Thoughts in Between

by Matt Clifford

TiB 100: How to design a new ARPA; Tyler Cowen on talent allocation; how the internet is changing public intellectuals; and more...

Forwarded this email? Subscribe here. Enjoy this email? Forward it to a friend.

What might a new ARPA look like?

Apparently Dominic Cummings’ (previous coverage) personal motto during the UK election campaign was, “Get Brexit done... then ARPA” - referring to creating a UK version of the US agency that had such extraordinary R&D output in the decades after 1945. The question of how to design such an institution is both interesting and important. No government plans have yet been published, but Policy Exchange, the think tank, just issued an excellent collection of essays on the topic. 

I recommend the whole thing, but I want to focus on two points that seem to have broader implications. First is the importance of a “venture customer” - i.e. an organisation that stands ready to buy the products of innovation. The authors emphasise the role played by the US Department of Defense in the original ARPA and speculate about possible UK equivalents. It’s a difficult problem. In my day job I observe that large organisations often lack the capability to buy from startups. I suspect we’d be net better off with fewer corporate venture capitalists and more corporate venture customers. 

Second is the importance of what we might call “meta-risk”. Accounts of ARPA always emphasise its unusual risk tolerance - more like a venture capital firm than a government agency in its preference for magnitude over frequency of success. But one major and underrated lesson from ARPA is that we need more risk tolerance not just within institutions, but in how we design and implement institutions themselves. There are, I think, many important institutions we're missing out on because of meta-risk intolerance.

How to get the young and ambitious to innovate

Tyler Cowen has an excellent new column on talent misallocation, which, as long time readers know, I believe is one of the most important problems today. Cowen’s thesis is that “task complexity” is increasing, which means it takes longer to accumulate the skills needed to perform the most interesting jobs. As a result, ambitious young people are drawn to jobs like consulting and finance that pay well in financial and career capital, but "don’t reward (or allow) adventure” - which makes the whole economy less innovative.

Cowen sees tech as the exception to this “high insurance value, low innovation” equilibrium. I’m not so sure. I worry the opposite: you can now be so richly rewarded for not taking risk in big tech (see, e.g., slide 44 in this presentation) and for not really innovating (see the now classic Jeff Hammerbacher line), that tech seems a more dangerous talent suction machine than any vampire squid for precisely the people whose skills are most needed to speed up technological progress.

The good news, of sorts, is that pay in big tech is now so high that, for people with the right skillset, only the possibility of truly astronomical outcomes is tempting - and that is only found in entrepreneurship. The even better news is that if you can normalise entrepreneurship for the most ambitious people (as we’re trying to do at Entrepreneur First), you create the downside protection that McKinsey et al provide as standard: failure costs less if you’re surrounded by people who can hire you (or fund you again).  

How does the internet change public intellectuals?

Tanner Greer has an interesting new essay on why the “shelf life” of public intellectuals is so short. He takes as his canonical example Thomas Friedman - New York Times columnist and much-parodied quoter of taxi drivers. Friedman, Greer argues, has been intellectually fallow for over a decade, after a productive period around the year 2000. Greer believes this is the norm among public intellectuals, partly because of the way the mind develops and partly because they cease to be exposed to new milieus.

I’m interested in how the internet might change this. It introduces at least three mechanisms that might affect how intellectual careers work. First, distribution is now much less dependent on gatekeepers. To a great extent, Friedman the public intellectual was created by the NYT; today, winning a slot in a mainstream media outlet is often recognition of an audience gained elsewhere (see, e.g., Tyler Cowen, Noah Smith or Nate Silver). Second, new monetisation mechanisms for intellectuals exist like Patreon (where, incidentally, you can join me in supporting Greer’s work) and Interintellect.

Third - and perhaps this is too optimistic - the internet enables the possibility of more accountability for pundits. Philip Tetlock’s superb work shows that political experts make poor predictions, but pay no reputational penalty. The first two mechanisms point to a market for intellectuals that’s far more responsive to demand and therefore, perhaps, to getting things right. Unless, of course, that’s not what we really want... 

Quick links

  1. Don't panic (about the wrong things). The most grounded coronavirus analysis I've found so far.
  2. Maybe it's because I'm (not) a Londoner? Is there a wage penalty for regional accents?
  3. Google Maps-based performance art. How to fake a Google Maps traffic jam.
  4. Time for some game theory. How to design (fun?) video games that label images at scale
  5. Meaty topics. The rise and rise of veganism.

Your feedback

This is the 100th issue - and pretty much the second anniversary - of Thoughts in Between. I'm quite surprised what an enjoyable and valuable part of my life it's become, so thank you for reading.

As an anniversary present, I'd love it if you'd forward it to a friend or two. And it's always great to hear from you: feel free to get in touch on Twitter or by hitting reply.

Until next week,