limits of automation

[Whereisup]

We have seen that computers can do any noncreative job. We already have cars which can drive themselves, so eventually, there will be no jobs for truck drivers. Automated equipment could do hair, cook food, run warehouses without people, etc. An advanced computer could repair other computers and simpler machines.

So essentially, except for high level executives, we could end up with everyone else unemployed. We would then have to figure out a way to get money to unemployed people so that they would be able to purchase the goods and services that the automated economy could produce.

However, and in addition, I don't think computers could do creative work, so there could be as many jobs there as we want.

The reason is that computers can't work with qualities and metaphorical ideas, and those are required to be effectively creative.

Let is look at an example of a quality problem. We see red as warm or hot, and blue as cool or cold.

However, in terms of photon energy of the light, objectively, blue light has more energy than red light. So objectively, red should be cool while blue should be hot.

We could arbitrarily program a computer to calculate that red is hot, but that would ruin information processing which needs to use objective wave lengths. So in encountering red, the computer would be faced with two or more possible ways to go, and couldn't decide which. We could program in what the computer should do with each red, but that would get into so many exceptions we ourselves would have to program in, that it wouldn't be practical.

But the problem isn't just color. It's probably not infinite, but the number of possible metaphorical ideas is at least extremely large. We humans can work with that by comparing qualities. However, qualities can't be expressed in digital mathematics, so computers cannot compare subjective qualities like humans do when they are being creative.

Thus, I think that humans will be needed for creative work. Thus a system which gets money to people to buy goods and services produced by an automated economy could do that by hiring people to be creative, or to study to learn a creative skill.

Jim

 

[SteadyMercury]

I don't get the red/blue temperature thing, why would you even try to make a computer judge temperature from a color palette when it could simply use a thermometer to measure ambient temperature or a metallic probe to sense how much heat is conducted with contact?

 

[william the wie]

I don't get the red/blue temperature thing, why would you even try to make a computer judge temperature from a color palette when it could simply use a thermometer to measure ambient temperature or a metallic probe to sense how much heat is conducted with contact?

Also the problem is not a shortage of jobs but a surplus of debt service payments due to government programs designed to encourage superoptimal purchasing of education and housing in particular.

 

[f252863]

Computers can do creative work.

There are books sold on Amazon today that are the work of only computers. They can also compose music.

 

[expat_panama]

...there will be no jobs for truck drivers. Automated equipment could do hair, cook food, run warehouses without people, etc...

People have been saying that for centuries but most Americans know better.

Real world is more automation = more employment.

 

[Whereisup]

I don't get the red/blue temperature thing, why would you even try to make a computer judge temperature from a color palette when it could simply use a thermometer to measure ambient temperature or a metallic probe to sense how much heat is conducted with contact?

Just to follow the example, if we wanted a computer to do original paintings, it would have to use colors as colors, but that would wreck it's abilities to determine wavelengths of light in terms of energy. Worse, there would be many other connotations of colors that the computer would miss.

More generally than the example, humans can do all kinds of things with connotations of words, metaphorical ideas, puns, and so forth that would simply be confusing/disruptive to a computer, which can only work with digital codes, and information which remains fixed and consistent.

That is one reason that search engines can use key words, but they cannot do a search with open ended questions which depend on subtler kinds of information processing.

Jim

 

[Whereisup]

Computers can do creative work.

There are books sold on Amazon today that are the work of only computers. They can also compose music.

Computers can do a kind of pseudo-creativity, but not truly original works that is of good quality.

For example, computers can compose music in the style of a past composer by counting the number of times the composer uses certain notes and certain sequences of notes. However, that computer composed music is limited and somewhat mechanical. A computer can't compose different shades and tints of music as the composer would have done if he or she had not died.

Stated another way, computers can imitate creative work up to a point, but cannot do creative work.

In the scientific area, computers can work with what has already been discovered, compute statistical values for a group of data, but the computer can't come up with serious new discoveries which leave the box. Only humans can make LEAPS into the unknown with seriously new hypotheses.

Jim

 

[Whereisup]

I don't get the red/blue temperature thing, why would you even try to make a computer judge temperature from a color palette when it could simply use a thermometer to measure ambient temperature or a metallic probe to sense how much heat is conducted with contact?

Also the problem is not a shortage of jobs but a surplus of debt service payments due to government programs designed to encourage superoptimal purchasing of education and housing in particular.

The sentence contains a contradiction. If there were more jobs with reasonable compensation packages, there would be no problem with debt service payments on GOVERNMENT debt as it is now. Debt as a percentage of GDP is still less than it was after World War 2, and we managed to solve that problem. Of course, there could be a problem if the debt were to become large enough.

We aren't even up to optimal on education, although quality is as much a problem as budgets. If we had a good enough educational system, the debt would no longer be a problem, unless it were greatly increased. The reason is that if we were adequately educating everyone, people would make better financial decisions and make more discoveries per decade, which would quickly increase our Gross Domestic Product by tens of trillions of 2013 dollars. Some tiny inventions only add some millions of dollars per year to the economy. Moderate inventions add billions of dollars per year to the economy. A few inventions, such as the invention of computers, add more than a trillion dollars per year to the economy. A better educational system could add a lot more to the economy than you think.

