What is technology? The informational layer of everything

Here is a puzzle from economist Paul Romer for you to consider:

I’ll claim that there’s a recipe out there that you could use to just assemble carbon, oxygen, and hydrogen atoms, and that if you just use the right recipe to put them together, it will make a factory that will be smaller than a car, that will be mobile, that will seek out some renewable input, that will convert [the input] into some chemicals that humans want, that will be self-healing whenever it gets injured, that will maintain sterile conditions, and that will even make a replica of itself when one generation breaks down. So the question is, could you really put something together out of carbon and hydrogen and oxygen and atoms that would be that sophisticated?

So, is it possible? Romer:

If you describe that just in the abstract, many people will think it’s not possible. But then you point out to them that it already exists in the form of a dairy cow. And then get people to think well, what is the recipe or set of instructions for a dairy cow? Well, it’s really just a DNA sequence that has a list of instructions for how to assemble raw materials together.

That’s from From Poverty to Prosperity, a series of interviews on growth economics and innovation by Arnold Kling and Nick Shulz. I bring it up not just because it’s a neat example, but to make a related point. What is technology? Here’s Merriam Webster:

1 a : the practical application of knowledge especially in a particular area : engineering 2 <medical technology> b : a capability given by the practical application of knowledge <a car’s fuel-saving technology>
2 : a manner of accomplishing a task especially using technical processes, methods, or knowledge <new technologies for information storage>
3 : the specialized aspects of a particular field of endeavor <educational technology>


Romer uses the metaphor of “recipes” to describe technological change. And the definitions above emphasize the role of knowledge. The point I’m making is that technology is in a sense just information. When we invent a new technology, we come up with a new recipe – a new set of instructions for combining things – that yields something useful. In this sense, information is at the center of technological change.

I bring this up first because I think it is clarifying. It helps us discard the notion of technology as being simply “high tech”, and leaves us with a broader, but to my mind much more useful and interesting definition. But there’s another reason I mention it. We’re seeing a fundamental restructuring of how we produce and distribute information. Yet too often we tend to think of “information” as this separate sphere of the economy. There are information goods, and there are regular goods. Software is in the former, cows are in the latter.

But the Romer example show that that is wrong. Once you start thinking of a cow as a technology, you realize that there is what we might call the informational layer of everything. And that means that the restructuring that we’re seeing that allows Wikipedia and Linux to be produced via non-market collaboration suddenly seems far less limited. It’s not just limited to a subset of goods that are “information goods” because everything is an information good.

Now, that’s not to say that we have the technology to make new cows (yet) or that if we could that it could be crowd-sourced. But I think it does give us a good way of looking both at technology and at the broader potential for new models of information production. There’s no such thing as an open source cow today, but that doesn’t mean there never will be.

Paul Romer and The Great Stagnation

Today I downloaded Tyler Cowen’s new e-book The Great Stagnation for $4, along with Amazon’s Kindle app for Android. It’s got both the publishing and economics blogospheres all aflutter so I’m looking forward to the read. But what kind of blogger would I be if I didn’t comment first, read later?

Cowen had an op-ed a few weeks back in The New York Times that laid out his basic claim. Here’s one bit:

The numbers suggest that for almost 40 years, we’ve had near-universal dissemination of the major innovations stemming from the Industrial Revolution, many of which combined efficient machines with potent fossil fuels. Today, no huge improvement for the automobile or airplane is in sight, and the major struggle is to limit their pollution, not to vastly improve their capabilities.

Although America produces plenty of innovations, most are not geared toward significantly raising the average standard of living…

…Will the Internet usher in a new economic growth explosion? Quite possibly, but it hasn’t delivered very good macroeconomic performance over the last decade.

Cowen is brilliant, and I’m looking forward to the longer version of his argument, but I want to consider a more optimistic take on our economic moment by another brilliant economist, Paul Romer. What follows is from Romer, interviewed by libertarians Arnold Kling and Nick Schulz in their book From Poverty to Prosperity (which I can testify is valuable and interesting even to non-libertarians).


…this may be the most important question in human history: why have we had technological change and why has it been speeding up over time? …

…it may be inherent in the process of discovery that the more we learn the faster we can learn. It’s a notion that was captured by Newton when he said that he could see farther because he stood on the shoulders of giants. That was the first model that I tried for the speeding up phenomenon, that the more we learn the more we can learn…

…So it’s that kind of analysis, thinking of ideas as recipes – really, instructions for combining together small numbers of physical objects – that persuades, I think, anybody who works through the logic that the number of things we could have even tried up to this point in time is so small compared with the number of things that are possible, that we’re just extremely early in this discovery process. For as far as you want to project into the future of humans, we won’t run out of new things to discover. And as I conjectured in the beginning, it may even be that the more we learn about this process – the science of DNA, the science of materials, or our understanding of quantum mechanics – the more we learn about this stuff, the better we get at finding new, ever more valuable mixtures…

…The second-round answer, which I think may actually capture more of the truth, is that it may get easier to discover as you learn more things, or it may not, but what we’ve done is created better institutions over time so that we now exploit the opportunities much more effectively than we used to.

I believe Cowen wants to challenge this both on whether technological change has in fact been speeding up, and whether the technological changes we’ve seen recently are the same kind of economic game-changers we saw in the 20th century. I don’t know what role he would attribute to institutions. Which model is more accurate? I have no answer to that. But the Romer interview at least offers an alternative conceptual lens to Cowen’s. (You can read the full interview with Romer here.)

Cowen also argues that America has eaten up a lot of the “low-hanging fruit”, in education for instance, as described in the clip below:

It’s an interesting point. But once again for a more optimistic take I’ll point to Romer:

We’re going through this shift in the economy where the fraction of human effort that goes into actually physically rearranging things – bending metal, doing manufacturing, and so on – is going down over time and the fraction that’s going into discovering the right formula or recipe is going up over time. And that’s a really good thing, because all of the value really comes from finding the new recipes. If you picture the innovative activity of one hundred years ago and you think of it as U.S. Steel, most of the workers there were involved in literally bending or melting metal – doing the physical rearrangement – and a relatively small number of people were coming up with, say, better ways to extract iron from ore…

…So we’re going through this transformation where a larger and larger fraction of the labor force is engaged in problem-solving, sifting through possible ideas, and a relatively small fraction of people actually stamps out the pills or bends the metal.

(That bit combined with some other comments offers an interesting take on Julian Simon’s work, both approving and disapproving of his work on population and innovation.)

So in education the question may be whether we can innovate fast enough to outpace declining marginal student aptitude. Romer’s model suggests that as more minds spend time thinking about how we educate, rather than spending that time grading tests, passing out papers, etc., that innovation could speed up. Cowen’s suggests that each student we attempt to get through school is likely to be harder than the last. Will more minds working on better educational recipes mean better education, and therefore more minds devoted to finding other useful recipes? Or have we already eaten all the low-hanging fruit?

Perhaps I’ll have more to say after I finish Cowen’s book.