How realistic is "Global Peak" series now?

[Starglider]

We could have had some ugly oil scarcity if enhanced extraction techniques hadn't worked or hadn't been developed in time. As I've said, for industry insiders and people who did serious research on the subject (not the general public and not most article writers: not me either) it was pretty obvious even in the early 2000s that fracking and related enhanced recovery techniques were ready for prime time waiting for just a price spike to get them over the hump into mass deployment. However in some alternate universe where the EPA pre-emptively regulated anything involving injecting 'chemicals' into the ground into oblivion (as the environmentalists wanted them to), R&D would have gone a lot slower, US supply would be lower and the Brent price spikes would have been considerably higher.

[K. A. Pital]

Probably peak water.

I am tempted to say "peak idiots", but that was already explored in Idiocracy in a brilliant way.

[Adam Reynolds]

I am tempted to say "peak idiots", but that was already explored in Idiocracy in a brilliant way.

I suspect a major problem will be peak employment. When robotics/limited AIs really can be cheaper and more effective than people, there is no reason to hire them in any real numbers*. An entire generation in developed nations will never be able to be fully employed through no fault of their own. This will require a major revamping of the dominant economic system of capitalism to suit this new era.

* In economic theory, this is called the marginal product of labor and capital. Assuming capital can fully replace labor, there is no economic incentive to hire people.

[K. A. Pital]

I consider marginalism a vain model, especially considering that adjusting wages to the marginal product of labour in labour-surplus economies drives the wages below subsistence.

Truth is, economics is a social science. The current order will be quickly overthrown if it were to seriously threaten the large-scale employment of its own populace. Think Greece or Portugal - but on a global level. Such governments (and corporations lobbying for such governments) will be crushed by popular uprisings. At least, that's what I hope will happen. If crushing poverty with total productive power and all economic control in the hands of capitalists happens instead, humans truly are the fools.

[Elheru Aran]

People need something to fill their time, whether they're paid for it or not... but pay does give them an incentive to keep doing it. So I suspect that even if much of the actual production is done by robots, there will still be jobs created that work around that. If someone is laid off from the factory where they made cheeseburgers, they may be able to find a job selling those cheeseburgers instead, or learn how to fix the machines, or do something else like running online service call centers (to throw out a vague example).

There's almost certain to be a transitional period, but how severe the effects of increased automation will be I couldn't say at the moment.

[Zixinus]

About labor: it will be more of a case of labor needing more investment to be useful. Machines replace manual labor where it is simple enough to be replaced by machines and instead create need for people that can operate, maintain and create the machines. Kind of like how plumbers replaced nightsoilmen. The problem will be that a person will need more and more education to be useful in the labor force.

[Ralin]

About labor: it will be more of a case of labor needing more investment to be useful. Machines replace manual labor where it is simple enough to be replaced by machines and instead create need for people that can operate, maintain and create the machines. Kind of like how plumbers replaced nightsoilmen. The problem will be that a person will need more and more education to be useful in the labor force.

If a machine can do the work of twenty people and requires five people to run, make and maintain it that's still -15 jobs. If enough new technologies swing in that direction we're still looking at a pretty severe deficit of jobs for people to do.

[Kingmaker]

About labor: it will be more of a case of labor needing more investment to be useful. Machines replace manual labor where it is simple enough to be replaced by machines and instead create need for people that can operate, maintain and create the machines. Kind of like how plumbers replaced nightsoilmen. The problem will be that a person will need more and more education to be useful in the labor force.

If a machine can do the work of twenty people and requires five people to run, make and maintain it that's still -15 jobs. If enough new technologies swing in that direction we're still looking at a pretty severe deficit of jobs for people to do.

If that were truly how it worked, we'd all be mostly unemployed already (see mechanization in, e.g. agriculture or manufacturing). There's not a fixed demand for labor, though. Heightened productivity means you produce more with the same labor inputs, or that you are able to use that labor elsewhere to do something else (possibly something new). Those 15 people may be jobless in the short-term, which obviously is not great for them, but that doesn't mean that the economy is running a net loss of employment in the long term. The only way they're going to be out of a job forever is if they are literally no longer worth employing (well, there are other possibilities as well, but they don't relate to robots).

[Adam Reynolds]

I consider marginalism a vain model, especially considering that adjusting wages to the marginal product of labour in labour-surplus economies drives the wages below subsistence.

Truth is, economics is a social science. The current order will be quickly overthrown if it were to seriously threaten the large-scale employment of its own populace. Think Greece or Portugal - but on a global level. Such governments (and corporations lobbying for such governments) will be crushed by popular uprisings. At least, that's what I hope will happen. If crushing poverty with total productive power and all economic control in the hands of capitalists happens instead, humans truly are the fools.

Indeed it is a problematic model, which was my point. I hope that something like this would happen, but America at least is so obsessed with capitalism I'm not so optimistic there.