But education doesn't actually cost as much as you think. The reason is that the teachers and other recipients of educational funds turn around and buy goods and services. That means that businesses sell more than they would otherwise, and employ more workers than they would otherwise. All that extra profit and extra employees then produces extra tax revenues which would not occur if we didn't have those teachers buying goods and services. Those extra tax revenues then offset a big chunk of the expenditures on education.

An economy is a much more complex structure than a family budget.

It needs to be organized efficiently of course. Some ways of spending money will yield more return extra tax revenue than other ways of spending money. As much as possible, governments need to spend money in ways that yield a large return extra tax revenue, while avoiding ways that yield a low return extra tax revenue.

But notice that the efficiency of various ways of spending in creating extra return tax revenue isn't something we see discussions. I think the lack of discussions means that the effect of various kinds of spending on future tax revenues isn't something that policymakers take into account at this time.

Jim

 

[Whereisup]

...there will be no jobs for truck drivers. Automated equipment could do hair, cook food, run warehouses without people, etc...

People have been saying that for centuries but most Americans know better.

Real world is more automation = more employment.

Serious futurists who have predicted a full automation economy have never set a specific date, so we are well within the prediction. Most Americans a hundred years ago might have predicted a few advanced machines, but there is no historical data that I know of that people were expecting today's modern computers and what they can do. If you had asked people if the new fangled cars would become able to drive themselves, people would have said "no" and almost none of them, or none of them, would have thought up that idea themselves.

There were a few times in history that specific machines were replacing workers, but those were specific problems. For example, nobody thought that a machine loom would cost farmers, blacksmiths, carriage makers, and so forth, their jobs. It was only the older workers on hand looms that were afraid of losing their jobs, and who did lose their jobs.

So people haven't been expecting a fully automated economy for hundreds of years.

Now, that they can see the technology being developed, engineers are predicting a fully automated economy. So if it is true that most Americans know better, then most Americans disagree with the experts in the relevant fields. And your "most Americans" don't have college degrees in engineering, and don't understand the technology. So why do you think that belief that there can't be a fully automated economy would be well founded?

Jim

 

[expat_panama]

engineers are predicting a fully automated economy

No they're not. If they were we'd have a link from some engineering society announcing it. There isn't because the idea's pure stupidity.

Just the same, engineers do like to automate jobs, and usually the type of work they automate is engineering work. Engineers were not idiots worrying that they'd automate themselves out of work but rather they were smart enough to know that by increasing and lowering the cost of engineering work that more people would have access to it and find they can't live without it. We got more engineers working now than ever.

 

[SteadyMercury]

Computers can do creative work.

There are books sold on Amazon today that are the work of only computers. They can also compose music.

They can be programmed to simulate creativity, not be truly creative.

As an example, say you have a website with a signup process, and you notice many people enter their email but don't complete their registration when they see that they realize they have to pay a signup fee. What do creative people do? They harvest those email to send out marketing spam touting the glory of joining up, and probably include a coupon code for a discount membership rate.

This is a (granted toy) example of human innovation. A simple, cheap, and obvious business/marketing decision but something a computer would never come up with on its own, at least today.

 

[SteadyMercury]

If you had asked people if the new fangled cars would become able to drive themselves, people would have said "no" and almost none of them, or none of them, would have thought up that idea themselves.

I disagree, people have been throwing around the idea of a future of automated and/or flying cars for quite some time.

There were a few times in history that specific machines were replacing workers, but those were specific problems.

I wouldn't describe it as a few times, more like throughout modern history. How can one attribute something like the steam engine to a specific problem? Is the problem specifically to use heat and water to make something move? From that you get everything from automated threshers to boats to trains to automated pumps and cranks, etc. things that people in all industries out of work.

 

[william the wie]

SM, it may not and probably will not put North American workers out of work but automation in the form of better ships have been putting members of the third world out of business with great regularity since the 1400s. Steam engines and similar breakthroughs have resulted in India and China going from either of them being richer than the area of the modern EU in 1400 to the two combined being less rich than the same area today.

And North American automation/productivity is on schedule to overtake Chinese wages in 2015 and Indian wages shortly thereafter. The problem being that 1,000,000 Chinese workers being replaced by say 10,000 Mexican workers or 1,000 Canadian/US workers is going to be a whole lot more noticeable in China than in NAFTA.

 

[HenryBHough]

Sadly Lenny Bruce's 1950's tale of the world's first automated airliner is nowhere to be found on the web. Anyone else remember it....and the understatedly spectacular ending?

 

[Whereisup]

If you had asked people if the new fangled cars would become able to drive themselves, people would have said "no" and almost none of them, or none of them, would have thought up that idea themselves.

I disagree, people have been throwing around the idea of a future of automated and/or flying cars for quite some time.

There were a few times in history that specific machines were replacing workers, but those were specific problems.

