Search Episodes
Listen, Share, & Support
Listen to the latest episode
Subscribe via iTunes
Subscribe via RSS
Become a fan
Follow on Twitter

Support Us:

Please consider making a donation to help make this podcast possible. Any contribution, great or small, helps tremendously!

Subscribe to E-Mail Updates

Related Readings
  • Answers for Aristotle: How Science and Philosophy Can Lead Us to A More Meaningful Life
    Answers for Aristotle: How Science and Philosophy Can Lead Us to A More Meaningful Life
    by Massimo Pigliucci
  • Nonsense on Stilts: How to Tell Science from Bunk
    Nonsense on Stilts: How to Tell Science from Bunk
    by Massimo Pigliucci
  • Denying Evolution: Creationism, Scientism, and the Nature of Science
    Denying Evolution: Creationism, Scientism, and the Nature of Science
    by Massimo Pigliucci

RS 205 - Michael Webb on “Are ideas getting harder to find?”

Release date: April 1st, 2018

Michael Webb

This episode features economist Michael Webb, who recently co-authored a paper titled "Are ideas getting harder to find?" It demonstrates that the number of researchers it takes to produce a technological innovation has gone up dramatically over time. Michael and Julia discuss various possible explanations for why this is happening, along with several challenges to his paper.

Paper Co-Authored by Michael: "Are ideas getting harder to find?"

Michael's Pick: "Truth and Power: An Interview with Michael Foucault"

Edited by Brent Silk

Music by Miracles of Modern Science


Full Transcripts 

Reader Comments (14)

Oh, really the fresh ideas are very hard to find nowadays..Thus, the info provides for us a great variety of options.
Michael Webb seems like a really capable economist. As Michael points out, technologies that have measurable metrics have almost necessarily reached a point of diminishing returns. Mega flops per Second for a CPU and the number of transistors per square inch on integrated circuits (Moore's Law) relate to a now well established technology with obviously diminishing returns. Corn seed yield relates to a genuinely ancient technology that also has obviously diminishing returns. Many FDA trials amount to bureaucratic nonsense, and fail to represent any metric of actual progress. The PhD time and Patent Application times also have much more to do with bureaucracy than actual technological complexity.

As Michael mentions, a technology that creates a paradigm shift would have much more effect on a given industry. However, can anyone really measure the difficulty of discovering a technology that creates a paradigm shift? For instance, optical computing would probably far surpass an previous achievement from semiconductors, and a new food source could obsolete corn entirely. No method exists, however, to measure the difficulty or probability of someone making such discoveries.
April 2, 2018 | Unregistered CommenterJameson
Yet another example of why economic models are of limited use. Near as I can tell, the underlying assumption is that knowledge develops along a what might be called a factory model, there is a direct correlation between "inputs" and "outputs." So, there should be a direct relationship between the "amount of effort" and the "range of outcomes."

This maybe good economics if you are making some sort of widget. But, do we really believe that science is merely making widgets? What I know of this history of science, the history of ideas, is that ideas develop in fits and starts. And, there is no linear relationship between effort and outcome. There is no predicting where the next break through will come from. And, there is no predicting how those break throughs will effect either current practice and or future developments.

There is no evidence that streaming this process will work, no matter how rational the arguments for streamlining might appear. It is more likely that "streamlining" will in fact lead to a dismantling of the scientific enterprise.
April 5, 2018 | Unregistered CommenterKen Taylor
Super-intelligent AI to the rescue!
April 5, 2018 | Unregistered CommenterMax
The example of Moore's law is a poor one for this argument because it's an example of requiring exponential effort to achieve ongoing *incremental* improvement, which is the case in many areas when you're improving an existing solution rather than truly coming up with a new idea. Of course many new ideas are required to solve all the issues in stepping Moore's law forward, but in each generation you have already squeezed out all the obvious (and many un-obvious) ideas from all the previous generations, so it's not surprising that "oh, let's just double it again" becomes exponentially hard.

It would be much more informative i think to look at the explosion of ideas that follows at the heels of each opening of a new frontier. Some examples of this today would be mobile device applications where you could look at the number of games developed for iOS and Android, the new successes in Machine Learning (since AlphaGo) and the explosive growth of new applications of these technologies, and the many new ideas in medicine and Synthetic Biology that are being developed with each new discovery about how cells work.

Unique ideas are more difficult to find simply because the number of creative people has gone from a few privileged elite to potentially billions of people around the world.

