Matt's Thoughts In Between - Issue #70

This week: Facebook starts competing with central banks; whether we should worry about AI suffering; a history of the 21st century; and more…

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

Facebook launched a new, global sort-of-cryptocurrency, Libra, that promises to shake up the world of payments. There has been tonnes of commentary from every angle - the NYT announcement piece is a good starting point. The most interesting idea is that Facebook is now competing head on with nation states by introducing a stable, KYC-compliant currency that is (somewhat) independent of existing central bank monetary policy.

Predictably, this is drawing fire from Facebook critics and competitors. Matt Stoller has a good overview of the anti-monopoly case against Libra and Facebook co-founder Chris Hughes (whose critiques of the company we’ve discussed before) is against it too. Traditional regulators have made unfriendly noises and Tyler Cowen makes the point that the banks will hate Libra, and seldom lose regulatory battles like this. Facebook’s best argument is probably, as Alex Danco argues, that Libra is the friendly, sanitised version of crypto that governments can get on board with (but note Eric Wall’s point that it’s really not that simple). 

Andreessen Horowitz is fond of saying that startups are in a race to get distribution before incumbents get tech. In a sense, Libra is the clearest signal yet that Facebook is truly the incumbent now. There is nothing particularly novel in the technology vs classic cryptocurrency, but Facebook is hoping that its massive distribution will succeed where Bitcoin has failed so far, and make a digital currency go truly mainstream.

I wrote several months ago about the case for worrying about the suffering of insects. If that seemed farfetched, what about the suffering of artificial intelligences? That’s the theme of this thought provoking interview with Roman Yampolskiy on the excellent Philosophical Disquisitions podcast. If AIs become capable of suffering, then there are serious ethical implications of progress in AI (This is a major theme of Robin Hanson’s weird and excellent, The Age of Em).

The key question is whether AIs can or could have conscious experiences - or qualia (see here for a podcast introduction to the the topic). Yampolskiy and has developed a surprising but ingenious technique for testing for AI consciousness: show AI models optical illusions and ask them what they see. For example, show them a still image that the human mind processes as movement. There is no movement in the image, so any perception of movement is entirely a qualitative experience created in the “mind”. 

Of course, this approach has limitations. It’s easy to imagine conscious minds that don’t process images or have experiences as humans do. But it’s nevertheless an interesting starting point for exploring the space of possible minds. Perhaps this seems a frivolous topic for research, but ethics have shifted markedly over the centuries, as we’ve (rightly!) expanded our circle of ethical concern to include an ever larger group of humans (and indeed other species). Progress in AI takes us into uncharted territory. It seems prudent to start thinking about the consequences.

Tanner Greer has a fascinating new post on what a contemporary history of the early 21st century might look like. I love the format he uses: rather than try to write a whole history in a blog post, he proposes a table of contents for a book-length treatment. This gives 50%+ of the flavour of what the book would be at (presumably) <5% of the work. (This reminds me of Borges’ technique of writing “book reviews” of imaginary books as a shortcut - e.g. Pierre Menard)

Historians love the question of periodisation: when do eras begin and end? Greer opts to begin the story of the 21st century in 2004 - “the high tide of America’s unipolar moment” - though see David Auerbach’s response for the case that it should start in 1994 with Newt Gingrich’s triumph in the midterms of Bill Clinton’s first term, and Winthrop Wickard’s for the claim that you have to go all the way back to 1989. 

For Greer, the last 15 years are a story of fragmentation and institutional decline. It’s striking that elites have become more pessimistic, even as they have become increasingly successful. The end of the information monopolies of broadcast television and newspapers mean that there can no longer be narrative dominance for a particular “mood”. There will always be someone - indeed, millions - on Twitter who do not share your optimism/pessimism and can inject their dissent into the narrative. As Greer says:

Thanks for reading TiB. Please do forward this to someone who might enjoy it - it’s the easiest way to help grow the community. And do feel free to hit Reply if you have any comments - I love reading your thoughts.

Until next week,

Matt