Tyler Cowen and Ben Southwood write,
To sum up the basic conclusions of this paper, there is good and also wide-ranging evidence that the rate of scientific progress has indeed slowed down, In the disparate and partially independent areas of productivity growth, total factor productivity, GDP growth, patent measures, researcher productivity, crop yields, life expectancy, and Moore’s Law we have found support for this claim.
I think of this in terms of what factors might cause science to exhibit increasing returns or diminishing returns
The most obvious source of diminishing returns would be the “low-hanging fruit” story. The problems that remain to be solved are really hard: human biology; human psychology; climate science; etc.
Another possibility is that scientific genius is limited. If you throw more money at mediocre scientists, you don’t get any results.
The most obvious source of increasing returns would be lower cost of new scientific tools. The cost of genome sequencing, for example.
Another source of increasing returns would be better communication technology.
If you want to speed up scientific progress, the tools that are available are probably institutional. What changes in institutions can we make that would either mitigate diminishing returns or promote increasing returns?
As the authors point out, science has become bureaucratized, with much of the funding coming from government grants that impose a lot of costs in terms of process and perhaps impede creativity. Having more science funded by patronage, with wealthy donors providing funding with fewer strings attached, might be an improvement–unless you think that the bureaucratic requirements add value and rigor to the process of choosing scientific projects.
Scientific progress may have slowed, but I think total factor productivity has little to do with science. Most productivity improvements are not drawn from science but rather result from innovative people trying to solve business problems.
The two giant policy opportunities for improvement:
– Free Market Health care: This is a well worn issue in this circle, but there is enormous potential here, to pave the way for higher returns in biology and health care. Some free market reforms are happening outside of the political sphere, but better policy would go a long way.
– Higher education reforms. All STEM knowledge/training is run by government owned or quasi-government owned universities. Notice that even math/computer training is tied to physical location: the University of Michigan, the University of Georgia. Something like mathematics, it shouldn’t matter where you are located, even what country you are in, and education shouldn’t be owned by government.
Necessity is the mother of invention.
I love this paper, which is both weird and wonderful since I had such low expectations for the paper, bordering on dread to even read it.
My low expectations stems from my strong disagreement on Cowen’s stagnation hypothesis. The paper’s conclusion is still one of stagnation but it cleanly lays out the argument and carefully discusses the major counter-arguments. In short, this is a beautiful framework on which to a build a discussion about economic growth and innovation.
Cowen and Southwood underestimate the enormous impact of web (~1995), mobile, and cloud (~2007 for both). Although they try to quantify the value of the Internet, they do so only in terms of customer facing front-end value and ignore the scale and efficiency benefits of the cloud back-end systems. This is best described with Facebook’s acquisition of WhatsApp for 19 billion for a 55 person Company with 500 million users. You can argue the value of each user but the web/mobile/cloud innovation is what has enabled 55 people to service 500 million adoring customers. Productivity?
This innovation carries over to the discussion about patents and scientific papers which doesn’t capture the breakthrough contribution of engineers like Jeff Dean at Google and James Hamilton at Amazon/Microsoft. Even a back of the envelope of the impact of Jeff Dean’s Google published papers that resulted in the Hadoop copycat technologies and the valuations of Cloudera and Hortonworks should be eye-opening.
Cowen and Southwood’s discussion of Moore’s Law correctly described the dimensions on which stagnation, or a slowing, has occurred but missed the new dimension of performance-per-watt for mobile devices represented by Intel being eclipsed by Arm based CPU’s and new startups like the Nuvia are threatening Intel’s dominance in the cloud data center. Intel is not failing to innovate, they are failing to innovate fast enough. It is hard to argue stagnation as this epic battle unfolds.
The bureaucratic reqs are NOT for more creativity, they are to eliminate fraud & cronyism, which are reduced but not eliminated.
More “Prizes” for results would be one way towards better results, with partial funding for partial results.
Another diminishing return story is the increasing amount of GDP which goes towards non-productive people, especially the retired but also the sick. This is a great civilization benefit, to have retired folks in middle class comfort, and live secure in the knowledge that such retirement will be available, but one of the prices paid is less push towards extra effort for most folks.
One of the most disappointing tools we don’t really have is reliable voice-to-text virtual assistants. Have you tried any lately? Directions in the car, voice commands for entertainment, these are highly limited. But related to this is the fact that more young people have terrible senses of direction, or map-reading skills. As “easy-living” tools become more available, more folk become more soft. In body and in mind.
There was a movie which very well showed soft people living soft lives. Animated, from Disney – the people were living in space, in easy life-assisted comfort. Mostly plump / fat, being driven by scooters from their sleeping quarters to the pool or wherever else they wanted to go. Wall-E. The movie was about the robots, not the people.
