A Niche in the Library of Babel

Paul Krugman and Business Week have recently referred to a grim new paper  – “Is US Economic Growth Over?” by Robert Gordon.  He’s an economics professor at Northwestern University, and notes that real growth peaked in the 1960s, and has been tailing off for some time now:

“British data for 1300-1870 come from Broadberry et al. (2010) and are ratio-linked to Maddison’s U.K. data through 1906. U.S. data are based on NIPA Table 1.1.6 back to 1929 and are ratio-linked to annual GDP from Balke-Gordon (1989)”

He doubts the reality of continuous economic progress, and instead says that there have been three distinct Industrial Revolutions:

1. 1750 to 1830: E.g. steam, railroads
2. 1870 to 1900: E.g. electricity, internal combustion engine, running water, indoor toilets, communications, entertainment, chemicals, petroleum
3. 1960 to present: E.g. computers, the web, mobile phones

What looks like continuous progress is the working out of the inventions of each revolution.   They take some time to really kick in, and that’s when growth happens.  E.g. automobiles were invented in the late 19th century, but didn’t have a broad effect until the 1920s, and didn’t become universal (at least in the US) until the 1970s.  Now there are more cars in the US (250M) than people over 16 (~240M), so more big growth is impossible.

There was little growth before 1750, and there’s no guarantee that there will be any in the future.  It’s already declining in spite of the Third IR.  That hasn’t had nearly the effects on ordinary life as the Second.   Cellphones are nice, but the difference between a cellphone and a telephone is minor compared to the difference between a telephone and a letter.   The Third IR did cause a growth bump in the late 90s, when PCs and the rising Internet really did make a difference, but that seems to have played out already.

The reason he gives is that the benefits of the Second IR were fundamental improvements in human quality of life, and can only be achieved once.   People can take care of sewage disposal once, and it’s done.    They can acquire easy and clean transportation,  or the banishment of night, or relief from farm labor, or freedom from the heat and cold of the seasons only once.   All these gains can be improved upon, and can be more widely shared, but at some point future gains just aren’t as great.   Electric lights can become more efficient, and safer, and cheaper and have better color, but the big change was from candles.

The Third IR added freedom from repetitive mental tasks, and greater access to information, but those just aren’t as big a deal, according to Gordon.   I partly agree.  I’ve spent most of my career designing computers, but I’ll be the first to admit that they aren’t as important as indoor plumbing.    To drive this home Gordon proposes a thought experiment, which I’ll rephrase slightly. Which would you rather have:

1. A PC, and an outhouse and outside well?
2. Or books and a bathroom?

Fond as I am of reading and writing blogs, I prefer to go inside in the winter!

He also believes that even the present anemic rate of US growth is slowed by 6 factors:

1. An aging population
2. A stagnant educational system (i.e. graduation rates are no longer rising)
3. Rising inequality (Note that the above chart is for average, not median, GDP per capita – median looks much worse)
4. Global competition
5. Bad energy and environmental prospects
6. High consumer and government debt

Other countries may be affected less.  E.g. Canada is refreshing itself with lots of skilled immigrants, and Sweden suffers far less from inequality and debt.

It’s a bleak picture, but at least in my own field I don’t think it’s right.  Advances in electronics have NOT come in clumps; there has been something major in every decade.   Here’s a list off the top of my head for the last century, where I’ve assigned the decade to the time of wide use, and not just first invention:

• 1900s – radio telegraphy, electric street cars, movies
• 1910s – vacuum tubes, radio telephony, vinyl records
• 1920s – radio broadcasting
• 1930s – radar, movie sound
• 1940s – digital computing, television
• 1950s – transistors, tape recording
• 1960s – communication satellites, integrated circuits
• 1970s – microprocessors/DRAM/disks => PCs
• 1980s – computer networking, CDs, VCRs
• 1990s – Web, cellphones, GPS, digital cameras
• 2000s – digital video (DVDs & streaming), smartphones, electric cars

Invention has been constant, and shows no signs of slowing.

