A doctor's appointment illustrates what's gone wrong with technology, and how we might fix it.

A little while ago, Vox journalist Matthew Yglesias injured his knee and needed weeks of rehab. It led to a small epiphany, which he recounted during a podcast discussion of electronic health records.

We have apps for booking doctor’s appointments, and you can learn about knees and treatment methods using your phone from the comfort of the sofa. Compared to 15 years ago, that’s wonderful. But (Yglesias continued) if you hurt your knee in Star Trek, you’d go to the sickbay and Dr Crusher would use a little machine that goes bleep bloop bleep and then your knee is fixed. That machine doesn’t seem to be on the agenda, and everyone is focused on making minuscule improvements here and there. Has something gone wrong with our ambition?

Dr Crusher examines a patient on the Enterprise

There’s a broader context for that question. We’ve become used to continuous improvement in living standards, and textbook economic models assume that GDP growth will continue forever, but it might stop soon. The debate about what might happen next was started by Robert Gordon’s working papers in the late nineties, which he expanded into a substantial book a few years ago.

You don’t need to read it, thanks to a productivity boost provided by Bloomberg’s Tom Orlik. He pointed out that you can get the entire gist of the argument by considering three sitcoms.

First, Little House on the Prairie. Rural life in the 19th century was horrible. There was constant physical labor; people were socially isolated; their diets were poor; they were poorly educated; there was no medicine.

Moving forward a few generations brings us to Happy Days. Everything has changed: they have electricity, appliances, lights, refrigerated food. They have access to modern medicine, and modern fashion. Everyone has a car, and enough leisure time to hang around at the milk bar all the time. They have Fonzie. The daily life of ordinary people has been utterly transformed.

One more step forward, to Friends. Things are a little different, in areas like mass tertiary education (Ross is a palaeontologist) but our heroes’ lives are on the whole pretty similar to the previous sitcom. The big discoveries have been discovered, and can’t be discovered again. Technology is now about grinding out small improvements, in barely perceptible steps, leaving our children’s lives essentially the same as ours.

Between the first two sitcoms, there’s an obvious list of differences that are impossible to avoid. To tell the difference between the second pair, you have to squint a bit. If that continues into the future, then no amount of arguing about productivity measurement will save us from the fact that GDP growth has virtually ended.

Orlik’s choice of Friends is smart, because it avoids assessing the impact of the Internet and contemporary tech. Gordon, for his part, has a superficial and inaccurate understanding of modern AI, perhaps because he first started thinking about these issues 20 years ago. Nevertheless, he makes a strong point about GDP and market output. Computers caused a surge in measured productivity in the nineties, which has since petered out. If the Internet was really bringing fundamental changes to our working lives, then output per hour would be surging. It’s not.[1]

Startups will not save us from technological stagnation, and in fact will make it worse. As people have learned how to run them well—which essentially means using the Build-Measure-Learn principles of the Lean Startup—they have come to excel at exploring the immediate next steps in technology space. Startups build and test things that have a decent chance of working or failing within about six months. There are good reasons for this, but at its worst it generates nothing but small twists on existing successful products. We’re making Uber for lawyers, Pinterest for programmers, etc etc. These are all risky bets for the individuals involved, but on a historical scale, they’re simply too conservative.

I think there are two things that have a reasonable shot at restarting the growth in living standards. One is large tech companies.[2] Across the private sector, only something the size of Google has both the scale and the incentives to experiment with genuinely risky long-shots. For example, Waymo took almost a decade of development and testing before it was ready to be shown to users.

Another possible source of substantial change is Missions, a concept developed by Mariana Mazzucato. The idea is that public agencies can set ambitious goals for innovation, which eventually translate into concrete projects. For example, something like this on the topic of climate change:

Or about public health[3]:

I think Mazzucato’s theory is tremendously exciting. It opens up a whole new perspective on industry policy, and at the same time revitalises the stale idea of ‘public-private partnerships’, while building leadership and human capital in the public service. And I love the way it criss-crosses industry sectors.

As Aaron Timms observed, “Many governments aspire to foster innovative businesses, but exhibit little flair for innovation themselves in their thinking about how to attack that task”. The Missions framework seems like it brings something genuinely new, and in conjunction with the R&D at large tech firms, offers the hope that we can restart GDP growth and make our children’s lives materially better than ours.

[1] Could this be due to mismeasurement? Measuring productivity in a modern services economy is conceptually tricky. The technical details are complex, and immensely boring. Even so, it doesn’t leave enough wiggle room. If contemporary tech was having an impact anywhere in the market economy, it would be freeing up resources which would show up in well-measured sections.

[2] Author interest disclosed.

[3] Both those diagrams come from a report that Mazzucato wrote for the EU last year.