Thoughts in Between
TiB 180: What governments should do about AI; Politics and reality; France's puzzling geniuses; and more...
Welcome new readers! Thoughts in Between is a newsletter (mainly) about how technology is changing politics, culture and society and how we might do it better.
It goes out every Tuesday to thousands of entrepreneurs, investors, policy makers, politicians and others. It’s free.
Forwarded this email? Subscribe here. Enjoy this email? Forward it to a friend.
Politics, reality and “virtualism”
We’ve talked a lot in Thoughts in Between about how the world is getting “weirder” (see, e.g. TiB 150, which is dedicated to this topic). One explanation is Bruno Maçães’ thesis that “virtualism” is the successor ideology to liberalism: pluralism persists not because of toleration, but because increasingly different groups in society can immerse themselves in distinct narratives and fantasies that become reality for them (There are some tragic examples currently in the news). I recommend Julian Lehr’s essay, “Is this real life?”, which we discussed in TiB 135, for a good discussion of the forces that have made this possible.
Perhaps that sounds absurd to you - or you think it’s a pathology on “the other side” of some political or cultural divide, but there’s increasingly strong evidence that our identities and affiliations shape our realities, whoever we are. For a long time, we’ve known that partisanship has influenced perceptions of political and economic reality, at least in the US. Democrats are much more likely to believe that the economic outlook is positive if there’s a Democrat in the White House, and vice versa. A new paper by David Brady, John Ferejohn and Brett Parker suggests that the effect has become much stronger over the last twenty years.
The authors find that the impact of partisanship on economic perceptions has roughly doubled since 1999. Moreover, it used to be that severe recessions caused perceptions to converge across the political spectrum, but even this effect has faded. Perhaps most interestingly, the paper reports a near-symmetric effect on both sides of the aisle; there is no “reality-based community”. I suspect the effect would be weaker in Europe, but I wouldn’t be surprised to be similar results in the UK. We should worry about these findings: without a shared fact base, political community is impossible.
Why is France so good at mathematics?
José Luis Ricon has a good post on what makes a country’s university system successful, based on his reading of Markets, Minds and Money by Miguel Urquiola. The key fact to explain for Urquiola is the US’s dominance in the sciences, even when you control for population and wealth. Urquiola thinks the answer lies in the US’s free market-oriented academic system, which allows resources to flow to the most productive researchers (at the cost of a much less equal distribution). José isn’t so sure and reviews potential counterexamples.
The whole post is worth reading, but the section on France’s outlier performance in mathematics is particularly intriguing (more here). France has roughly the same number of Fields Medals as the US, despite far less funding for mathematics and fewer mathematicians*. Not only that but, as José points out, the effect is concentrated at the super-elite level: neither France’s school performance (as measured by PISA) nor its high school Olympiad record are outstanding. The best explanation seems to be cultural: maths is unusually prestigious in France and as a result many more of the best students want to study it.
I think this cultural point has broader implications. Something I think about a lot (because of my day job) is which countries and cities are likely to become strong startup ecosystems over the next decade. One framework goes something like this: invest in places where educational elites are inclined towards STEM subjects, but where traditional elite career paths are becoming structurally less attractive (especially government - see TiB 61 for more on this). France is a prime example, as is Singapore. I expect Paris’s startup success to continue - with many mathematician-founders.
*Though there’s a whole Twitter account dedicated to taking down the Fields Medal
TiB podcast: what governments should do about AI
This week’s TiB podcast is a conversation with Jess Whittlestone and Jack Clark, two experts on AI policy and ethics. Jess is a researcher at Cambridge’s Centre for the Study of Existential Risk and Jack is a co-founder of Anthropic, an AI research company whose launch we discussed in TiB 167 (I also highly recommend his weekly newsletter, Import AI, which is an invaluable resource). Our conversation focuses on their new paper, “Why and How Should Governments Monitor AI Development”, which I’d encourage you to read.
The core thesis is that - as we’ve discussed many times before, most recently last week - AI is a transformational technology with broad impact across society and the economy, but one that is currently dominated by private sector actors. If governments want to be able to shape, harness and regulate it, they will have to develop new capabilities. In the paper, Jess and Jack call for governments to begin by investing in initiatives to measure and monitor AI research and deployment. On the basis that “what gets measured gets managed”, they argue that this will help governments to reduce potential harms (such as important decisions being made by biased or non-robust systems)
In the conversation we discuss:
- The capabilities governments need in AI
- Practically how governments can start down this track
- The risks if governments allow the private sector’s capabilities to outstrip theirs
- How to rank different countries’ AI capabilities
- … and much more
- Fat chance. Very long, but completely absorbing thread (with links) on the mysterious causes of obesity.
- GPT-3 does geopolitics. Fascinating / disturbing thread on what happens when you ask OpenAI's chatbot about Xinjiang.
- Supply chains smoked. Striking chart on the surging price of container shipping (but what explains the cheaper anomalies?)
- Sitting on a goldmine. Is Afghanistan sitting on $1 trillion of mineral resources?
- Against Foundation Models. The counterargument to last week's piece on foundational models in AI (see also this paper)
Thanks for your support
Thanks for reading. If you enjoy Thoughts in Between, please do forward it to someone who might like it or share it on Twitter or LinkedIn.
If you comments, questions or recommendations, feel free to hit reply.
Until next week,
PS: 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. It makes a big difference.