Thoughts in Between

by Matt Clifford

TiB 97: the end of the tech cycle; the limits of explainable AI; organising genius; and more...

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

Where are we in the technology cycle?

A lot of important questions - both socio-political ones and, more parochially, for VCs - rest on whether you believe Big Tech will be disrupted by new entrants over the next, say, decade. You’re much less likely to think regulating or even breaking up Facebook et al is a good idea if you believe their dominance will be short lived. And, for VCs, the types of companies that are attractive investments are quite different if you believe we're due a new computing paradigm.

This week two of the best tech analysts, Alex Danco (here) and Ben Thompson (here and here), address this topic. Both look at the question through the lens of Carlotta Perez's Technological Revolutions and Financial Capital, which remains the essential framework for thinking about how tech interacts with finance. (See Jerry Neumann's excellent summary here or this podcast). Both conclude that we’re into the deployment phase of the “computing” technology cycle, where a tech paradigm matures and the role of speculative capital (and the prospect of astronomical returns) diminishes. 

Danco notes that VC has already started to transition: SAAS funding today looks more like “production” than “financial” (i.e. risk) capital. What does this mean for VC over the next decade? I suspect a hollowing out of the middle, with the best (risk adjusted) returns for innovative production capital (like Stripe Capital or Capital-as-a-service) and for super early stage exploration of frontier technologies that can seed new paradigms. It goes without saying that I’m biased here...

GPT-2, chess and the limits of explainable AI

Scott Alexander has a fascinating blog post on teaching GPT-2, OpenAI’s Natural Language Processing model (which we’ve discussed before), to play chess. The model doesn’t have a concept of a chess board or the rules; it’s learned to play entirely by reading millions of games written out in chess notation. As a result, it makes a lot of mistakes, but it’s still a striking result.

How much should we be impressed by this? As Alexander notes, it’s hard to say. Gary Marcus, an AI pioneer and noted deep learning skeptic, is, well, skeptical. But even if you don’t think this is a meaningful AI milestone, there’s lots to ponder here. Matt Levine’s take on the implications for finance made me think. Levine notes that “black box” trading algorithms are criticised for opacity and, facetiously, suggests that a new approach would be to train GPT-2 on written stock recommendations and get it to write the human-legible investment memo for you.

The joke, of course, is that while such a memo might sound plausible, the reasons it gave for investing in a stock wouldn’t be the actual reasons for the model selecting the stock in a causal sense. We’d have “explainable AI” with spurious explanations. But is that so different from the human mind? Dan Sperber and Hugo Mercier’s excellent book, The Enigma of Reason, suggests the primary evolutionary purpose of human reason is not to make better decisions, but to convincingly justify decisions after the fact. Perhaps in the space of possible minds, GPT-2 is closer to us than to something truly alien. 

Request for input: a reading group on "organizing genius"

Last year I hosted a group in London to read through Peter Thiel’s Stanford class on Sovereignty, Technology and the Limits of Globalisation. It was one of the most interesting things I did in 2019 and I met lots of great people. I’d like to do something similar in 2020 with a different syllabus. The topic I have in mind is something along the lines of “Organizing Genius” - that is, exploring “What are the organisational forms and practices that maximise the output of talented people?

The title comes from this book, which was recommended to me by Demis Hassabis, co-founder and CEO of DeepMind (and one of our investors at Entrepreneur First). It’s a topic I’ve been thinking about for a while, but I’ve not yet dedicated time to thinking about it systematically, so this is a good excuse. I’ve not been able to find a syllabus directly on this topic, so I thought I would try to pull one together myself - with some help from you... I've had a go at putting together a draft reading list here. Have a look.

This focuses on cases studies of organisations that combined extreme talent concentration and extreme achievement. It’s very much a work in progress with lots of blind spots, so I’d love recommendations, either on groups to study or relevant reading material (e.g. my friend Arnaud already shared some great ideas for some more theoretical material on what makes super productive groups form). Just hit reply if you have ideas! If you’re interested in joining in the reading group, I plan to share a sign-up form next week. 

Quick links

  1. Let me complicate that a bit. The rise and rise of nuance. (I feel it's necessary to link to this paper)
  2. Big questions in theology. Can you convert elephants to Christianity? A (surprisingly?) interesting Twitter thread.
  3. It's getting hot in here. Why did medieval trials by ordeal work?
  4. Non-dancing Queen? Rob Wiblin on why Anglo cultures don't dance and probably never will.
  5. Soaring sales. Ethiopia's surprising biggest export.

Your feedback

Thanks for reading Thoughts in Between. If you enjoy it, it'd be great if you'd forward it to a friend or two. It's always great to get feedback: feel free to get in touch on Twitter or by hitting reply.

Until next week,

Matt