Thoughts in Between

by Matt Clifford

TiB 216: AI and the future of art; Rawls and web3; the Ottoman Empire and the printing press; and more...

Does DALL-E mean the end of art?

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.

A few weeks ago, in TiB 210, we talked about DALL-E, OpenAI's new model that "draws" impressive images based on human-supplied captions. This thread has some striking examples (Here’s an explanation of how it works that I came across since). I somewhat facetiously used the title, "AI comes for the illustrators", but of course developments like DALL-E (and Google has released its own version) do raise profound questions about the future of art and artists. Erik Hoel has a new post on this topic, which is worth reading. Erik was previously on the TiB podcast and we discussed his related essay on AI-novels in discussed in TiB 169.

He gives his piece a brilliant and provocative framing:

You were born into a world where most things were made by human consciousness. You may die in a world where nothing is made by human consciousness

Erik draws a distinction between the extrinsic properties of art (what it looks like, essentially) and its intrinsic properties (what it means to the artist and their communication of that to the viewer). He notes that critics from Aristotle to Tolstoy have argued for the primacy of the latter (Tolstoy: the "aim of works of art is to infect people with the emotion the artist has experienced"). But art generated by AI can, of course, have no such properties, at least until AIs become a lot more advanced than DALL-E...

Does this matter? Hoel says yes and his conclusion is pessimistic. But other responses are possible. AI artists may become a tool for human artists, as Ted Underwood notes (Erik considers this too and calls it the "best case scenario"). Nabeel Qureshi expands on this: maybe AI will do for art what it did for chess - simultaneously make humans "obselete" and yet usher in a golden age for humans too. And Matt Clancy points out that the more you believe in the value of art's intrinsic properties, the more optimistic you should be that we'll still value human art. This is an important debate to have now; I expect we'll be having it about many other domains over the next decade.

What web3 can learn from John Rawls (maybe)

We've talked before about the intersection of political philosophy and emerging technology - an underated topic in my view. Back in TiB 190, we looked at Iason Gabriel's work on the relationship between AI and John Rawls' Theory of Justice (ToJ), arguably the most influential work of political philosophy of the 20th century. As I said back then, much of political philosophy since 1971, when ToJ was published, has been responses to or critiques of it from angle or another. This week I came across a new invocation of Rawls' relevance to emerging tech - Li Jin's exploration of how ToJ should shape Web3.

Li argues that if Web3 is to achieve its potential and fulfil its mission, it needs to ground itself in Rawls' difference principle, one of the most important ideas in ToJ. In short, the difference principle specifies that inequality is morally justified only to the extent it benefits the worst off. So, for example, we might tolerate entrepreneurs becoming very wealthy if that creates the incentives for people to build the products and services (and the economic growth) that raise the absolute position of the least advantaged. Li argues that this idea should be at the heart of a lot of web3 projects, as a way to operationalise web3 values such as self-determination, democratisation and rewarding participation.

I'm not sure that I share Li's conclusions but I think the piece does force us to ask an interesting question: how important to human flourishing is the fundamental design of the web? Is it so core that it should be subject to the moral demands of the veil of ignorance - and that there's a moral imperative to replace "web2" with web3? I find the communitarian turn in crypto (see TiB 200) absolutely fascinating; it seems a far cry from its anti-state / cypherpunk origins (not that that's disappeared). But my guess is that it shows just how early things still are. Current market conditions notwithstanding, everything is still up for grabs.

Lessons from history: live near a printing press?

An occasional favourite TiB topic is the sometimes surprising long-run impact of historical events, from the Reformation to the Cultural Revolution. We've talked about this several times (see, e.g., TiB 9099152179), but I particularly recommend the discussion in TiB 194, where we discuss how robust these effects are, if you're interested in this sort of thing. I came across a new and interesting paper in this genre this week, by Bogdan Poescu and Mircea Popa, which looks at the long-run impact of Ottoman rule on human capital development in Europe.

The authors exploit the fact that different regions within modern nation states had varying levels of exposure to Ottoman governance (ranging from 0 to 600 years - see this chart) to explore what impact this had on economic development. The reported results are pretty striking: a difference of 300 years of Ottoman exposure is associated with around 20% lower GDP per capita today. This is roughly the same magnitude as the standard deviation of GDP per capita in these regions, so represents a significant effect.

As we discussed in TiB 194, it's frustrating when these sorts of studies publish eye-catching headline effects, but don't provide a plausible transmission mechanism. The authors here, though, supply an interesting hypothesis: the Ottoman empire saw late adoption of the printing press, which in turn delayed the emergence of mass literacy (There's a very good discussion of this in the paper). This chart, for example, shows a fairly strong relationship between a town's physical distance to a printing press in the 16th century and its literacy in 1930! It's worth pondering what similar long run effects we may be sowing the seeds of today. Perhaps, in the spirit of the today’s first section of AI art, the lesson is: embrace information revolutions!

Quick links

  1. Tinder box? Interesting thread on dating apps and sex inequality
  2. "The ghost of a codec". The "music" that's stripped out by the MP3 compression algorithm. Surprisingly compelling.
  3. The world is missing out on some of its best founders. Evidence from East Germany.
  4. "Value add". In a shocking development, it seems some venture capitalists overestimate how helpful they are!
  5. Bear with me. The Chinese real estate market does not look healthy.

The bit at the end

It's great to have so many new readers this week (thanks Tyler!). If you got this far and liked it, please do share with a friend.

As always, feel free to 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.