Thursday, 16 February 2012

Further Collapse of the economy due to the increased automation of jobs by software?

Some commentators reckon that many of us will end up put out of work by sofware in the near future and that thus all of the wealth will inevitably end up in the hands of a very few (say the famouse "1%). That sounds plausible but is very simplistic, *Marxist* and not at all taking into account the realities of the situation:

Even today most work is already done by computers.
I doubt that all of the jobs will be eliminated for a simple reason: the banks can't allow it to happen or they will collapse.
More likely we will continue with the scenario we have now of boom and bust and the average knowledge worker fitting the "I will pretend to work and you will pretend to pay me" scenario whereby
they are effectively thinking most of the time then pushing buttons for a small amount of time to produce products/deliverables.
Additionally unless we get strong AI, where is the basic research going to come from to produce new generations of products?
The "rich" can't simply gobble up all manufacturing capacity and that's all there is to the economy.
Even today the vast majority of the economy is in services. Some of that can be further automated but the thinking cannot yet be.
Many companies biggest resource (tired old cliche but still true) is their intellectual property generated by humans.
Are the rich so much smarter than the rest of us that they can generate *all* of the intellectual property all by themselves unassisted?
In the other extreme nightmare scenario where the rich don't give a shit about services and new intellectual property and are interested in shrinking
the economy down to automated manufacturing and high end services with no new intellectual property due to lack of scale (i.e. stagnation) and the rest of us are in blinding poverty:
I reckon that would be a recipe for revolution, never mind the fact that the aggregate economy would shrink and there would be less rich people.
It's not in the interests of the rich to put everyone out of work. We're a "resource" and the more productive we are the richer they are.
I suspect that instead we may just see more booms and busts instead as the only way to drive the fake economy and create jobs will be to print money
and force it through handfuls of pre-picked "winners" like the zombie companies of japan.
On the other hand one other super optimistic scenario might be in the absence of strong AI we simply incrementally upgrade the tools that each worker is using.
We're making the assumption that only university educated people can use automated software or hardware tools.
That's not even true today. Call centers use highly automated systems. Yes we may get to the stage that many of the "scripted" call centers could be
automated and that would thus throw people out of work, right? Well it comes down to a combination of trickle down economics combined with a zero sum game.
If the game is zero sum (i.e. no net new profit caused by further automation) then yes, when their jobs are automated out of existence there will be no new
jobs. In all likelihood, however, the profit margin of companies who use automated software tools instead of people will increase because otherwise why do it?
It doesn't make sense for a company to invest in automated software tools to do a job if humans are cheaper. So we can definitively say that profit margins will
increase. This will means the owners have more money to spend. Now here's a question: Do higher income people or business owners spend most of their money on a. manufactured product
or b. services? Here's a further question: do higher income people spend a higher or lower share of their income on *personal* services?
The answer is of course more. Therefore those displaced from low end service jobs will find themselves doing more personal services which are not automated.

On the other hand, what about the higher end jobs that currently require a university degree or significant training.
Surely some of those jobs will be eliminated by more intelligent automated tools?
Think again. If the tools are semi intelligent themselves then with adequate training even dummies will be able to operate them in all but the most limited set of circumstances since.
Instead the dummies will find themselves enfranchised in much the same way that high paying manufacturing jobs in the city raised the income of poor farm workers
who left the land to find work in factories and in places like Detroit they ended up being middle class instead of working class or poor.

