[p2p-research] Tick, tock, tick, tock… BING

[Paul D. Fernhout]

Kevin Carson wrote:
> On 12/9/09, Paul D. Fernhout <pdfernhout at kurtz-fernhout.com> wrote:
> 
>>  Marshall Brain was the first person I read who was really explicit, step by
>> step, about the link between automation and joblessness at all levels,
>> especially in Manna. It had been said before for a long time, but he really
>> seemed especially clear about it. And he makes clear an economic link in the
>> sidebar here, written around 2002:
>>   http://marshallbrain.com/robotic-nation.htm
>>  """
>>  The "Jobless Recovery" that we are currently experiencing in the U.S. is
>> big news. See for example The Mystery of the 'jobless recovery':
>>   "Consider these facts: Employment growth at the moment is the lowest for
>> any recovery since the government started keeping such statistics in 1939.
>> The labor force shrank in July as discouraged workers stopped seeking
>> employment. The number of people employed has fallen by more than 1 million
>> since the "recovery" began in the fall of 2001." [ref]
> 
> Paul, as interesting and well-argued as your posts on the subject are,
> I'm afraid I'm extremely skeptical that automation is the primary
> reason for the jobless recovery, or that that will be a primary cause
> of structural unemployment in the near term.

There are two issues. One is employment as a "lagging indicator" after GDP 
growth, in part for reasons of cycles and stuff like lags in retraining and 
people making new businesses in response to "creative destruction". But, as 
I said, it is lagging further and further behind, to the point where it is 
going to catch up with the next recession. :-) And part of that is from what 
I think is "limited demand".

The second big area is how automation has affected jobs, and I think that is 
growing in importance, and this "Great Recession" is probably the tipping 
point on it. Or if not this one, certainly the next in another five years or so.

More by Marshall Brain:
"Marshall Brain Makes Splash With Structural Unemployment Message at 
Singularity Summit 08"
  http://ieet.org/index.php/IEET/more/brain20081026/
"""
Robot lovers and outsourcing opponents could soon have something in common: 
the fear that their jobs are at stake.
   In the future robots will take over many tasks performed by American 
workers today potentially leading to increased unemployment, says Marshall 
Brain, founder of How Stuff Works and author of e-book Robotic Nation.
   “In theory we should all be able to go on a perpetual vacation as robots 
do all the work,” Brain told attendees in a presentation at the Singularity 
Summit in San Jose. ” Instead because of the way the economy is structured 
right now, when robots arrive it will have devastating effects on all of us 
because there will be so many unemployed people.”
   The implications of an increasingly automated economy could be dire for 
society unless we restructure our economy, he argues.
   While productivity has risen over the year, worker compensation has not 
kept pace and there is increasing concentration of wealth, says Brain.
   Developments in robotics means that technologists could be creating a 
second intelligent species, claims Brain. “So far no credible evidence to 
indicate that there is more than one intelligent species in the universe, 
which is us,” he says. “But that changes with the robots.”
   Computational power has significantly increased in the last two decades 
and is growing exponentially. In 1992, computers could perform about 300,000 
operations per second. By 2022, it is likely to jump to a trillion 
operations per second and by 2042 it could be a quadrillion operations per 
second, predicts Brain.
   “A $500 machine that can do that, whenever it happens, combined with 
vision and natural language processing could change how we look at robots,” 
he says.
   Potential applications of robots could then be in use as automatons in 
fast food restaurants, transportation, education, construction and retail 
among other areas. We will have robotic cashiers, robotic stocking, 
sweeping, help and cart retrieval at Wal-Mart,” says Brain.
   To deal with that version of the future, he suggests society should 
redesign the economy to get the benefits of automation.
   His solution? Spread the benefit of productivity to everyone by breaking 
the concentration of wealth, increase pay and reduce the work week. Sounds a 
lot like socialism, doesn’t it?
   The idea provoked a question from attendees. When industrialization first 
occurred there were fears of massive unemployment which never panned out. 
Why will the integration of robots into the workforce be any different?
   “We didn’t create a second intelligent species 150 years ago,” says Brain.
   “Now we are doing that with intelligence that will get better and better.”
"""

> For one thing, replacing humans with industrial robots is the kind of
> expensive, capital-intensive investment that the old mass-production
> industrial core was prone to in its heyday.