As for economics, there is a classic joke that summarizes the problem. A physicist and an economist are trapped on a desert island with a can of food they are attempting to open. The physicist claims that he can use a complicated system in which he drops a coconut on the can and open it. The economist states that this is ridicules, the solution is that he will just assume he has a can opener.

If that were truly how it worked, we'd all be mostly unemployed already (see mechanization in, e.g. agriculture or manufacturing). There's not a fixed demand for labor, though. Heightened productivity means you produce more with the same labor inputs, or that you are able to use that labor elsewhere to do something else (possibly something new). Those 15 people may be jobless in the short-term, which obviously is not great for them, but that doesn't mean that the economy is running a net loss of employment in the long term. The only way they're going to be out of a job forever is if they are literally no longer worth employing (well, there are other possibilities as well, but they don't relate to robots).

It's worse than just a short term problem in that sense. Things like designing said robots require skills that a large portion of the population actually lack the skills to do. Even doctors and stock brokers could easily be put out of work by computers that can do their jobs better than them. Remember Watson, the IBM supercomputer that won at Jeopardy? It's primary job was and is learning how to diagnose patients better than a doctor. It's not far off.

Even if it were merely a short term problem, in America, right now, 3.5 million drive for a living, one of the largest blocks. What happens when driverless cars(or delivery trucks) put them all out of work. What insurance company would ever consider allowing a person to drive when a computer doesn't make mistakes as often? It won't happen overnight, but as it happens, more and more people will be put out of work and not have skills to do anything about it.

[Lonestar]

What happens when driverless cars(or delivery trucks) put them all out of work.

For a lot of freight, I strongly suspect they'll act as security guards riding along. Obviously not all, but a fair chunk of it.

At least until we get security droids I guess.

[Patroklos]

Someone still has to take it to your door, get the signatures, etc. etc.

[Gandalf]

Theoretically the van could send you an alert telling you to come outside, and then you could collect it yourself. They could use the electronic signature devices, record the interaction, and then be most of the way there.

[K. A. Pital]

Long-range shift driving will be obsolete with automatization. It is possible for one man to do the work that previously required two.

Those 15 people may be jobless in the short-term, which obviously is not great for them, but that doesn't mean that the economy is running a net loss of employment in the long term. The only way they're going to be out of a job forever is if they are literally no longer worth employing

Actually, no - the worth of their employment can go down and their life standard with it. Which is pretty much something we have witnessed throughout these decades - the real incomes of the working class are either stagnating or decreasing. This is also evident in a longer-term evolution of purchasing power of said incomes. Which is what I've alluded to on the previous page. High tech - low life.

[Jub]

Someone still has to take it to your door, get the signatures, etc. etc.

Not as much as you might think. Loading/unloading and warehouse management can already be automated and some of the busiest warehouses in the world are already running virtually human free. Door-to-door mail delivery is on the chopping block in many places as a cost cutting measure and if you were willing to design around it an automated mail truck could easily load and unload roadside communal mail boxes. For parcel delivery, things get a lot simpler when you either pay the premium for drone delivery to your property or get an email telling you to pick up packages at the post office/storage depot between 9 and 5 Monday to Saturday. Even the storage depot could be mainly automated with customers running a smartphone app or manually entering their delivery confirmation at a help terminal before an automated system retrieves the package and send it up to the desk. You'd probably want one or two customer facing staff members to help customers use the system, a manager to run the location, and maintenance guys who are probably outside contractors. Seven regular employees per location tops, with most of them being part time minimum wage cashier analogs.

Things like cashier jobs are already being replaced by self-checkouts, and the only thing stopping them from spreading are unions. Online shopping is only growing, and we're already seeing companies being driven under by this. Game and music stores are already just barely hanging on, and game stores might just go the way of the video store if digital distribution keeps growing the way it has been. Things like clothing stores could probably be made to sweat if a Kinect-like product could take a body scan and find clothing that fits were to hit the market. Fast food is already basically a robotic assembly line and it won't take many more price cuts before kitchen renovations start to coincide with large-scale layoffs.

I'm not saying that all of these things will happen, but loads of manual labor and customer service jobs will be within spitting distance of the chopping block within the next decade or two.

[Patroklos]

There is a difference between set robots lose on a controlled, carefully mapped and regulated space graded and organized specifically for their purposes and limitations and my house which has 24 cracked and crumbling non standard steps up a 45 degree incline and curves 90 degrees before requiring you to go through a latched gate, around my porch and ring my manual 1970s door bell. If there are UPS robots doing that then we probably have automated mechs fighting our wars for and the end is near.

There are lots of things robots can and will be doing shortly no doubt. We have been exaggerating and predicting the coming of our robot slaves ans shortly there after masters for many decades and so far our track record isn't so good. On a more contemporary not driver less cars are almost here, just like they were 20 years ago.