Now you may be sceptical especially if you think that the economy is zero sum. In fact, the economy always has been about the growth of some sectors
of the economy and the collapse of others. That's because of the continual development of new products as scientific research advances.
There is currently massive change in China but apparent stagnation in the Western world. Western commentators seem to think that there
is a global problem of stagnation. There is not. There is a temporary imbalance whereby China and other "developing" countries are cheaper
because of labor and/or better more modern supply chains. That does not mean that the *global* economy is shrinking. On the contrary.
The real challenge facing us in the West is to develop new industries. New industries are based on new products and are created in a process
called "creative destruction" whereby the old non-competitive industries collapse and are replaced by the new. Horse drawn carriages were replaced
by automobiles. How many people can name the most successful manufacturer of horse drawn carriages in the 19th century? Not many. It's gone.
But GM, Ford, Nissan et cetera are here and are shortly to be in fierce competition with Chinese competitors who may or may not be more effective than they are.
Likewise, the likes of Bell and AT&T etc have had to contend with the rise of cellphones and cable companies have had to contend with the rise of the internet.
Telegraph companies had to contend with the telephone before them et cetera et cetera.

Right now the two main risks to the economy growing are nothing to do with the automation of jobs by information enhanced software tools.
Those two risks are the debt overhang from the housing bubble in the western (and especially English speaking) countries as well as shifting growing demand
for transporation away from oil as conventional oil supplies peak and start to decline.

In all cases, the problems need to be solved by *more* innovative products which in turn will generate new industries which we desperately need in order
to replace the industries we cannot compete with against rising stars like China.

Where will these new products come from?
Basic research and design.

In fact in an abstract sense the core of the economy itself *is* is R & D and *everything* else is support for R & D. Let's take a look at how that part of the economy
will be affected by increasing automation:

Those of us who are university educated are in no danger of being automated out of existence, we'll simply be using more and more and more powerful tools and adding much more value
than we currently are. Take a university researcher for example. One researcher is useful, but much of the work is currently spent searching literature and collating and correlating
the existing research. The internet has now enabled lay persons to do the same thing with the use of google and simply doing a mathematical combination of all the relevant keywords.
A layperson could not understand all of the texts thrown up by the correlated keyword searches but could certainly cut down the amount of time spent on this by said researcher.
If that job were automated further so that some semi intelligent tool could correlate successfully and pull up everything specific to what the researcher is looking for,
then they could spend more of their time on the experimentation. The experimentation itself can be speeded up significantly also by automation.
So does that put the researcher out of business? On the contrary. It increases productivity *massively* and the pace of scientific progress will jump by orders of magnitude.
Now could we somehow find some basic ways to plug in laypersons to this process? Indeed we can and we have. There is a game that currently exists for folding proteins.


ClarenceComments said...

I don't think you are correct about this, because I AM convinced that "strong AI" (not quite self-aware but able to do complex tasks and basic research) is just around the corner.

Here's one example:

Add in more and more capable, cheap, and self-repairing computers and robots and you get a very bleak future for human labor.

In the future, about the only useful thing that the average human will be able to do will be , in my opinion, personal services. And you only need so many of them.

This doesn't mean that everyone is necessarily going to be poor or enslaved (though it could if some of the "peakers" and "doomers" are right or if certain sociopathic oligarchs get control somehow), but I think it does mean that our economy in the next 15 to 40 years is going to need a major reconstruction.

Things go well: guaranteed incomes, localized mass production, possibly stuff like Marshall Brain talks about in his robotic future articles.

Things go poorly:
Oligarchs controlling vast armies of serfs/slaves and getting most of their income from automated production they control , Or some AI really does go intelligent and for logical or illogical (fear) reasons decides to enslave or cull humanity.


DB said...


I think you missed the point entirely. I'm saying that the kind of example you quoted will vastly increase the capabilities of the researchers and vastly more basic research will be done before. This is what I meant when I said "and research could be automated further". Moreover I'm saying that since it is in fact basic research itself that drives new products and it is new products that drive creative destruction which leads to new paradigms, vastly greater productivity and new industries I'm also saying that this will in fact make us all *richer* rather than poorer.

To illustrate how I'm right: think how often each of us who are somewhat intelligent and somewhat creative have had an idea that would help either your personal situation, someone else's or even be potentially mass market. Then what stopped it is the lack of funds for basic research. Basic research is expensive. If the cost of basic research is brought down significantly then the scale of turnover of new paradigms will speed up yet further.

Personally I call that beneficial.