I also think talking about "ideas" is hazardous, since in many areas ideas are worth nothing until you research/develop/implement them, so the economics of ideas are much more complex than what the image of "getting an idea" suggests. Even incredibly great ideas are very often ignored even when you shout them to the world, and until you at least develop a proof-of-concept to show people, you won't get much value out of the idea itself. Genius being 90% perspiration and all that.

Another great example of the modern explosion of ideas is in art, especially that created using computers, as the internet means that anyone in the world can be on an equal footing to compete in this field. It's instructive to look at a forum like the CG Awards at cgsociety to see the incredible diversity in artists and where they come from.

April 5, 2018 | Unregistered CommenterGavin Scott
Definitely depends on the domane. Transistor count & biology have stagnated. Space travel is lightyears beyond what it was 10 years ago, with far fewer people employed on the task.
April 5, 2018 | Unregistered Commenterlion
Actually, Michael Webb's model does work really well to predict streamlining. Furthermore, if we the number of scientists and engineers overall, don't we also increase the probability of a genuinely new breakthrough technology.

Max, great comment !
April 6, 2018 | Unregistered CommenterJameson
The second pick was “images of organisation” by Gareth Morgan.
April 23, 2018 | Unregistered CommenterPaul
This is strikingly similar to the work that Joseph Tainter has been doing since the 80s, though he relates the decreasing marginal returns to the ever increasing complexity born out of problem solving.

He believes it's a matter of energy economics and makes a system dependent on perpetual growth doomed to collapse.

Since our global economic system is dependent on exponential growth (by way of compound interest on debt/investment) and our political systems are dependent on global economic stability, we will absolutely require a new technological frontier to mine new ideas, resources and economic growth from or likely face a rapid reduction in complexity (a collapse).
May 7, 2018 | Unregistered CommenterGooseus
Informative topic on Rationally speaking.
May 11, 2018 | Unregistered CommenterJenny
Thanks for your information. It is very necessary for me.
May 28, 2018 | Unregistered Commenterbloxorz
Perhaps the issue that’s “holding us back” is that we simply lack the mindset and tools for combining insights across disciplines. That is, putting on and taking off glasses is an absurdly slow way to switch paradigms when one’s field of vision is a kaleidoscope of ideas — the answer is there in plain spice sight! Where does one get “glasses” to pull that vision into clear view and give us the starting point for writing it out in the languages of science so it can be tested and actually built?
August 1, 2018 | Unregistered CommenterRon Carlson
When I first ran across Webb's hypothesis and analysis in NBER, I had a real problem reconciling his decreasing number of Ideas at a time in history with an exponential increase in human knowledge, communication ability, and the number of existing ideas whose combinations make new ideas. The number of possible ideas is growing factorially with the total number of existing ideas which is much faster than even exponential growth.

The problem is in his definition of an Idea as buried in economic math. His ideas are really applied innovations where an idea passed through all the filters of physical reality, economic feasibility, market viability, financing, and then all the regulatory and standards limits while overcoming the objects of existing vested interests and become an economic reality. The rate-limiting steps in creating an economic Innovation are not "Ideas" in the common usage of the word.

I did a very long comment on his previous publication of this data. In the first part, I looked at some of my views on the problem of trying to understand how you can get a long-term exponential growth rate in the per capita economy when each sector of the economy is on a long-term sigmoidal growth rate curve and always has been. In the second part, I did two analysis of areas of the economy where I have direct scientific knowledge with Moores law and GMO crops. He appears not to understand the details of how technology evolves and how research depend upon thousands of other areas of research ideas and when you are on the spear point at the boundaries of human capacity, you have to add these secondary areas to your budget and that gives the appearance of decreasing total factor productivity that he claims. He didn't look at the ratio of the number of Ph.D. researchers in solid state physics (the heart of moors law) to experts in other areas like clean room design, hard UV light production. Instead of utilizing clean rooms from hospitals and biosecure facilities they have moved past their capabilities and are now feeding knowledge to hospital operating rooms and biosecure facilities. This doesn't mean that the solid state physics researchers are decreasing in total factor productivity, they just had to bring another factor into their accounting.
< >
August 28, 2018 | Unregistered CommenterDallas Weaver, Ph.D.
Using Moore's Law as an example feels incredibly wrong to me. The graph is only linear if you put it on a LOG scale. The real graph is rising exponentially, which means that that progress is accelerating rapidly. Every year its way faster than it ever was before. And it's not really about ideas.
September 26, 2018 | Unregistered Commentersilver

PostPost a New Comment

Enter your information below to add a new comment.

My response is on my own website »
Author Email (optional):
Author URL (optional):
All HTML will be escaped. Hyperlinks will be created for URLs automatically.