[K. A. Pital]

Driverless cars were not "almost there" 20 years ago, not by a long shot.

[Zixinus]

If a machine can do the work of twenty people and requires five people to run, make and maintain it that's still -15 jobs. If enough new technologies swing in that direction we're still looking at a pretty severe deficit of jobs for people to do.

I was talking from the perspective of creating new labor. In fact I think that there will be (probably already is) a problem where with the current education system we cannot educate enough people sufficiently for them to enter the useful labour poll while still having people that are educated but not enough to be (easily) employable. Thus we have have a labour shortage and unemployment at the same time.

[Simon_Jester]

I consider marginalism a vain model, especially considering that adjusting wages to the marginal product of labour in labour-surplus economies drives the wages below subsistence.

That's not an irrelevant model, that's an important result of the model.

See, marginalism models tell you exactly what a free market economy is going to do, which is important because it tells you what will happen if you leave the system alone. If the supply-demand model predicts an unacceptable outcome (e.g. wages dropping to below subsistence levels), then you know in advance you're going to have to do something or experience the consequences.

Whereas if we were entirely ignorant of such things, we would just slam into this chaotic mess with no chance of addressing the issue in advance in an orderly fashion.

Truth is, economics is a social science. The current order will be quickly overthrown if it were to seriously threaten the large-scale employment of its own populace. Think Greece or Portugal - but on a global level.

I've had arguments about this with some of the free-market fanboys. What it comes down to in my opinion is this. Market models do not really address what happens if social conditions become bad enough to trigger a revolution. But they do give you a lot of information about the conditions capable of triggering such a revolution, which is useful knowledge and a positive reason to have the market model.

People need something to fill their time, whether they're paid for it or not... but pay does give them an incentive to keep doing it. So I suspect that even if much of the actual production is done by robots, there will still be jobs created that work around that. If someone is laid off from the factory where they made cheeseburgers, they may be able to find a job selling those cheeseburgers instead, or learn how to fix the machines, or do something else like running online service call centers (to throw out a vague example).

There's almost certain to be a transitional period, but how severe the effects of increased automation will be I couldn't say at the moment.

The big issue is that there are so many categories of labor which are likely to be replaced or cut back in the next few decades. And there's no strong evidence that new categories of job are going to emerge to take the place of the reduced jobs, which is what happened in the transition to the factory economy and (mostly) in the transition to the service economy. We may just plain reach the point where job destruction does not lead to a balancing job creation.

And there's the education issue on top of this; it may well be that in 2040 it's nearly impossible for someone with an IQ of 90 or below to find work... and then in 2060 it's impossible for someone with an IQ of 105 to find work... and so on.

Those 15 people may be jobless in the short-term, which obviously is not great for them, but that doesn't mean that the economy is running a net loss of employment in the long term. The only way they're going to be out of a job forever is if they are literally no longer worth employing (well, there are other possibilities as well, but they don't relate to robots).

Thing is, that's actually in danger of happening. We're producing large numbers of children today who simply do not have skills much beyond "ask Siri to do it." In twenty years' time, Siri will be a lot more capable and flexible. They won't.

If AI on the Siri or granddaughter-of-Siri level is combined with technologies now in development (like networks of drones to physically carry stuff around, like improved robot vision so robots can see and navigate and function autonomously in a 3D environment on their own)... At some point, the only people who are employable are those with fairly impressive levels of critical thinking, analytical reasoning, and so on. Or those with exceptional people skills.

Anyone who lacks those characteristics, will have a hard time finding a job that can't be done by machines.

If a machine can do the work of twenty people and requires five people to run, make and maintain it that's still -15 jobs. If enough new technologies swing in that direction we're still looking at a pretty severe deficit of jobs for people to do.

I was talking from the perspective of creating new labor. In fact I think that there will be (probably already is) a problem where with the current education system we cannot educate enough people sufficiently for them to enter the useful labour poll while still having people that are educated but not enough to be (easily) employable. Thus we have have a labour shortage and unemployment at the same time.

I'm smack in the middle of this.

The fundamental problem we have is that we're trying to educate everyone, including people who are at best marginally employable even now because they have severe behavioral problems or are just plain unmotivated. These are going to be some of the first to lose their jobs to automation (say, by 2025). And that's if they haven't already lost them, which some have, which is why we currently have high-ish permanent unemployment.

And because of this, we have to dilute and divide our resources, and accept excessive levels of disruption to the learning environment, as the cost of doing business. Thus, the children close to the median in aptitude and motivation have bad examples set in front of them on a regular basis, and don't get the level of rigorous, disciplined instruction that would prepare them best. Then they leave high school inadequately prepared and lacking critical thinking and analytical skills...

...At which point they become the next wave of marginally employed people, who will be losing their jobs around 2035 instead of 2025.

