Thoughts in Between

by Matt Clifford

TiB 191: Top 10 papers of TiB, including AI and science; where great founders come from; the "need for chaos"; and more...

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.

TiB vacation project: the best papers of the last 4 years

I'm was on vacation last week and tried to shift most of my reading to fiction for a few days, so for this edition I thought I'd try something different as a "holiday special".

A lot of what I cover in TiB is new academic papers across my interest areas, so what follows is my "Top 10 papers of TiB" - my favourite articles from the last nearly-four years of writing the newsletter that I think are worth a re-read (Plus Quick Links as normal).

Here goes, in the order in which we discussed them...

Some people just want to watch the world burn...

Petersen, Osmudsen and Arcenaux on how a psychological trait they dub the "Need for Chaos" (or "burn it all down"), and not partisan interest, is the best predictor of spreading political misinformation. A crucial lens through which to view recent Anglo-American politics. (Discussed in TiB 81)

How the Catholic Church rewired our minds

Henrich et al on how the Catholic Church rewired our minds: A remarkably ambitious paper that argues that the Catholic Church's long promotion of the nuclear family over extensive kinship ties explains why European-descended societies tend to be "WEIRD" psychological outliers relative to global norms. (Discussed in TiB 90)

Stagnation and scientific incentives

Bhattacharya and Packalen on stagnation and scientific incentives: The authors argue that a major cause of scientific stagnation is the focus on citations in evaluating researcher performance - which pushes academics towards lower risk, incremental research and away from the more fundamental but riskier bets on which growth relies. (Discussed in TiB 103) See also Cowen and Southwood on the slowing of science (TiB 92) and Wu, Wang and Evans on the value of small teams in science (TiB 52)

Pro-authoritarian technologies

Caplan on pro-authoritarian technologies: There's a lot of discussion about whether AI makes totalitarianism easier (e.g. automated surveillance). Bryan Caplan argues that the most dangerous pro-authoritarian technology is actually life extension, as it's leader succession that is the greatest weakness of dictatorships. (Discussed in TiB 93)

Discovering new materials with machine learning

Tshitoyan, Dagdelen, Weston et al on using machine learning to discover new materials: Amazing paper in which the authors use natural language processing to read, unsupervised, the entire corpus of academic articles in materials science and then use the model to "discover" novel materials with useful properties (Discussed in TiB 72) See also Eric Schmidt and Maithra Raghu on how deep learning is changing science (TiB 114)

The end of disruption

Bessen, Dent, Kim and Righi on the end of disruption: Four economists show that, despite the rhetoric from my own industry, disruption has actually declined since the advent of the software revolution. It's not, they say, because of lax anti-monopoly enforcement or lobbying, but because big firms today have deeper moats built of intangible assets. (Discussed in TiB 132)

The hardware lottery (or why the best ideas don't win)

Hooker on "the hardware lottery": An fantastically thought provoking paper on why the best ideas don't always win; in machine learning at least a (software) idea needs not only to be good, but to fit with the prevailing hardware paradigm. A great read with far reaching implications. (TiB 133)

Why we dream

Hoel on why we dream: Erik Hoel, a neuroscientist and fiction author, proposes a fascinating hypothesis: we dream to inject noise into the "training data" our brain uses to learn, in order to prevent overfitting and improve our ability to generalise (Discussed in TiB 134)

"Forced entrepreneurs"

Hacamo and Kleiner on "forced entrepreneurs": Contradicting the idea that there are "born entrepreneurs", the authors find that when banks and other high wage employers higher fewer staff, more of these people become founders and they are more successful on average than "voluntary" entrepreneurs. As I say all the time, the world is missing out on some of its best founders! (Discussed in TiB 165)

The roots of exceptional performance

Güllich, Macnamara and Hambrick on what produces exceptional performance: This study finds that, contra popular belief, focused and deliberate practice (the famous "10,000 hours" claim) in athletics is associated with strong performance at junior level, but negatively correlated with exceptional performance at senior level. The very best athletes tend to specialise late. (Discussed in TiB 177)

Quick links (from this week)

  1. The terrifying power of conformity. What percentage of people do you think will let a researcher search through their unlocked phone if asked to do so for an experiment?
  2. The fall and fall of US immigration. Trump effect or secular shift?
  3. Goods, services and pandemics. One explanation for supply chain chaos in a single chart.
  4. Not a super forecast. What chance do superforecasters put on US-China conflict in the near/mid-future? Sobering.
  5. Partisanship shapes reality, part 1089: Republican evaluations of the US economy today are more negative than those of either party when Lehman collapsed (NB: this is a symmetric phenomenon)

BONUS: Adventures in technological sovereignty. Daniel Ek invests €100m in defence AI firm Helsing. Hopefully more in this in a future TiB edition. (Disclosure: I am a small investor)

Thanks for your support

Normal service resumes next week. Thanks for reading. As usual, if you enjoy this, please forward to a friend or share on Twitter.

And if you have comments, questions or recommendations, drop me an email. I'm always happy to hear from readers.

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.