Summary: US growth is slowing when it should be accelerating as we shake off the effects of the crash. The possibility of a fifth year of slow growth strengthens fears of stagnation like that afflicting Japan since 1990. Yet there’s good reason to suspect that a new industrial revolution has begun, potentially generating incredible new wealth — if we manage the process well politically. Which future is correct? Both of them. {1st of 2 posts today.}
“There are decades when nothing happens; and there are weeks when decades happen.”
— Attributed to Lenin.
Contents
- We’re becoming Japan.
- We’re accelerating to take off speed!
- The 3rd industrial revolution?
- Conclusions
- For More Information.
(1) We’re becoming Japan.
Compare growth in per capita GDP of America and Japan. We following in their footsteps.
- US since the crash: 1.4%/year (2010-2014: 1.7, 0.8%, 1.6%, 1.5%, 1.6%).
- Japan before the crash: 2.0%/year (2003-2007: 1.2%, 2.6%, 1.9%, 2.1%, 2.3%).
This is a big story. It’s called secular stagnation (see the posts describing this theory, with links). Readers of the FM website have known about this since 2010, with more details given in 2013 and even more last year. Larry Summers introduced it to the world in 2013. It’s still controversial, as seen in Ben Bernanke’s rebuttal this week (see Larry Summers devastating reply). I suspect time will prove Summers is correct.
Also, Japan has still not pulled out of their stagnation, despite the 3 arrows of Abenomics.
(2) But we’re accelerating to take off speed!
No, we’re slowing, as shown by the Atlanta Fed’s GDPnow forecast for Q1 of zero growth. Yet the experts remain hopeful for a better year than 2014. The Fed foresees growth in 2015 of 2.3% – 2.7%. But then in September 2012 they expected growth in 2015 of 3.0 – 3.8%. The February survey of Professional Forecasters shows a median expectation of 3.2% for 2015 (note this calculates annual GDP slightly differently than does the Fed). I expect they will be disappointed, yet again.
(3) What about the 3rd industrial revolution?
I and others have said that the 3rd industrial revolution has begun. Some even expect the singularity to occur during the lives of some alive today, as the rate of tech evolution goes vertical — especially in the breakthrough fields of AI, nanotechnology, genetics, and fusion (see the details here).
But these things take time to develop. Despite frequent claims, tech is not developing faster than in previous revolutions. For example, the below graph shows the rate of internet adoption was slightly slower in its first decade than that of radio and television (but faster than that of the telephone, which required more infrastructure). For details see Gisle Hannemyr’s “The Internet as hyperbole: A Critical Examination of Adoption Rates“. The start date assumed for the internet is 1998.
But once these developmental processes start, they move with astonishing speed. America’s cities were wired for electricity and telephone over 2 decades. For example see this photo of Niagara Falls in 1904, with each factory generating its own power. In 20 years they were all using grid power — as were tens of thousands of factories across America.
There are scores of technologies in such early stage roll-outs — 3-D printing, new health care tools, robots for manufacturing and services, electric and self-driving vehicles, software automating countless jobs — now appearing on a small scale. It will take years for their effects to shake the economy. But the revolution will arrive.
(4) Conclusions
The coming years of stagnation will put immense social stress on America, especially as the 1% continues to boldly exploit their increased wealth and power (social but not political stress unless we organize and act together in response; flocks of sheep don’t have politics). Beyond that we probably will experience a shift from stagnation to rapid growth, with political and social effects that we can only dimly foresee. I suspect it will be as turbulent as the years of dizzying change in the west between 1870 and 1920. That history has been sanitized for Americans, but it was ugly and bloody (see Prof Eric Loomis’ this day in labor history series).
We have time to prepare. Only large-scale organization will prevent the 1% from taking most of the gains from the new tech, as they have captured most of the slower productivity gain during the past 30 years. Will we use this time wisely?
(5) For More Information
If you liked this post, like us on Facebook and follow us on Twitter. See all posts describing how the 3rd industrial revolution has begun. And posts about the theory that we’ve entered a period of secular stagnation.
- Start here: 50 years of warnings about the next industrial revolution. Are we ready?
- The promise and peril of automation.
- Looking at America’s future: economic stagnation, or will computers take our jobs?
- 50 years of warnings about the next industrial revolution. Are we ready?
For deeper analysis see these books:
- The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies
by Erik Brynjolfsson and Andrew McAfee.
- Average Is Over: Powering America Beyond the Age of the Great Stagnation
by Tyler Cowen.
- Rise of the Robots: Technology and the Threat of a Jobless Future (2015)
by Martin Ford.
The Singularity is a con job, a classic example of a degenerating research paradigm like alchemy or phlogiston. AI is a scam going nowhere, nanotech has been defined down to triviality until it’s nothing like K. Eric Drexler’s Engines of Creation, genetics research applications have stalled out because of the problem with identifying metabolic pathways and the issues of complex multiple alleles and epigenetics, and fusion power has always been “20 years away” for the last 60 years.
One of these groups of things is not like the other group of things.