In Central Europe, they could get some Western movies, including those which showed modern stores filled and stuffed with goods to buy. Movie after movie had store after store — but none of the movies was about how much stuff was in the stores. Some commies tried to say the movies were “lying”, but most believed the commies were lying, about that and much else.
If the goal of most voters is to have an easier life, which seems likely, there will be less push for more scientific advances. Along with low-hanging fruit, there is a culture about facts & ideas that supports more or less scientific truth.
Did you see the ACLU tweet that included “Men who get pregnant and give birth are men”? With these kinds of changes, culture is going backwards far faster than new truths are going forward.
That is an interesting perspective, seriously. Natural Language Processing (NLP) is one of the technical areas that has innovated at pace far beyond my wildest predictions. Starting with Siri and Apple’s inability to extend its early lead and now has been eclipsed by Alexa and the Google Assistant. The growth in Amazon’s Echo ecosystem at first blush contradicts your observation but I think the discrepancy comes down to how we underestimate how hard it is to build good NLP and NLU (Natural Language Understanding) technology that feels “reliable”.
The NLP innovation continues at break-neck speeds with break-throughs with on-device NLP (the current systems do the NLP processing in the cloud) that leverages silicon Deep-Learning/Machine-Learning. I guess the question is when/if they will bridge your perceived “reliability” gap.
“Having more science funded by patronage, with wealthy donors providing funding with fewer strings attached, might be an improvement–unless you think that the bureaucratic requirements add value and rigor to the process of choosing scientific projects.”
—
Bot science is taken over the process with automatic funding via crowd sourcing. Also the bots are able to use pay for use contracts which get customers and developers on the same page from the start. If you notice, the fintech field is spread all over, no longer do we concentrate in one place. Also notice the financial bots are scaring the bejesus our of central bankers. The is the bot advantage, and it is not stopping.
Humans used to fly in commercial aircraft across the oceans at supersonic speeds. Not any more.
People actually reached the moon and walked across it, 50 years ago. There’s some talk about doing that again in another decade or so, but not much money for it. Notions of mining the moon and asteroids for useful resources and building human settlements on the planets haven’t advanced beyond the talk stage since Apollo days.
Used to be undersea bases humans lived in for weeks or even months – “Sea Labs”. The last of those was closed over 20 years ago. The CIA made a pretense of sea bed mining back in the 1960s when they had a scheme for dredging up a sunken Russian submarine; we haven’t made any progress toward actually realizing this idea.
Back in the 1950’s there was talk of drilling a hole thirty miles deep to reach the layer where earth’s crust meets its mantle. The Russians got to about 20 miles down, if memory serves, but nobody’s gone deeper in the last 60 years.
Back about 1985 a fellow named Eric Drexler came up with an idea — nanotechology — about building smaller and smaller devices until industrial processes could be performed at the level of molecules. Still a nifty idea, but not one industry or governments really wants to spend money on.
Schools of fish seem to be getting smaller. The fish in them seem to be decreasing in average size. We’re utilizing more varieties of fish, true, to make up for the over fishing, but at some point this reaches a dead end. Maybe we ought to worry a bit about improving ocean fishery productivity? Ho ho ho.
Got any ideas about dealing with global warming? Got any ideas about convincing people global warming is an impending danger? The science is so certain after all!
Where are our flying cars?
This is progress?
mike shupp:
supersonic airliners: inefficient for commercial air travel, great for signalling.
moon walks: ok for science (once), great for sabre rattling, great for televised events.
sea lab: irrelevant with advances in autonomous vehicles
sea floor mining: Probably not cost effective. Doesn’t deep water drilling count as a far more technical achievement? If Andrew McAfee’s “Less from More” is correct and we are dematerializing, how can space and/or sea floor mining make economic sense?
earth crust boring: boring seems to be advancing pretty well if the Eurotunnel is considered advanced tech. I’m not sure the cost/benefit analysis works well for mantle exploration compared to remote imaging/sounding technology.
nanotechnology: nano machines seems to be a bust but microfluidics is impressive and viral delivery systems seem to accomplish some/most of the promise of nano machines.
overfishing: we don’t seem to do well with tragedy of the commons type problems but aquaculture, including microscopic autotrophs looks promising.
global warming: it is real but I’m skeptical of the models that claim it will end in catastrophe. Useful models might be a good first step.
flying cars: they are called airplanes and helicopters.
I don’t know if Vaclav Smil has done the type of progress/stagnation analysis of technology that we are talking about, but I trust his still of analysis more than the signaling-centric approach of Peter Thiel and his flying car tagline.
So, advances in aeronautics? Not worth having.
Manned space exploration and settlement? Pointless.
Undersea settlements or scientific bases? Irrelevant.
Undersea mining and other sea utilization? No economic sense.
Geology as experimental science? Ruled out by cost/benefit analysis.
Nanotechnology? A bust.
Overfishing and ocean utilization in general? A tragedy of the commons.