As above, it is getting cheaper and cheaper. In the 1980s, we had a GE Robot 
Vision system in the robot lab I managed. It was a huge box almost the size 
of a refrigerator. With addons, it must have cost approaching US$100K, which 
was worth twice as much then with inflation. Now, you can do the same with 
some US$30 webcams and a $300 laptop (well, and maybe a USB electrical 
interface card). So, a drop in price by almost a factor of one thousand. 
Granted, the laptop and webcams are not industrially hardened, so you might 
have to add a little more considering the industrial nature, but it is at 
least a factor of one hundred reduction in cost overall, and the computers 
now could do so much more (that was black and white, now you can do color, 
that had a few vision system operations, now you can do much more). In many 
ways, there is just no comparison with how much more functional 
off-the-shelf computer systems can be now compared to those old wire-wrapped 
special purpose circuits. I've seen advertisements now for shoe-box sized 
devices with cameras you just clamp on your production line to do parts 
inspection and train them by example for probably (not sure exactly) only 
several thousand dollars. This is work that it used to take an employee to 
do. Now you can replace many of those inspectors with a small box.

Robots themselves are dropping in cost. Just look at the toy section of the 
store. For $50 you can get things that walk and fly. You can get more 
complex robots for US$2000 or so (the old Aibo robot dog). Humans are 
expensive considering health insurance, retirement, training, and so on -- 
an employee typically costs in the USA about US$100K to US$300K depending on 
what they do. Even at US$100K a flexible robot is cheaper than the person if 
it is low maintenance and can be used for just one year. And robots are 
cheaper than that.

The free market is going to drive this. We are one by one passing tipping 
points for all sorts of things. People are either replaced entirely or the 
jobs are deskilled so cheaper workers can do the work.

> But the old manufacturing corporations are deliberately eschewing
> investment in capital-intensive factory machinery, for the reasons
> described by Piore and Sabel, and instead outsourcing production to
> small-job shops in distributed supplier networks.  Those job-shops may
> be more technically sophisticated than the factories they're
> replacing, but it's the kind of sophisticated machine design that
> expands the machine's usefulness as a craft tool in the hands of a
> skilled worker.

Even without robots, 3D printing, ShopBots, improved CNC machines, better 
design, better materials, and so on are all reducing production costs.

Basically, you can re-engineer many processes by better designs and better 
materials to remove most of the craft aspects. And 3D printing is going to 
do much of the stuff previously done as craft. Within twenty years most of 
that work will be, at best, a human interacting with some sort of complex 
robot-like device, where the team can output what now would take ten or a 
hundred people.

Manufacturing is, effectively, being automated away in the near future (even 
as, like agriculture, maybe 1% of the workforce might remain in it as 
"managers" of the robots). It's still interesting to think about how to help 
that process along, but people are actively doing it to cut costs. And this 
recession is probably inspiring a lot of companies to take the next steps to 
cut even more costs.

The next frontier is really services. And those are harder, because they are 
often less structured. You can control almost everything in a factory for a 
factory robot in a way that you can't for, say, a home-visiting nursebot. 
Teleoperation may help with that though; for example, one visiting human 
nurse could, from a central location, use a tele-operated nursebot in each 
person's home to do the nurse's rounds without travel time, probably 
doubling or tripling productivity, especially if more and more they could 
hand-off routine activities like changing bandages and maybe even run 
several nursebots at once. So, possibly, one nurse could do the work of ten 
nurses with telemedicine. You may think that is sci-fi, but the US military 
is actively developing related technology, and using some of it now as are 
others.

