Thoughts in Between

by Matt Clifford

TiB 211: Experiments in governance; red-teaming science; fixing Twitter; 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.

Twitter governance is hard: Elon Musk edition

Elon Musk announced a bid to buy Twitter (we talked about Musk and Twitter a couple of weeks ago). I was naive, or perhaps just British, in that I didn't expect this to be a full-blown US culture war issue... but it has. Crudely, one camp believes Twitter has a tendency to censor right-leaning political speech (and that Musk will end this) and the other believes that this is not true - or if it is, it's for good reason. Unfortunately most of the commentary is consequently tedious, as nearly everyone's arguments are entirely predictable based on their broader political stance.

It's nevertheless worth reading this thread and this thread by former Reddit CEO Yishan Wong, as well as Marc Andreessen's annotations as examplars of two of the sides of the debate. And, of course, you should read Matt Levine. The most interesting thing I've read on the topic, though, is this post by Anthony Lee Zhang. Zhang points points out that Twitter will always represent a difficult governance challenge because its political and cultural importance dwarf its ability to monetise:

Considering Twitter as an individual profit-maximizing firm, if you buy Twitter and redirect eyeballs to your preferred target, you will likely lose money, since your tweaks to the feed ranking algorithm likely decrease engagement somewhat and thus decrease ad revenue. But if your goal is, for example, to destroy democracy in the US, or push some other kind of political agenda, you may be willing to sacrifice a lot of revenues on Twitter to do this

The profit-maximising version of Twitter is likely not the same as the free speech-maximising version of Twitter or the democratic stability-maximising version of Twitter or the personal abuse-minimising version of Twitter. What version will Elon Musk's be? Even if he wants to "open source the algorithm" (whatever that means...), my guess is that neither he nor anyone else knows.

PS: Ben Thompson's take on what a Musk-run Twitter could look like is, unsurprisingly, very good

Actually, *all* governance is hard

This question of how to "do" governance for high-stakes organisations like Twitter - and, probably more importantly, the kinds of AI research organisations we talk about a lot here - seems both important and under-explored. Holden Karnofksy has a really excellent new-ish post on the topic, which I highly recommend. Karnofsky points out that for organisations that might plausibly have the potential to change the world forever, none of the standard models of governance are very satisfactory. As he shows, each of "shareholder voting plus board of directors", "non-profit-style board" and "founder control" fall short when the stakeholders are the entire world.

What other options are there? As one example, Karnofsky links to this fascinating paper on decision making under moral uncertainty, which argues for what it calls "proportional chances voting". There's a good description in this footnote, but in short:

Let's say that 70% of the Parliament members vote for bill X, and 30% vote against. "Proportional chance voting" literally uses a weighted lottery to pass bill X with 70% probability, and reject it with 30% probability... A key part of this is that the members are supposed to negotiate before voting and holding the lottery. For example, maybe 10 of the 30 members who are against bill X offer to switch to supporting it if some change is made.

(The example is for a Parliament, but the proposal could apply to any kind of organisation) The idea is that this forces compromise and the weighing of all stakeholders' views and so reduces the risks of the "tyranny of the majority".

To be clear, Karnosfky doesn't argue that proportional chances voting - or anything else - is "the answer". His point is that there is remarkably little innovation in governance (or even a comprehensive literature on the topic) despite the upside of better models being significant. I came across this wonderful thread by Henry Farrell this week on social science books that ought to have wider readership. My own suggestion for such a list would be Legal Systems Very Different From Ours. It's a book that opens your mind to just how large (and weird) the space of possible solutions for governance is. It would be nice to see it explored.

Why we need "red teams" in science

We've talked a lot here about ideas for improving how science works (see, e.g., TiB 134, 165 and 153), so it was great to have Stuart Buck on the TiB podcast this week. Stuart has spent most of this career exploring this question. He was until quite recently VP of Research at Arnold Ventures, where - among other things - he contributed to the creation of organisations like Vivli and the Center for Open Science. He's now the executive director of the Good Science Project, a new organisation funded by Patrick Collison and an ACX grant to work on improving US science funding.

A lot of our conversation focuses on the twin problems of reproducibility (are published research findings true?) and innovation (are published research findings important?) in science and the potential tension between them. We explore some of Stuart's ideas for improving things, notably the ideas that we should demand more null results and encourage the practice of "red teaming" apparently important results. This idea causes us to touch on one of my favourite social science books, The Enigma of Reason (see also this review of it that Stuart references), and the question of when we can rely on our rationality to correct mistakes versus enlisting the scrutiny of others.

I highly recommend Stuart's essay for Works in Progress on these topics, if you're interested in this theme. We also discuss:

Quick links

  1. So far from God. Fascinating thread on religiosity in Europe (Why is the Netherlands an outlier here?)
  2. On doing nothing. Great discussion about why we moderns are so bad at doing nothing, given that humans in general are good at it…
  3. German exceptionalism. Excellent thread putting the geopolitics of Germany, Russia and energy in historical context.
  4. Low hanging chlorine. A surprising and optimistic finding on the impact of water treatment.
  5. Focus! Our confidence in our ability to multi-task is inversely correlated with our actual ability to do it.

Just a newsletter writer standing in front of a reader...

... asking you to share this. Forwarding to a friend or a tweet is particularly effective.

And feel free to hit reply if you have comments, questions or suggestions.

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.