FM’s second paragraph is reasonable and accurate — but it describes current technologies, not pie-in-the-sky singularity fantasies like superintelligent AI, or Drexlerian nanotech assemblers that can take apart a lump of coal at the molecular level and reassemble it as a diamond, or gene therapy that can create mutants with an IQ of 500, or pipe dreams like too-cheap-to-meter fusion power.
Let’s go into details: AI is a scam because scientists have proven unable to define or objectively measure intelligence. The Flynn Effect shows that lQ tests have such large socially constructed elements that they are, in practical terms, useless. AI as a research paradigm depends on building machines that exhibit human or superhuman intelligence — but if a quantity like “intelligence” can’t be defined or measured scientifically, obviously the research paradigm is at a dead end before it starts. Black children raised by black families but adopted by whites before age 7 show no difference in IQ from white kids, while black children raised by black families after age 7 show a big IQ gap. ItaIian immigrants in the 1930s showed a Iarge IQ deficit compared to immigrants from Norwegian countries. And so on. A vast mountain of evidence shows that IQ tests exhibit a large social component. Whatever IQ tests measure, it’s not just pure problem-solving ability, but some form of socialization as well.
So much for the AI pipe dream.
Studies show that even dropping the cost of generating electricity to zero wouldn’t have much economic impact on American society. The costs of distributing and modifying electricity from high-voltage to low voltage at the end user would make electric power about as expensive as it is today. The main impetus for fusion power involves the dream of independence from imported oil — but current nuclear fission pebble-bed technologies or thorium breeder reactors allow that without failed dead-end pipe dream fantasy technologies like fusion power.
So much for the fusion power pipe dream.
Nanotechnology in Drexler’s sense is physically impossible because of stiction at the molecular level caused by Van der Waals forces. Dreams of rod-and-ball molecular mechanical computers or molecular memory devices turn out to be physically impossible. This means that micro-robots cruising through your bloodstream repairing your cells are also physically impossible. As a result, “nanotechnology” has been redefined successively down until it now means “small particles,” which has no connection with K. Eric Drexler’s original definition of nanotechnology as molecular mechanical computers.
So much for the nanotechnology pipe dream.
Gene research has enjoyed great success. Turning this research into practical therapies has hit a brick wall, however. Complex diseases like cancer involve the interaction of many gene loci — there is no one single “cancer gene.” Ditto diseases like arthritis, lupus, and so on. And turning off one cancer gene often has huge effects on other genes, so complex as to be impossible to predict, since genes not only code for proteins, they regulate other genes. To make matters worse, creating a therapy for a disease requires that doctors interfere intelligently in the metabolic pathway of the proteins expressed and regulated by genes. Because it’s so hard to figure out these metabolic pathways, gene therapies constantly exhibit the problem that they work beautifully in the test tube and fail completely in live organisms. Worse yet, epigenetics means that genes change the proteins they express an the metabolic pathways they regulate according to the environment, which means a complex interaction twixt environment and genetics that researchers haven’t even begun to unravel.
So much for gene therapy.
Thomas,
Your bizarrely sweeping dismissals of these technologies — especially weird since you don’t specify a time limit to them — are similar to countless similar ones in the humor bin of history. Airplanes and space flight are impossible! Nuclear weapons are impossible!
All we know is as a general rule the people dismissing these technologies know a lot less about them than the scientists and engineers working on them
The general rule is that people overestimate change in the short term and underestimate it in the long term.
It seems to me that FM and TM should first clear up definitions before putting forth arguments. For instance, I have seen nanotechnology refer to everything from very small molecular complexes with catalytic-like effects to microscopic engines à la “Fantastic Voyage”. The former kind is already being deployed with success, the latter is pure unattainable SciFi in my opinion.
Same problem with AI: it means everything from automated pattern recognition, which works fairly well in specialized domains, to expert systems. The history is littered with expert systems that gave excellent results in medicine, network configuration, etc, even superior to human specialists, enabling cost and time savings, but that for many reasons (frame problem, knowledge acquisition bottleneck, etc) were abandoned despite their benefits.
As for fusion: after many decades, research has made no substantial progress towards an operational prototype able to generate more power than it consumes for any significant period of time. What I have read about past and current fusion research indicates that the approach favored till now (and still being pursued with that huge international fusion center being built in France) just cannot scale to achieve a steady state. When I was a teen, fusion was touted as delivering energy “within 50 years”; fusion proponents have not changed their tune — it will still deliver plentiful energy 50 years from now. Their track record is rather unconvincing, but fusion sure keeps high-energy physicists occupied.
guest,
It is the rule with comments, that people focus on tidbits while ignoring the main points.
The point about the breakthrough tech is peripheral — or less — to the post. In the relevant time horizon — concerning the next decade or two — none of these will have any significant economic impact. I mentioned it to provide a larger perspective framing the debate about secular stagnation vs a 3rd industrial revolution.
If people wish to discuss the singularity, I suggest going to the post discussing it in detail: Has America grown old, and can no longer grow? Or are wonders like the singularity in our future? — rather than a post where it’s one sentence out of a thousand words.
