The per capita rate of innovation (2004)

http://brucegary.net/book_GE/Chapter_20.html

by derwiki • 12 years ago

17 23 12 years ago

dang• 12 years ago

We changed the submitted title ("Global literacy rate over time (and more)") to what seems to be the author's phrase for the key idea of this piece. Better title suggestions welcome.

derwiki• 12 years ago

Sure. I posted mainly because I thought literacy rate over time was interesting, but wanted to post more context than just the graph. Looks like others found the rest of the page interesting.

api3• 12 years ago

I've had an intuitive sense for a long time that the innovation rate has fallen dramatically since the 1950s. What we're doing now really feels like small incremental improvements to technology that was all invented prior to 1970.

Has anything actually been invented since 1970?

Relevant:

https://www.youtube.com/watch?v=yJDv-zdhzMY

protonfish• 12 years ago

The Internet and everything in it?

ori_b• 12 years ago

Was created as ARPANET in the 1960s.

zenogais2• 12 years ago

It's important not to confuse the rapid production and reproduction of novelty with innovation.

api3• 12 years ago

All invented and in many cases first prototyped in the 60s.

sp332• 12 years ago

Alan Kay asked Stack Overflow the same question. https://stackoverflow.com/questions/432922/significant-new-i...

api3• 12 years ago

The responses are not encouraging. They're either incremental improvements on things invented before 1980 (often before 1970) or the poster is misinformed and the idea pre-dates 1980.

The WWW is actually a 1960s idea. Watch Engelbart's demo.

Sometimes I wonder if the problem is the massive growth of free market ideology (market fundamentalism) since 1970: Ayn Rand, the Reagan Revolution, libertarianism, etc. The free market is heavily biased toward extracting pre-existing value and things with short time horizons. It rarely spends money on very speculative R&D, and the culture that it begets is a hustler culture of quick-to-market, quick-to-exit, etc. Things like the Engelbart demo are the product of very smart people with very rough, flexible deadlines sitting around and thinking a lot. There's no way to monetize that. A related problem is that invention itself is hard to monetize except via patents... the innovator often loses in the market to fast followers.

In machine learning terminology we could say that the free market is a greedy hill climbing algorithm prone to converge on local maxima and unlikely to cross fitness valleys.

darkmighty• 12 years ago

1) The source, "Asimov's Chronology of Science and Discovery" is obviously biased towards older innovations. It's difficult to judge a priori the impact of current innovations. (and it's pretty outdated by now)

2) Breakthroughs are simply very impactful innovations -- of course earlier innovations are going to have an advantage: it's like saying "The total number of descendents per capta has been dwindling since the dawn of ages. We're doomed!". Claiming the rate of breakthroughs should remain constant is ridiculous.

3) Additionally, inventions are information. They're cumulative and knowledge instensive. We can't expect them to occur at a constant rate simply because it will get harder and harder to learn eveything required to invent some really new technology/scientific discovery. I like to think of human knowledge as a chunk of information that we compress and pass along. There's only so much we can compress, and our brain hasn't really changed significantly to accomodate and absorb more information faster. The result of course is specialization, and again specialized innovations by definition have limited scope.

4) To claim we had no specialized breakthroughs recently is even more ridiculous. Take the computing aspect of engineering alone (and math) and you probably get more small breakthroughs than there were inventions at all in the 18th century. Telecommunications also developed completely in the last few decades.

tdaltonc• 12 years ago

Where is this data coming from? And why are there no confidence intervals on his measures?

ignostic• 12 years ago

I do hope everyone realizes this is silly.

Everything in this comparison relies on the quality of data. The "number of innovations" data comes exclusively from a book by Isaac Asimov with under 1500 entries. To make some sort of grand pronouncement about a date range for the end of humanity based on Asimov's very subjective and unscientific qualification of what makes an innovation is absurd.

It's interesting that we think incremental improvement matters less than a brand new idea. Even steel tools built upon prior knowledge, and took decades to get right. Machines were incremental improvements from manually powered systems, which gradually evolved into the complex machiens we have now.

Some of these changed may feel like evolutionary innovations rather than revolutionary ideas, but that's how technology develops. It's easy to look at past innovations from the Greeks, because we see how the technology went from 0 to amazing. But we do it in retrospect, and speaking in terms of centuries, which makes innovations appear to spring out of nowhere more than they actually did.

gwern• 12 years ago

It is a bit silly, but if you use a variety of encyclopedias and source texts with various temporal cutoffs (not just this one, but Simonton's work or Murray's _Human Accomplishment_ where he ignores any item after 1950 while drawing on published books), you do pretty consistently get the late 1800s as the apex for per capita.

77777772• 12 years ago

I agree and disagree. :) I think it is silly only if you assume it is incorrect. But, although not from a robust source of data it still could be the correct result. And in that case, I couldn't call it silly. But who knows, the word "innovation" is subjective, so you could probably argue against any dataset, pretty easily.

That being said, I agree that identifying something as an innovation is going to be heavily weighted to past accomplishments. At this point, innovations in theoretical physics are so complex that I can hardly begin to really grasp the importance of discoveries beyond my niche and I myself have a Ph.d. in physics.

homerowilson• 12 years ago

Yeah. Got to love that 10th order polynomial fit...

77777772• 12 years ago

The attempt to give a formal treatment to data as subjective as "innovations" is probably not much more than a fun exercise. And, from that perspective, I did have fun reading it. :)

nazgulnarsil• 12 years ago

The full book is somewhat interesting: http://books.google.com/books?id=RNhLAW615QgC&pg=PP1&dq=inau...

He paints an excessively pessimistic picture IMO.

Nevermark• 12 years ago

The estimated 50% chance of human population collapse between 2140 and 2600, seems reasonable.

Our power to destroy ourselves with weapons or accidents, or use biological and information technology to productively enhance or replace ourselves is growing at an exponential rate.

Any scenario where humanity reaches the year 3000 without either having destroyed ourselves, or enhanced ourselves beyond recognition, seems unlikely to the extreme. The only scenario I can imagine where radical technological change does not result in a radical change to the human condition is if a totalitarian government managed to enforce a no-progress civilization for hundreds of years.

I am marking the years 2140-2600 on my calendar so I can remember to look back and review the results of this prediction using my post-human thinking apparatus. Hopefully I didn't extinct myself.

selimthegrim• 12 years ago

For a second there I read, "enhanced ourselves beyond _redemption_..."

danelectro• 12 years ago

A lot of innovation is individual as opposed to institutional.

Motivated, technically qualified individuals have always been "out there", but now they are lost in the wash with so many highly degreed average people dominating both institutional and venture efforts.

The traditional "great man" innovators are more out-of-place than ever in this landscape, where quantity of technical people is the limiting factor for shareholders requiring rapid growth.

If structures allow a return to having the quality of the innovators prevail over the quantity, more of the outstanding unrecognized individuals may be able to contribute again like what was seen 50 to 100years ago.

Some shareholders would not be able to accept the longer incubation periods and slower exponential growth necessary, but not all shareholders are the same. It could happen.

Ask yourself, if DaVinci, Newton, Edison, Einstein or somebody like that were unknowns but wanted to innovate for you, or wanted you to support their efforts (which might already be in progress), which of your exclusionary tactics would eliminate them in the first or second round?

I've said it before, technology has progressed but it's clear that PEOPLE used to be more advanced than they are now.

The per capita rate of innovation (2004)

Related Stories