TiB 119: Internet diplomacy in 2020; AI that writes poetry; startup ecosystems and art history; and more...

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

China, Trump and the weirdness of internet diplomacy

The Economist reports that authoritarian regimes are enjoying America’s current turmoil - and especially anti-protestor violence, which represents a welcome opportunity to normalise or obscure their own (much worse) behaviour. China in particular has used the trauma of 2020 to step up propaganda campaigns, both overt (e.g the so-called “Wolf Warrior” diplomats) and covert (e.g. the tens of thousands of fake accounts Twitter took down recently).

This is a complex area, though, with many odd elements. It’s pretty clear, for example, that Wolf Warrior diplomacy doesn’t work. This paper shows that it actually hardens Western attitudes towards China among those exposed to it (By contrast, messages that emphasise China’s foreign aid and investment soften feelings). But it seems to be what Xi wants, irrespective of the impact.

Most interesting, though, is the spotlight current events put on Trump’s complicated relationship with China. On the one had, Trump's sympathy for authoritarians - and Xi in particular - is well documented, as this excellent podcast with Evan Osnos makes clear (as do Trump’s past comments on Tiananmen Square). On the other hand, China’s liberal dissidents have a lot of sympathy for… Trump! This thread and this paper suggest that their valorisation of the West and the concomitant belief that the West has no injustice makes them reflexive defenders of Trump and the US. The internet makes everything weirder - and that's certainly true of diplomacy.

One step closer to general intelligence?

OpenAI announced GPT-3, a new natural language processing (NLP) model with 175 billion parameters and impressive capabilities. We talked about its predecessor, GPT-2, before (here, here and here), which could already (sort of) play chess, write poetry and do arithmetic. The new version has made progress on at least the latter two of these; Scott Alexander gives some striking examples here (If you read one piece on GPT-3, read that).

OpenAI acknowledges that there’s no fundamental breakthrough here. What’s new is sheer scale and the performance boosts that seems to bring, depending on whose analysis you follow. There’s a debate on the future of artificial intelligence that (crudely) divides those who think that we need radically new techniques to crack “artificial general intelligence” (AGI) and those who think we just need to scale existing techniques. We talked about this last year and the “bitter lesson” that "building in how we think does not work in the long run”.

If you’re in the latter camp, the case for worrying about GPT-3 is best expressed by Gwern’s excellent essay - or in single tweet, here: scale seems to improve performance a lot and GPT-3 shows no sign of an upper limit on this. Not everyone agrees: there’s a case for viewing GPT-3 as a let-down, as argued in this post. But, as Gwern says, these capabilities would have seemed astonishing, even to leading reseachers, just five years ago. As one of the best essays on AI safety has argued, there’s no fire alarm for AGI.

What startup ecosystems can learn from art history

I wrote last month about the critique that there are “too many entrepreneurs”. This prompted some great conversations, including an excellent salon on the topic with InterIntellect and this thread from my friend Roy Bahat. Like Roy, I believe - no surprise, given my day job - that we actually need many more of some kinds of entrepreneurs.

More controversially, I believe that one thing that has artificially reduced the size of startup ecosystems is that a subset of those already on the inside prefer to keep entrepreneurship outside the mainstream. This explains lots of silliness, like the idea that entrepreneurs are fundamentally, even genetically, different from other people (even though many of the most successful deny it) or the rhetorical insistence that founders need be “crazy”, “misfits” or rebels. To me this is special pleading, no different from any guild that naturally wants to restrict entry to its profession.

On this theme, this week I came across a superb series of short essays (start here, then here) by Bryan Kam in which he compares my thoughts on ambition and entrepreneurship to the history of art. Kam argues that every group of innovators (in art, artists) needs a group of interpreters (in art, critics) if they want to influence the mainstream - but that the innovators always resent the critics despite this. I suspect the same is true of startup ecosystems: the old guard always resent the "tourists”, even though they’re a sign - sometimes even a source - of success.

Quick links

BONUS: I recorded a podcast episode with my friend Azeem Azhar of Exponential View and we discussed many themes from TiB. You might enjoy it.

Are urban Chinese richer than Americans? It depends what you measure (and how much you care about inequality).
That arc of the moral universe sure takes its time. Support for mixed race marriage was a minority opinion in the US until... guess the year.
Marc his words. Superb interview of Marc Andreessen by Sriram Krishnan (and great Twitter summary here)
The... lightbulbs have ears? Amazing/terrifying technology that allows conversations to be recovered from light variations in bulbs.
Where does top AI talent come from? Great set of visualisations of patterns in global AI talent.

Your feedback

Thanks for reading Thoughts in Between. If you enjoy, I'd love your help growing the readership: the easiest way is to forward it to a friend or share it on Twitter or LinkedIn.

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.

I'm always happy to hear from you. Feel free message me on Twitter - or just hit reply.

Until next week,

Matt Clifford