The subject of the post is of extreme importance to America, and probably to many of the people reading it.
I can’t speak for nanotech. I would note in regards to nuclear power that it took little more than a decade to go from Fermi pile to the first commercial plant. On the other hand Tokamak have been around since the mid 50’s and the we are still stuck at the experimental stage. Granted, the priority was not the same but the amount money and brain power thrown at the problem over the decades has still been quite significant. Quite simply the physical process is very hard to control and bar surprises it will require several decades of further study and experimentation while the reactors themselves will be very expensive to build and operate. We are talking about the second half of this century for something commercial and even then they will probably require public propping up.
I would not be over optimistic about experimental fission designs either. Extensive testing, certification etc. take time in such a field and it may well turn out that there are unforeseen issues. Then you have to sell it to the public.
Speaking of which public opposition against experimentation and practical application of biological knowledge, from vaccines to OGM to animal tests, is spreading like wildfire and is already being codified into law in some countries.
Both guest and Marcello are correct here, while FM is (rarely for him) dead wrong.
Words matter. The definition of words matters.
Guest remarks:
Yes, that’s exactly the problem here — with sweeping SciFi hype like “Some even expect the singularity to occur during the lives of some alive today, as the rate of tech evolution goes vertical — especially in the breakthrough fields of AI, nanotechnology, genetics, and fusion” FM clearly implies something like the Singularity generating shockwaves of disruption in our society. But the reality remains that the kinds of technologies FM describes in his second paragraph (3-D printing, new health care tools, robots for manufacturing and services, electric and self-driving vehicles, software automating…jobs) are mundane incremental advances that will clearly not produce a “3rd industrial revolution.” Software won’t automate hairdressing. 3-D printing won’t print out brldges.
FM cannot on the one hand invoke the Singularity in paragraph 1 of his article, then blithely dismiss the Singularity in his comments with the casual assertion “If people wish to discuss the singularity, I suggest going to the post discussing it in detail…rather than a post where it’s one sentence out of a thousand words.” The entire thrust of his current article is that SciFi whiz-bang technologies will produce a technological and economic Singularity. But this is vacuous hype, and I’ve been hearing it since the 1950s — superintelligent AI! Fusion power! Micro-robots in our bloodstreams! Domed arctic cities! Flying cars! It’s all garbage, classic Heinlein-type tripe.
FM falls into the classic Heinlein fallacy when he claims “The general rule is that people overestimate change in the short term and underestimate it in the long term.” This claim can be quite true or utterly false — depending on the exact time frame in which you apply it.
Example: “The general rule is that people overestimate change in the short term and underestimate it in the long term,” is true from the period 1900 to 1970. First airplane to 1st moon rocket in just 67 years! But the same rule turns out be completely false from the period 1970 to today. Following that rule from 1970 onward, we would confidently expect manned space missions to have reached all the planets in the solar system by now and the first starship to be a-building today. Instead, manned space missions stalled out after the moon and no starships are even theoretically feasible today, or in the foreseeable future. This projection is a failed and faulty exponential forward extrapolation of a trend which actually shows a sigmoid curve, for solid physics reasons. The rocket equation is brutal and sharply limits our physically possible maximum specific impulse and thus our payload-per-unit-mass and rocket thrust.
Likewise, Wirth’s Law (software gets slower faster than hardware gets faster) has killed off hard AI, and as guest notes, “history is littered with expert systems that gave excellent results in medicine, network configuration, etc, even superior to human specialists, enabling cost and time savings, but that for many reasons (frame problem, knowledge acquisition bottleneck, etc) were abandoned despite their benefits.” Expert systems turn out to be fatally brittle and insanely expensive and time-consuming to engineer…aside from the blunt fact that no one really knows what intelligence is, and the brutal reality that whatever biological systems are, they don’t work like computers.
The confusion of terminology here is pernicious. FM explicitly invokes the Singularity, when when he gets serious knowledgable criticism of these pipe-dream SciFi fantasies from commenters, FM backs off and switches to mundane technologies like “3-D printing, new health care tools, robots for manufacturing and services, electric and self-driving vehicles, software automating countless jobs.” 3-D printing can’t print out laptops, new health care tools won’t empty bedpans, robots for manufacturing and services can’t replace janitors, electric and self-driving vehicles are limited to very short range trips at a max of 30 mph, and “software automating countless jobs” can’t replace a waitress.
Confusing sigmoid growth curves with exponential growth curves is a big mistake, and trying to move the goalposts when you get called on it is an even worse mistake,
Thomas,
That is quite a reading FAIL, in almost every way.
I am not a fan of the technological singularity theory — over any relevant timescales (I.e., a few generations). I did not “invoke” it ( whatever that means), it was one sentence in a long post, and it is quite irrelevant to this theme of this post.
As for the technologies you so casually dismiss — robots (I.e., the next wave of automation of industry and services), 3-D printing, etc — you are as usual quite *confident* that almost every expert in this field is wrong and you are right. Possible, but I doubt it.
Pingback: Celebrate our bright future on New Year’s eve! – Weather Brat Weather around the world plus