Matt's Thoughts In Between - Issue #49

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

Will tech cause a trade union renaissance?

I wrote last week about the political consequences of institutional decline and mentioned in passing the role that trade unions formerly played in building meaning and identity. The Economist this week has an interesting piece on the decline and potential rebirth of unions, focused on the role of technology.

The motivating case study is of YouTube creators quasi-unionising to protest the removal of adverts from - and hence their ability to monetise - their videos. (Arguably, something similar is happening to Patreon, though the motivation is very different and arguably from the other side of the political spectrum). 

Tech has both undermined and potentially provides new models for trade unionism. The Economist points to the ways in which tech enabled better output measurement (which in turn enabled performance-related pay and a diminished sense of solidarity) and reduced the political importance of physical assets (e.g. coal mines) that had previously provided unions with leverage. But it does cut both ways - in theory, tech should make some kinds of organising much easier - and I think it’s likely we’ll see more pro-worker tech in the coming years (e.g. Entrepreneur First alumni Portify). It’s an important topic and I’m intrigued to see how it plays out.

I, for one, welcome our new Starcraft overlords

I’m surprised that DeepMind’s AlphaStar’s comprehensive defeat of some of the world’s leading Starcraft players didn’t make more of a splash in the mainstream media. There’s a good explainer here. It’s worth watching some of the video in the first link, even if you have zero interest in the game itself, as an illustration of the AI’s capabilities.

As this graphic illustrates, there are at least some dimensions in which Starcraft is significantly more challenging than Go, which DeepMind defeated with AlphaGo to much acclaim back in 2016. In particular, Starcraft is a game of imperfect information, unlike Chess or Go, and so there is so single optimal strategy (though see here for a skeptical take). The DeepMind team they worked on AlphaStar did a Reddit Ask Me Anything on the topic and they have interesting insights on the topic.

One of the most important questions in AI remains how generalisable DeepMind’s methods are. Do we need a more fundamental breakthrough to achieve AGI or will deep learning just achieve milestone after milestone in fields that each time we think are the preserve of humans alone? MIT Technology Review has an interesting piece this week that provides some historical perspective and suggests deep learning will need to give way to new approaches - but the reality is that we don’t know whether this time it really is different.

"Influencer incubators" and the new career

As I’ve written about before, I’m very interested - both for self-interested and intellectual reasons - in the broad idea of “individuals as an asset class”. It strikes me that there are more and more fields in which it makes sense for individuals to share risk and upside with organisations who can (hopefully) greatly improve their productivity. Computer science education and company creation are perhaps the most advanced areas, but they just scratch the surface...

I was therefore fascinated to read this thread on “influencer incubators” in China. Do read the whole thing, but the basic idea is that (a) it’s increasingly lucrative for a certain type of individual to make a career of monetising their social media influence; (b) there are a set of skills and tools that make success in this field more likely; and (c) returns in this field are (perhaps) power law distributed, so it makes sense for third party investors who can help with (b) to make a lot of bets in the space. 

As so often, China is way ahead of the curve here (I’ve not come across any comparable Western organisations, but let me know if I’m missed them) and it does seem a glimpse of the future. It’s interesting to speculate about which other fields might benefit from this approach. One journalist and author who wrote an in-depth profile of EF once said said he’d like to see the same idea applied to writing books - and presumably sports and art might also be a good fit. I’m excited to see how this develops in the coming years.

Quick Links

1. Wrong password? Come on in. A surprising security result.
2. The path to power. Amazing graphic of the pre-Congress careers of members of US Congress
3. Spoiler: the Black Death was really bad. The UK population over the very long run
4. Move over Apple. Stunning embroidered computer.
5. It's a small world after all. Beautiful animation of global air and sea flows

Your feedback

Thanks for reading. If you like TiB, you should forward it to a friend who might enjoy it or share it on social media. If you have feedback or suggested links, please feel free to hit Reply

Until next week,

Matt