TiB 129: Authoritarian regimes and AI; the scenius of Silicon Valley; "Optimal Kickstarter"; and more...

Welcome new readers! Forwarded this email? Subscribe here. Enjoy this email? Forward it to a friend.

We’ve talked a lot about the relationship between new technologies and types of political regime. South East Asia-based analyst “Pondering Durian” has a new post this week on “Technology-Governance Fit", which riffs on Peter Thiel’s line, “Crypto is libertarian; AI is communist”. One of their main conclusions is that authoritarian countries like China have a natural advantage relative to the US in AI.

I’m skeptical. First, I buy Henry Farrell’s argument, discussed previously, that there may be systemic problems with the training datasets available to authoritarian regimes. Second, I disagree with the idea, which Pondering Durian mentions in passing, that advanced AI will obviate the need for markets to allocate resources. We looked at the problem with this argument here, drawing on Cosma Shalizi’s superb essay on the topic. Third, it just seems there’s a lot of evidence, both civilian and military (great thread), that the US retains a lead in AI. 

I do think, as we’ve discussed before, that China’s ability to co-opt top machine learning talent to national priorities is likely its biggest source of advantage relative to the US. But a new report by CSET suggests that even that may be less pronounced than I had thought. That’s not to say that the Technology-Governance Fit framework is without value. But I suspect it’s in biology, above all genetic engineering, rather than in AI that the West may find its values most at odds with pushing the boundaries of science. 

Economist Dietrich Vollrath has a new and excellent review of Christopher Leucyer’s 2007 book Making Silicon Valley. It touches on a number of favourite TiB themes, particularly the causes and history of economic growth and the importance of talent. If you share my interest in these topics, it’s worth your time.

Vollrath notes the similarity of Lecuyer’s explanation for the success of Silicon Valley and economic historian Joel Mokyr’s thesis on the origins of the industrial revolution, above all the centrality of networks of talent:

This may remind regular readers of the idea of “scenius”, or communal genius, discussed previously here. 

It’s an important question whether it’s possible to build a “remote-first” scenius. During the height of the coronavirus lockdown, I wrote about the potential benefits of building tools and infrastructure that would allow talent outside major hubs to participate more fully in innovation. The internet has already done a remarkable job of enabling like-minded people to build relationships that transcend physical distance. If Vollrath (and Lecuyer) are right that “the network preceded the breakthroughs, not the other way around”, that’s a very positive sign. 

We’ve talked before (most recently a couple of weeks ago) about the importance of public goods. I believe the world suffers from massive under-supply of public goods because the most obvious provider - the state - often can’t or won’t supply them. Is there a way for non-state actors to step in? 

One interesting possibility is Alex Tabarrok’s idea of Dominant Assurance Contracts, in which entrepreneurs seek pledges that, if a minimum funding threshold is not met, are not only zero-cost (à la Kickstarter) but actually receive a payout from the entrepreneur. This has lots of positive effects described in the linked post (If you like this, Tabarrok wrote a whole book of similar ideas). I was reminded of this this week because Matt Clancy wrote a great post on an “Optimal Kickstarter” design, based on this paper by Glen Weyl et al, for providing public goods.

The idea is that you collect pledges as normal, but a philanthropic funder backs individual pledges according to a formula that, under certain assumptions, results in the optimal provision of public goods. If this seems very esoteric, note that this sort of thing is already being trialled in the world of Ethereum. In a world where the largest philanthropists are increasingly likely to be maths/engineering aficionados, I suspect this kind of mechanism may turn out to be much more consequential than it looks today. 

Thanks for reading Thoughts in Between. It’s very much a weekend project, as my day job is pretty busy, so I’d love your help to grow the readership. The easiest way is to forward this to a friend or, even better, share it on social media.

Lots of newsletters get stuck in Gmail’s Promotions tab. If you find it in there, please help train the algorithm by dragging it to Primary.

I love getting email from readers. Feel free to hit reply or message me on Twitter.

Until next week,

Matt Clifford