float is electric. Plenty of clean ways to charge that, especially considering they're unused during most daylight hours.

I believe this person is describing delivery by way of some kind of boat, rather than automobile.

Nah, milk float is a term in British English: https://en.wikipedia.org/wiki/Milk_float

You might also be able to buy glass bottles of milk at the grocery store when you're buying other groceries already, which would save the milkman the extra trip. You'll get your deposit back on the next grocery trip.

Dock workers (almost entirely replaced by containerization).

Ops guys (almost entirely replaced by containerization).

(I kid, I kid)

Some things come to my mind:

- wooden barrels / casks for wine beer

- coppersmith (not sure about the exact term. i mean people who create pans and pots by hand)

- basketmaker

- coachbuilder

Wooden barrels for wine are still very much in use, and still largely made by hand. Source: I've been to a lot of wineries and talked with a lot of people who make barrels.

A lot of old professions are still reflected in surnames. A barrel-maker is a Cooper. Coachbuilder would have been Wainwright.

Butner, a button maker

I feel so sorry for William Shatner.