But what of his larger point, that the Second IR satisfied human needs in a way that the Third has not?   I think there’s growth here that isn’t showing up in the statistics.   Most people have seen a great expansion in freedom in the last 50 years.  The 1960s may have been great for monetary gain for white straight males, but minorities, gays, and women would much rather live in the 2010s.    Even the elderly are far better off – they’re healthier, more mobile, and more engaged.  The metric of dollars per person is not capturing this.    It can capture the relief of physical discomfort that the Second brought on, but not the relief of mental discomfort that we’re seeing today.   It should matter that people feel less oppressed.

Has Third IR technology aided this expansion of freedom?  Perhaps, because people become more tolerant as they are exposed to a broader range of experience.   Recorded music, television, and now the Web have made all of the world’s culture available everywhere.    This can be overstated, as Wired-style cyber-vangelists tend to, but to the extent that the Third IR has been helping this process, it has been improving the quality of life as much as the Second.

Cesare Candi harp guitar, ~1900. Yes, it can be built, but why?

Two years ago I was asked to look at the assets of a failed startup.    All its equipment and intellectual property were about to be sold for a pittance, so I was supposed to see what could be saved.  The startup had found a way to build a cheap  infrared camera, one that could see much deeper into the infrared than ordinary cameras.  A typical silicon sensor can see into the near infrared (a wavelength of about 1000 nm), whereas this could see down into short wavelength infrared (SWIR,  about 1600 nm).   There’s a lot of this kind of light around even on dark nights – it’s emitted constantly by a layer of the stratosphere.   They thought they could build a great surveillance camera, one that didn’t need any illumination at night.

It turned out that almost no one cared about that.   If you put a surveillance cam outside, you also put a light there.   The sort of places you really want to cover are truck yards, where it’s easy to steal a lot of stuff, and those always have lights.  The people who really care about covering dark areas are ones who don’t want to advertise their presence.  That’s basically military bases in hostile areas, and that’s not much of a market.

So they went under.  The technology was great – they had found a way to put a layer of germanium on a standard silicon sensor (germanium has a much lower bandgap than silicon, and so can sense lower-energy photons), so it was a simple add-on to existing tech.  The people had come out of Bell Labs, and one of them was an IEEE Fellow, which is a real honor.  I asked them what else you could do with such a sensor.  “Well,” said the former CTO, “it turns out that SWIR light goes right through healthy teeth, but is absorbed by the water in cavities or other damage.”

I almost fell over.  They had a way to do dental exams without X-rays, and they spent all their money on surveillance cams for Afghanistan.  What a waste!  “Wouldn’t that have been a better application?” I asked.  “Our CEO and VCs were committed to the surveillance idea, so we never pursued it,” said the CTO.  And now they can’t.

I’ve seen this happen lots of times.   Someone builds something really nice that really works, but its basic purpose was wrong-headed.   A few other examples:

• A wireless link capable of gigabits per second that’s meant to replace a $5 HDMI cable.
• A full custom chip for a proprietary disk interface (DSSI) that’s only going to be used on a few thousand workstations.  That’s about $1000 of design expense per chip for a chip that only costs $20.
• A fully-programmable parallel image processor that spent all its cycles doing standard operations like video compression.   That stuff should have been done in a standard block, leaving the processor to do novel operations that would distinguish the product.
• A chip that can do the latest and greatest video compression algorithm (H.264 SVC), one that’s so difficult that no one can actually decode it.  Well, they could if they had the same custom hardware, but no one will buy that until video is available in this format, and no one will do that until everyone can decode it.  Chicken, meet egg.

The engineers involved in all these projects did their jobs well.  These projects were completed, and worked.  It was the people at the top who screwed up.   They asked for the wrong thing to be built.

That’s not always fatal.  I once heard Wally Rhines speak about this.  He was the manager of the group at Texas Instruments that built the first major digital signal processor (DSP) chip, the TMS32010, in the early 80s.  He’s now the CEO of a large chip design software company called Mentor Graphics.  DSPs are now a $10B product segment and are used in a vast range of applications, from cellphones to disk drives to audio.    Even way back then TI knew they had a hot product on their hands.  It was obvious what DSPs would be used for – speech synthesis.   Everyone would love to have talking cars!   TI had actually had a big success with the Speak & Spell, a toy that let you type a word and then hear it, and so knew that this was the wave of the future.   Ten years later they looked back at the top 100 applications of DSPs, and speech didn’t even make the list.  They guessed completely wrong, but the design was so flexible that it could be used for lots of other things.  Few projects get to be so lucky!