TiB 101: How Confucius would design an AI; why everyone is so incompetent; can hedge funds disrupt venture capital; and more...

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

Incentives, institutions and the decline of competence

Last week’s Iowa caucuses were a fiasco. Riffing on this, Claire Berlinski has an interesting piece on the decline of competence in public life. Her argument is that post-war US business and political culture was heavily focused on problem solving (see this long but excellent essay for more), but this fell away because it was not formalised in education and other institutions. As Berlinski says, given the overlap between people who run elections and those who look after nuclear weapons, this matters a lot.

It’s hard to measure “competence”, but there is general feeling that Western state capacity has fallen relative to, say, China or Singapore (see, e.g. China’s much feted rapid hospital building following coronavirus). What do they do differently? I have two hypotheses. First, Singapore at least gives politicians much more direct (and incentive compatible) ways to monetise competence than the US or UK. Perhaps this explains why so many US policy makers get law degrees, a fact Berlinski bemoans: it’s a reliable way to generate income when you’re not in power.

Second, I suspect that the political rewards for competence in the West are muted by polarisation (see here for an interesting discussion). If elections depend mainly on tribal affiliation, you’ll get little credit from the other side for competence. Designing institutions that reward competence in public life feels like an important and neglected problem.

What might Confucian AI ethics look like?

How China might approach AI ethics seems an underdiscussed topic - at least relative to both the amount of debate about AI ethics more broadly. Jeff Ding has covered this many times in his excellent newsletter, but in general it’s hard to find English language resources on the topic. I enjoyed, therefore, this interview with philosopher Pak-Hang Wong on the superb Philosophical Disquisitions podcast. Wong’s work looks at what Confucian ethics (of which Xi Jinping is a fan ) has to say about emerging technologies.

Wong looks at this through the lens of what he calls “moral deskilling” (see this paper for more). Many of us are familiar with the basic idea of deskilling: as robots and other agents take on more human activities, humans “forget” how to do them. Moral deskilling extends this idea to the ethical sphere. If, for example, robot care workers become the norm, will we also forget moral mindsets and behaviours that today we consider virtuous?

Western philosophers have considered this (see, e.g. this discussion of Shannon Vallor’s work on TiB a year ago), but Wong argues that Confucian ethics has a different framework for thinking about virtue - one more concerned with ritual and practice - that is likely to shape the design of Chinese AI services. I’m just scratching the surface, so I also recommend Wong’s essay on what “responsible innovation” looks like outside a liberal democratic framework. In a world of national technological sovereignty, understanding moral codes different from our own is going to be important. 

Can quant funds succeed in venture capital?

My day job means I’m very interested in new models for building companies. So I was intrigued by this piece about how Two Sigma, one of the most successful tech-driven hedge funds, is moving into venture capital. There have been some attempts to apply the sort of quantitative methods that have been so successful in public markets to VC (e.g. SignalFire), but in general quants have had little impact in the sector.

What’s particularly interesting is that Two Sigma is looking not only to deploy capital into existing companies, but to use data to identify opportunities and build companies from scratch. The “studio” model of company building has become popular in recent years, with firms like Atomic having remarkable success - but this is the first time I’ve seen a quant approach (and certainly from such a big name).

The article notes the growing trend of hedge funds creating private equity funds and points to Elliott as a recent example. This is a good opportunity to revisit a favourite TiB topic: Softbank’s Vision Fund (previous coverage). It was reported last week that Elliott has taken a large position in SoftBank and is pushing Masa to pull back from such huge tech commitments. Vision Fund II was already looking ropey, but it may well now be dead. If so, it’s the end of a (very weird) era in venture capital. 

Quick links

1. No to Nato? Remarkable survey data on who Europeans would side with in a US/China conflict.
2. Why corp.com is so dangerous. Fascinating story about why we don't want that domain to fall into the wrong hands...
3. Learning about coronavirus. Lingumi (an Entrepreneur First ed-tech company that operates in China) shares data on what learner behaviour tells us about the epidemic.
4. Plus ça change? How early modern universities competed for talent
5. In praise of clairvoyance. Good thread on what we know about Superforecasters.

Your feedback

Thanks for reading TiB. As ever, if you enjoy it, the best way to help is to forward it to a friend or two or to share on social media.

And it's always great to hear from you: feel free to get in touch on Twitter or by hitting reply.

Until next week,

Matt