Discover more from Res Extensa
Monthly Reading, September 2023
Res Extensa #40 :: Learning with YouTube, what happened to the industrial R&D labs, quantifying innovations, and Formula 1
There’s some confidence that comes with seeing something vs. reading about it.
The comfort that comes with knowing that advice is always at hand eases entry into adopting maintenance-mind in general. Instead of suppressing awareness about things going wrong with your devices, you become calmly vigilant for troubling signs and ready to act on them as needed. You take charge of your stuff.
People say apprenticeship culture has waned in the last several decades. True, in the formal sense. But I think modern apprentices are getting tutelage from a new master in the Internet.
Here's a wild fact: in the late 1960s, Bell Labs, the basic research-focused corporate laboratory of AT&T, employed 15,000 people. Entire high-tech companies aren't even that big these days.takes a look here at what happened to these big corporate labs — the likes of Bell, Xerox PARC, DuPont.
No one is quite sure why the lab model failed. It’s obvious that a scenario where Xerox is paying scientists to do research that ultimately mostly benefits other firms, potentially even competitors that help to put it out of business, could never survive. Similarly, the tension between managing scientists with their own pure research goals in such a way that they produce something commercially viable, while still leaving them enough latitude to make important leaps, seems huge. But these problems were always there in the model. What is harder to identify is an exogenous shock or set of shocks that changed the situation that existed from the 1930s until somewhere between the 1960s and the 1980s.
Of course lots of basic research is still taking place inside corporations, at places like Google, Apple, TSMC, or Nvidia in tech, or DuPont (still), BASF, and 3M in the material space. But for the most part that research is entirely private. Outside of some examples in machine learning and AI, there's a lot more R&D that never leaves the private confines of the parent company. Apple has at least a few Bell Labs-sized subdivisions internally, but their work doesn't have external market benefit beyond an indirect influence (like how the iPhone's digital keyboard became the norm).
He cites some depressing statistics about how much graft and waste is embedded in this crusty, overgrown, bureaucratic ecosystem:
You've now spent $30 million of your $100 million on indirect costs and $22 million on grant applications. Over half your budget is gone and nobody's actually done any science. And there’s no guarantee the rest actually ends up funding worthwhile research. Government grant applications are peer-reviewed by grant committees that notoriously disagree with each other, penalize risky and interdisciplinary research, maybe toss a few grants to their friends, and fail to predict which projects ultimately end up being useful. Something like 10-30% of the papers that come out of your funding won't be cited within five years, or possibly ever.
Getting good at academia is like getting good at Scrabble: it's just a made-up game, so many of the skills you acquire while playing it are useless anywhere else.
My friendstarted a newsletter devoted to Formula 1. He's only a few weeks in and already off to a strong start. As a newbie to following the sport, he's my resource in DMs on race days. Now I can get some more in-depth perspectives, too. This piece is about the slog that is the mid- to lower-point scoring drivers and teams, and how everyone is working their ass off to eke out incremental improvement — a good lesson for all of us:
So despite being an incredibly glamorous and romantic sport, one of the things I love most about F1 is how unromantic everyone remains about their progress. They celebrate the incremental progress, knowing the best way to make huge gains is to start by making small ones.
But our goal is to try and think about the science of innovation. And this means we can test innovations interconnect in an endless chain, each building on prior advances; combinations across disciplines drive major leaps forward; and the complexity of innovations has increased over time, requiring more interdisciplinary expertise. Mapping the evolution of innovations illuminates innovation's compounding nature.
Because the data doesn't exist anywhere to investigate these networks of innovations, he started his own database to reconstruct a temporal web of how and when innovations happen. You might also be able to add a geographical element (at least for the more recent discoveries) to try and uncover causal factors due to scenes forming in specific places.
Res Extensa is a reader-supported publication. Subscribe to receive new posts and support my work.