Just one example of what is under development:
http://telemedicinenews.blogspot.com/2009/04/darpa-advancing-programs.html
"""
According to Colonel Ling, DARPA’s Trauma Pod Program places surgical teams 
where and when they are most needed. Basically, the program takes medicine 
to the patient by not using humans and integrates tele-robotic and robotic 
medical systems. The initial phase has successfully automated functions 
typically performed by the scrub nurse and circulating nurse. These 
functions are now performed by semi-autonomous robots working in 
coordination with the tele-robotic surgeon. So far, surgeons have performed 
complex surgical procedures on a simulation mannequin by operating a robot 
using the Trauma Pod operations console. Two procedures were performed 
without the aid of a scrub nurse, and the system correctly changed tools and 
dispensed supplies with 100 per accuracy.
"""

Or what is used now:
http://telemedicinenews.blogspot.com/2009_04_01_archive.html
"""
Today in Alaska, telehealth has had had a huge impact and is used to 
identify many health issues much earlier. The network handles 12,000 cases 
per year, has 248 sites with care not only being delivered in Alaska but 
also delivered to other places such as Greenland, Panamanian prisons, and 
schools in Ohio. Both speakers emphasized that much has been accomplished 
and now telehealth provides care for many health issues including post 
surgical follow-ups, wound closure examinations, and post cochlear implant 
assistance. In addition, travel time is greatly saved since 8% of specialty 
consults are prevented with 20% travel time reduced for primary cases. This 
has saved AFHCAN $3.5 million in treating 43,800 patients. For every dollar 
spent by Medicaid on reimbursement, $7.95 is saved on travel costs.
"""

Cost savings also implies less jobs, if you assume demand is limited.

> More than anything, I think structural unemployment results from
> excess industrial capacity and the lack of sufficient demand to run
> industry at capacity.

That is obviously part of it.

But I think that is the older part of it.

The newer part is just the same thing that happened with agriculture going 
from 90% of employment to 1% of employment. Manufacturing in the USA has 
been going from 30% fifty years ago to 12% or less now, and is just going to 
keep going down. That just leaves employment in services. But services are 
much more volatile because you don't really need most of them. So, we saw a 
quick domino effect in the recent recession that surprised many economists 
in how fast it happened, as people just cut out so many services (plus cut 
back on luxury manufactured goods and high end food products like meat).

Retraining time is another part of it as people shift industries. But the 
fact is, many people who like manufacturing and have been doing it for a 
lifetime are not going to be happy about, say, a job as an insurance agent 
or mortgage broker (if we even need more of those).

> And even for mass-production industry, it's of questionable economic
> benefit.  For one thing, a robot hand with sufficient manual dexterity
> to perform a wide range of delicate operations is great, but until
> there's also a robot with the processing capability of the human mind
> that's capable of the craft skills and judgment to run a range of
> general-purpose machinery, the way a Japanese worker does on the shop
> floor, robots are probably the least thing folks in Detroit have to
> worry about.

Well, it's true people can be flexible. But they also get tired and bored. 
And they generate lawsuits. And they can quit after you trained them. 
Automation is a safer investment in that sense. And most products can be 
redesigned to be produced by automation. We really have not tried that hard 
so far. Flexibility is one issue -- needing to change production quickly, 
which is interlinked with other aspects of economics. But as software 
improves, and sensing improves, and digital design standards improve, it 
gets easier for automation to change what it does quickly.

I think the biggest difference is you are assuming automation will do it 
all. But even if it just amplifies one person on the shop floor, so they can 
do ten times as much, by basically being the manager of robots instead of 
doing stuff themselves, then that gets rid of a lot of jobs. So, in that 
sense, the issue is an improving partnership between humans and machines. 
But, that means that we need less humans in the factories.

> The Japanese deliberately chose Taichi Ohno's reinvented version of
> craft production with flexible machinery and a skilled work force,
> over robotization and deskilling of the work force.

Well, considering they now are struggling with an aging population and 
predictions of a labor shortage due to that, how long will that last? 
Besides, what you are describing is essentially a social policy of "make 
work" in a sense. How long will that last in a market system? I'm not saying 
crafts are bad -- I agree it's great for people to use their minds and hands 
together to do things. But now we're talking broad social policy (or EF 
Schumacher's Buddhist Economics), not free market economics.

--Paul Fernhout
http://www.pdfernhout.net/