I wouldn't describe it as a few times, more like throughout modern history. How can one attribute something like the steam engine to a specific problem? Is the problem specifically to use heat and water to make something move? From that you get everything from automated threshers to boats to trains to automated pumps and cranks, etc. things that people in all industries out of work.

There were only a few Luddite movements in history in which people feared they would lose their jobs. So it was a few.

It is more common in history for people to be excited about new technology. There is material im biographies of inventors which shows that situation.

Jim

 

[Moonglow]

AS far back as the '70's and '80's they have tried to create an automated brick laying machine, but still nothing.

 

[Whereisup]

engineers are predicting a fully automated economy

No they're not. If they were we'd have a link from some engineering society announcing it. There isn't because the idea's pure stupidity.

Just the same, engineers do like to automate jobs, and usually the type of work they automate is engineering work. Engineers were not idiots worrying that they'd automate themselves out of work but rather they were smart enough to know that by increasing and lowering the cost of engineering work that more people would have access to it and find they can't live without it. We got more engineers working now than ever.

That's not the kind of thing that engineering societies do. It's individuals who
report results and sometimes make predictions.

One piece of data is that if you have been following the financial news, you might remember that over the past decade, a very large number of large corporations have downsized due to automation, firing significant fractions of their workforces.

And then, the Wall Street Journal is quite good in terms of accurate reporting and opinion, and the Journal has been talking about major long term losses in employment from automation.

The question isn't whether or not there is a trend moving towards full automation. The question is whether or not there is anything that could prevent full automation.

In terms of non-creative jobs, there doesn't seem to be anything that automation is unable to do. We have increasingly accurate computer vision and computer understanding of language. We're not perfect yet, but there seems to be no reason that computer vision and other computer perception won't soon be adequate to automate any non-creative job. Also, automation can accomplish any movement, and better than humans in accuracy, so there seems to be no reason that the movement part of non-creative jobs can't be automated. Note that the robot surgery arms can cut tissue more accurately than human surgeons can, for example. Computer movement technology is very good.

So logically, what engineering problem could there be that would prevent full automation?

Jim

 

[Whereisup]

SM, it may not and probably will not put North American workers out of work but automation in the form of better ships have been putting members of the third world out of business with great regularity since the 1400s. Steam engines and similar breakthroughs have resulted in India and China going from either of them being richer than the area of the modern EU in 1400 to the two combined being less rich than the same area today.

And North American automation/productivity is on schedule to overtake Chinese wages in 2015 and Indian wages shortly thereafter. The problem being that 1,000,000 Chinese workers being replaced by say 10,000 Mexican workers or 1,000 Canadian/US workers is going to be a whole lot more noticeable in China than in NAFTA.

Before people get scared, I would like to point out that there are many ways to get a good income to people put out of work by automation. I will give just one example, but there are other possible methods as well.

First, we require businesses to give wages to the the automated machinery, set at a specific amount per hour of human labor each machine replaces. Then, since machines don't need money to buy houses go out to dinner,and so forth, we have a machine income tax set at 100% of the wages the machines are given.

Then, the government would hire everyone put out of work by automation to do creative work in science, scholarship, and the Arts, or to do human relations work like having conversations with people in convalescent hospitals, supplementing the computer learning of students, and so forth, or just to study, even for a lifetime. People who are disabled would of course just be supported, and people could still retire as now.

People then would use the money to buy the goods and services that the automated machines produced, closing the cash loop. The businesses which sold the products and services would use some of the money to pay the machine wages and taxes, use some of it to buy new machine tools as needed, and keep some of it as profit.

Since a fully automated economy could produce as many goods and services as people want, within reason, everyone could have a good standard of living and profits could simultaneously be high. It would be necessary to use alternative energy, but of course, the automated machinery could manufacture as many windmills, solar power facilities, and so forth as were needed.

By a good standard of living, I mean at minimum an American upper middle class standard of living. Within reason probably means that everyone could not own his or her own jet airplane. Everyone couldn't have natural jewels, because automation could only produce artificial gemstones. So everyone could have as many artificial rubies and emeralds as he or she wished, but there would only be a small number of natural gemstones available.

Notice that the above plan could be implemented slowly as the economy automated. The machine wage could start low, just high enough to give jobs to those people unemployed by automation now. Then as more people were driven out of work by automation, the machine wage could be increased. That would be smoother than a few massive shifts.

Of course, what is essentially the same thing could be done in other ways. The above example mainly shows that it would be easy to handle full automation, whether the above idea itself is used, or some alternative.

I don't insist that it has to be my idea. Any different idea which works is fine with me.

Jim

 

[expat_panama]

...over the past decade, a very large number of large corporations have downsized due to automation, firing significant fractions of their workforces...

People say that, and too often words go one way and actions go another. Let's deal with actions.

The bls.gov site reports that 144,386,000 Americans are employed --mostly by large corporations-- and ten years ago the number was 138,424,000. What we got is that "over the past decade" any factory automation/layoffs that we've had were more than offset by hirings at the companies that are automating the older factories.

 

[Norman]

Why are you confusing "What computers can do now?" with "What computers can do in the future where everything is automated?".

Anyway, who cares about jobs if the computer takes care of your job?