Thoughts in Between

by Matt Clifford

TiB 152: AI, compute and national security; should Bezos run the vaccine programme; information revolutions and evolution; and more...

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

Why policymakers should worry about compute

I missed this story from a few weeks ago about the OECD’s efforts (in collaboration with, among others, Jack Clark, whose Import AI newsletter is a must read) to create an index of the computational power (“compute”) available to different nation states. The goal is to help policymakers make investment decisions relating to their national AI strategies - which at least 80 countries now have, according to the article.

The strategic significance of compute, which we discussed in TiB 128, remains underrated outside tech circles. Its importance rests on an idea known as the “scaling hypothesis”: this holds that what’s required to achieve artificial general intelligence is not a major theoretical breakthrough, but rather scaling our current machine learning models by orders of magnitude. Some recent AI achievements (e.g. GPT-3) provide some evidence for this. And one key to scale in AI is much, much more compute. Gwern has the definitive write up of the scaling hypothesis, which I highly recommend. 

The upshot is that in a world of AI nationalism, compute is a key strategic resource (see also my podcast conversation with Ian Hogarth). We’re currently in a strange transition period where machine learning models are expensive to train relative to the budget of a startup (e.g. DeepMind’s AlphaStar model reportedly cost $26m to train), but small relative to those of nation states. But in a world where a single model could soon cost a billion dollars to train, policymakers will have to think carefully about how to build or access the resources they need to compete.

Information revolutions and evolution

Joe Henrich has a short and excellent piece on the long run impact of the Reformation on culture, society and - fascinatingly - biology. We talked about Henrich’s work before in TiB 90 (featuring this extraordinary paper) and TiB 79. His The Secret of Our Success was one of my top five books of the last decade

This new piece, adapted from his latest book, argues that the emergence of Protestantism 500 years ago led not only to mass literate societies, but also caused changes in our brains relative to our pre-literate ancestors. These changes in turn enabled and reinforced the cultural centrality of literacy. “Gene-culture co-evolution” is the big idea that runs through Henrich’s work: that is, the idea that genetic evolution and cultural development are mutually intertwined, with certain kinds of cultural organisation making particular genetic profiles more adaptive and vice versa. 

Given that I’ve argued before (e.g. TIB 88, TiB 90, TiB 131) that the Reformation is the best historical analogue for understanding the present day, it’s interesting to speculate about what sorts of cultural/biological co-evolution our own information revolution might promote. I suspect 500 years from now, humans will be very different (if we survive that long…). But even in the shorter run it shouldn’t surprise us that the internet - and the ways it changes how we store and process information - has far reaching consequences for our bodies and our institutions

Podcast: technology in the fight against COVID-19

The latest TiB podcast episode is a conversation with Rowland Manthorpe, who is technology correspondent at Sky News. Rowland’s reporting over the last year has, unsurprisingly focused on COVID-19. In this conversation we discuss the ways that technology has underpinned society’s response to the pandemic - but has also frequently disappointed. The canonical example, at least in the UK, is the failure of the “Test and Trace” system first announced in May 2020. Rowland, who broke the story back then, runs through some of its challenges, as well as the reasons for its success in South Korea.

Rowland points out that areas of relative success in the pandemic have tended to come from the health infrastructure investments that countries made decades before (e.g. life sciences or the network of research nurses in the UK). The incentive for such investment is often weak, however. In a post-pandemic world it will be interesting to see whether a new appreciation for resilience over efficiency (see TiB 112) can prevail over what will doubtless be loud calls for fiscal austerity. I hope so.

Other topics we discuss include:

This was one of my favourite conversations to date. Hope you enjoy it.

Quick links

  1. Remote possibility. Excellent thread on the future of post-pandemic work.
  2. Your brain was not built for this. A series of lovely counter-intuitive mathematical facts.
  3. An old saw. Which industry has the biggest negative externalities? A candidate...
  4. The golden age of globalisation? Almost one in ten people globally emigrated between the 1840s and the 1930s.
  5. Mate! Fun collection of one-liner chess variants.

BONUS: My InterIntellect salon on “Power in the Middle Ages” is tonight. Still time to sign up if you'd like to join.

What do you think?

Today marks the third(!) anniversary of Thoughts in Between. Whether you've been reading from the beginning or since last week, thank you for reading. This is very much a weekend project, so I feel lucky to have so many engaged and thoughtful subscribers..

If you enjoy TiB, I'd love you to share it, either on social media or by forwarding to a friend. Reviews of the podcast are also very welcome.

Feel free to email me or find me on Twitter if you have any feedback or questions.

Until next week,

Matt Clifford

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.