Global warming? Not really a settled science.
Flying cars? Already here, a century ago!
This is progress?
Exactly. Wonderful progress by objective measures. Manned X seems to be a common theme but unmanned/remote aeronautics, space exploration, undersea science, and geology are advancing rapidly. The path may not be what you imagined but the destination is the same. The rate of progress is only disappointing if you believe in the predictive power of science fiction writers or that all technical measures must increase continuously. I’m not disappointed with the progress of the canoe, it is a near perfect technology.
Measures of global warming and increasing CO2 concentrations are very real but the predictive climate models don’t pass the sniff test, in my opinion, but that is a separate discussion. The consensus is that the science is settled. My view is non-consensus but technically grounded.
This is a direct link to a mp4 file with a talk Freeman Dyson gave on Living Through Four Revolutions that finishes with a question about the progress of space science.
Yeah, we don’t need guys clumping about in space suits to bring back data from space. I et that, but so what? I mean, that’s science, which is sort of the topic, but there’s another issue which is what the human race does with science which is quite different. Would you argue that the whole point of Columbus’s journey to the New World, from the standpoint of history, is that it finally proved the world was round? Or maybe that human beings lived outside of Europe and Asia and Africa? That that was all? That that should be all, that no one but lunatics has ever imagined that human existence might have been diverted from one track to another by Columbus’ discovery? That we would have had World War I and II with the same participants and the same outcome without the influence of the New World? That the sort of life realized by hundreds of millions of immigrants over the last two centuries was never affected by their movement to the Americas? That the same inventions would have been made, the same sciences developed in just the same way if we’d all had the sense to stay in the Old World?
I disagree. I see technology as stagnating, and I strongly suspect stagnant technology and science lead to stagnant futures, for the human species and the individuals who comprise that species.
Columbus did not set out to prove the earth was round. Freeman Dyson, in the linked talk, specifically speaks about the social side of space exploration which was very much what the Russian space program is; the equivalent of an international sporting event. He says that has its place but should not be confused with science. He also tells a funny story about the International Space Station and its “science” narrative.
The consensus around here seems to support stagnation. I’ll happily take the non-consensus side, especially if Freeman Dyson is a fellow traveler.
Ah! You include yourself in the Dyson Sphere.
low hanging fruit gone,
only so many great minds, and
time/effort to simply understand state-of-art increases.
cranking public funds up? expect diminishing returns
I think that the inability of other countries to pick up the slack suggests that only a few nations have the institutions to push innovation forward in the “macro” sense that Mokyr talks about. That means problems in social institutions, the increased unwillingness to have unvarnished meritocracy and to take risks, the significance of increased democracy promoting redistribution over invention in the developed world, changes in cultural attitudes towards disruptive tech, and incompetence of government at the most elite levels of otherwise effective state bureaucracies in non-democratic regimes are enough to limit the pace of innovation even if the low-hanging fruit thesis is incorrect.
Whether the cultural, social, and institutional forces are headwinds or tailwinds, having lots and lots of elite-level prolific geniuses is going to help.
We have slow and crude level technology for how to make lots of geniuses – or lots of anything – which is via human hubandry, and which is how China made tall, world-class basketball players.
And we are approaching the capacity to genetically engineer lots of geniuses, for example, by spell-checking out genetic load to get the best complementary set of alleles.
The “ethics” debate on this is going to go like it did with Nukes. Every country with a capacity to dedicate nation-state-level resources to the problem will say, “If we don’t do it, they will, and we’ll fall behind in the race, so … ”
And so, whichever countries already have a good genetic depository of genius samples, have plenty of economic and scientific resources, and a geopolitically “serious country” governing class with an iron will to power, and little to no moral squeamishness about these things … oh, who are we kidding. “And that’s how China wins the future.”
Sounds right.
You don’t mention the 30-60 million “excess” Chinese now 18-55 who can’t find wives. This thanks to the one-child policy plus cultural desire for boys.
China today is looking an awful lot like Japan of the early 1930s — with increasing militarism combined with feelings of superiority. Plus clear Chinese equality, if not superiority, in lots of tech.
Tho it’s not at all clear China can win the future if another Great Leap of Culture occurs, perhaps with the slogan
Communist Billionaires are not Real Communists
Death to Billionaires – No False Communists …
Hong Kong, Uighurs; even Tibet — not yet the normal way a Country of the Future treats its own people.
Still, the Chinese seem to be world leaders in population surveillance & control, sure to be popular among other wannabee despots.
And, truly fearful if they avoid an internal revolution.
Is scientific progress any different than the rest of the culture? Look at our stagnant politics, increasing cost and time for public works, sluggish legal system, bureaucratic medical system, etc. Some of that is aging of the population (esp. the huge number of aging boomers), but some of it just seems to be complacency and pessimism. We’re barely trying to improve some of these systemic problems.