[K. A. Pital]

That's not an irrelevant model, that's an important result of the model.

I did not mean to say that it is irrelevant as a model, but that it describes a behaviour that is not evidenced in real life. This means the model's applicability and predictive power suffer. It cannot adequately explain the behaviour of the labour market, neither can it adequately predict what happens in this market, because it requires using concepts from sociology. If we admit that wages are, in fact, to a large extent determined by class struggle - or call it collective negotiations and politics - then we can no longer rely on marginalism to predict what wages will be like. The problem, of course, is that even factoring out politics (because usually surplus labour territories are the Third World, where minimum wages do not exist and theoretically nothing precludes downward adjustment of wages), they still do not go down below subsistence. Which means that even in this situation with no barriers, marginalist prediction fails. This is remarkable.

[Simon_Jester]

That's not an irrelevant model, that's an important result of the model.

I did not mean to say that it is irrelevant as a model, but that it describes a behaviour that is not evidenced in real life. This means the model's applicability and predictive power suffer. It cannot adequately explain the behaviour of the labour market, neither can it adequately predict what happens in this market, because it requires using concepts from sociology. If we admit that wages are, in fact, to a large extent determined by class struggle - or call it collective negotiations and politics - then we can no longer rely on marginalism to predict what wages will be like. The problem, of course, is that even factoring out politics (because usually surplus labour territories are the Third World, where minimum wages do not exist and theoretically nothing precludes downward adjustment of wages), they still do not go down below subsistence. Which means that even in this situation with no barriers, marginalist prediction fails. This is remarkable.

Marginalism does indeed break down in situations where the underlying economic assumptions become invalid (i.e. the people selling labor will riot, or the people buying it go out of business, rather than have work be done for the suggested wage). This is very similar to the way models in physics can break down when, for example, the oscillations of a pendulum become large, or when the speed of a body moving through air approaches the speed of sound.

Unfortunately, there are many political ideologues who do not understnad how to respect the limitations of a model. Physics does not suffer so much from this because physicists have relatively little stake in clinging to an inapplicable model. No one feels better or prouder for pretending that subsonic aerodynamics are an adequate tool to model a supersonic aircraft. People do feel better or prouder when they pretend that market economics (which they understand and somehow identify with, as a sort of Renfield-Dracula relationship) are applicable to environments where a market is unsustainable and will result in riots, economic collapse, or other breakdowns.

[Welf]

If a machine can do the work of twenty people and requires five people to run, make and maintain it that's still -15 jobs. If enough new technologies swing in that direction we're still looking at a pretty severe deficit of jobs for people to do.

So what? This also means that 15 people never have to work in their life and still can have a full income - provided you put a 75% tax on the produced output. Or you can go from a 40 hour work work to 8 hour work week. That does sound pretty nice. And in fact, that is what the advanced European societies did. In 1960 a German worker laboured 2.163 hours per year, in 2010 1.371. At the same time redistribution increased and and more and more of the economy shifted to social services like health care and care for the elderly. This is the way go, and it is feasible.

[Simon_Jester]

Yes; the big problem is making this actually happen, when there are powerful economic blocs who oppose it because they want control of the productive machinery to translate into the greatest possible share of wealth.

[K. A. Pital]

The big problem is not mistaking real achievements for something else (falling real incomes and the rise of poorly guaranteed part-time employment).

I am not sure a 1960 worker had to face lease labour, temp labour and part-time employment on the same scale as a typical worker in Germany does now. Are you?

I need the data to be able to say "that is the way to go" - before making a judgement on whether it is desireable. That something is feasible does not actually mean it is the way to go.

"8 hour work week" is a nice idea only if it is paid as a 40-hour work week was back in the day.

[Patroklos]

Driverless cars were not "almost there" 20 years ago, not by a long shot.

Thats the point, despite plenty of people predicting they would be at the time.

[Arthur_Tuxedo]

The big difference between Germany and the US, unfortunately, comes down to racism. It's not a coincidence that the dismantling of the social safety net occurred directly after the civil rights struggle, and the nations with the most generous social services are also the most ethnically homogenous. Most of the people voting to slash social services don't even have negative views of minorities, just unconscious biases that make them think that welfare recipients are lazy and undeserving.

As for mechanization of labor, there are a lot of white-collar jobs that are under more threat than blue-collar ones. It will take a major leap to replace a person installing solar panels on a roof, but replacing a neurosurgeon who makes $700K per year with a machine that is thousands of times more precise is just over the horizon. Lawyers and legal professionals will no longer be required to go over "boiler plate" documents in the near future, freeing up more time for the lawyer to do productive tasks but making the assistant's job obsolete. Compare that to a plumber, whose economically viable robotic replacement is much farther away. Automation will hit every class of worker, which actually gives me hope since it greatly increases the odds that the political establishment will do something to soften the blow.