TiB 157: History's most pivotal moments; how SpaceX won; AI + UBI; and more...

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

OpenAI CEO Sam Altman has a new piece on how to harness the power of AI for everyone. In short he advocates something like “AOC meets Henry George”: a wealth tax charged in equity, plus a land value tax to fuel a universal basic income (UBI). I don’t think there is anything intellectually radical here - as we discussed back in TiB 15 and TiB 16, these ideas have been around for a long time - but it’s fascinating as a showcase of how mainstream these previously fringe ideas have become among tech elites.

What are the dangers of this approach becoming the default way we deal with AI as a public policy issue? I think the best critique comes from this 2019 essay by Vi Hart*. I confess I didn’t expect to be persuaded by it, as I tend to be reflexively anti-anti-tech and anti-anti-capitalism. But it’s neither of those: it’s technically sophisticated, deeply reasoned and radical in its implications. It’s very long (probably needs an hour), but I think for at least some of you it will be the best thing you read this month, as it was for me.

The core argument is that “AI + UBI” as a policy prescription is seductive but profoundly ideological. It cedes too much power to the makers of AI and places too little value on the ongoing importance of human labour. The history of labour, Hart argues, shows that valuable work appears low value in a market economy in conditions of monopsony - and that’s the situation for the work that humans do to make AI possible. Hart points to alternatives to UBI, such as new institutions like the “Mediators of Individuals’ Data” proposed by Glen Weyl and Jason Lanier (which in turn remind me of some of Albert Wenger’s ideas in World After Capital). There are no easy answers here, but I am persuaded of the importance of keeping open a plurality of solutions to the challenges (and benefits) that AI will doubtless bring.

We’ve talked before (see TiB 100 and TiB 151) about the importance of “venture customers” - i.e. buyers willing and able to provide real market “pull” for products that don’t yet exist - in innovation. The canonical example is perhaps NASA’s demand for integrated circuits for the Apollo missions, which reduced their price tenfold over five years (see TiB 124). Although this is one of the most important roles governments can play in innovation, there’s relatively little experimentation in the space.

It’s therefore exciting to see this excellent new case study by Eli Dourado on NASA’s Commercial Orbital Transportation Services (COTS) programme - an experiment in procurement innovation. COTS’ goal in 2006 was to acquire a new way to resupply the ISS, given the imminent obsolescence of the space shuttle. It turned out to be extremely successful: the goal was achieved (twice - SpaceX and Orbital Sciences both provided working solutions) at a cost an order of magnitude lower than usual.

Dourado identifies three reasons for its success versus the old “cost plus” model. First, partner companies like SpaceX had to pay a share of their development costs themselves, so they were incentivised to build something of lasting commercial value. Second, payments were fixed sums and made only on successful reaching milestones. Third, NASA specified only the outcome, not the means, required. Lots more here. Procurement is unglamorous, but extremely powerful. We need more such experiments.

I’ve been enjoying The Rest is History, a new-ish podcast from Tom Holland and Dominic Sandbrook. They had a fascinating episode last week on one of my favourite topics - counterfactual (or “what if?”) history. Counterfactual history is somewhat disreputable among professional historians (“pure speculation”, etc), but enormous fun. It’s also great fodder for fiction, as discussed in the episode: Fatherland and Dominion are two personal favourites.

One of the reasons historians usually dislike counterfactual history is that it tends to ignore the big macro forces that shape the course of events. For example, to ask, “What would have happened if Napoleon hadn’t lost at Waterloo?” is to ignore that there are good structural reasons why he lost that wouldn’t disappear if you changed the outcome of a single battle. As Holland and Sandbrook emphasise, the broad shape of history would usually remain the same if you reversed some of the surface phenomena.

I think they take this too far though. History is a complex system marked by a high degree of path dependence. It seems plausible that we would get some wildly different outcomes if we “re-ran” history (see TiB 132 for previous discussion). Holland and Sandbrook do concede some individuals have low “replaceability” - they cite Jesus, Mohammed and Churchill (in May 1940) as examples. “What if Catherine of Aragon had given Henry III a son?” is likely another. I suspect it will matter a lot who creates AGI. If you have favourite examples of your own, I’d love to hear them. please do hit reply and share.

BONUS: The latest TiB podcast episode is a conversation with leading tech journalist Steve O'Hear on the future of European tech, Clubhouse, diversity in tech, and much more. I think it’s a really fun one; have a listen.

Thanks for getting all the way to the end. If you enjoyed this, I’d love you to take 30 seconds to forward it to someone else who might too.

Your comments, questions, recommendations, etc, are always welcome - feel free to hit reply.

Until next week,

Matt Clifford