🧵 View Thread
🧵 Thread (38 tweets)
1/ Recently finished "Scientific Freedom: The Elixir of Civilization" by Donald Braben.It's a sobering (but hopeful!) exploration of the stagnation in what I would call "paradigm shifting research" and what to do about it. BOOK REPORT THREAD 🧵
2/ Braben points out that the number of scientists has exploded to the point that just increasing the quantity of funding without changing *how* it's assigned, won't make any difference. Note that he points to the same dynamics outlined in "The Decline of Unfettered Research. https://t.co/GEeLmTwOZ1
3/ There seems to be a fundamental tension in government-sponsored research. On the one hand, we want the government to be accountable for every cent it spends. On the other hand, worrying about the ROI of research from day one can end up strangling it in the cradle. https://t.co/THl40o2WJs
Braben fully embraces the role of individuals in paradigm shifting research, and argues that modern culture has shifted away from enabling these people.This again mirrors "the decline of unfettered research" but with different conclusions. https://t.co/H9k7SboHMx
5/ The universe is shaped by nonlinearities. https://t.co/IrBajxAioD
6/ Five years seems like an important timescale in human affairs. Freeman Dyson argued that it was the maximum amount of time that we could plan with any precision. It also feels like the amount of time you need to become an expert in something. It's also ~4000 work hours. https://t.co/pu4oZaATiB
7/ Here I strongly disagree. Many people (myself among them) assume incorrectly "well if there are people with disposable income and an idea is good enough it will get funding." https://t.co/x4OGdjM0Fd
8/ <Checks Calendar> 😭 https://t.co/Px7V7T7rXD
9/ See: The Szilard Point https://t.co/YbaYIJ0xLA https://t.co/BusnsHv3Ci
10/ Like so many other perverse-incentive-filled platforms, the only way we're going to get past peer review is if we figure out something that is not just better, but *more convenient.*This will be hard. https://t.co/XvKmsHOBVd
11/ I don't always 100% agree with nobel prize winners, but when I do... https://t.co/DFE1YqN9JS
12/ The extended metaphor of technical language as an actual language is powerful.https://t.co/iXkywl7TMP https://t.co/ijrIrf8UrD
13/ So much going on here. New, eventually correct theories often have serious flaws. We need enough slack in the system to iron them out. But in a competitive, peer-reviewed system, that slack is non existent. https://t.co/JjQy7BJ5E9
14/ Why is there so little slack? Aha! https://t.co/QhlPSptoa9
15/ It is tricky to prove anything conclusively in this area. It's fraught with counterfactuals, nebulosity, and the streetlight effect. I appreciate the willingness to say "either you buy it or you don't." https://t.co/HDlOWzk1TU
16/ Another tricky thing is that for a funding organization to avoid having an "à la carte menu" it is *a lot of work.* You need to be willing to put in a ton of time to understand *how* a researcher is thinking. Braben was willing to do this in Venture Research. Most are not. https://t.co/DAS6NCVJJ2
17/ Repeating the importance of the individual. It's not a popular sentiment now, and yes, we all stand on the shoulders of too many giants to count *but* at the end of the day, ideas need to come from a single brain. Ideas don't create products but they do create paradigms. https://t.co/5VO7hFuKva
18/ LETS ARGUE. https://t.co/qTcWqzEWtJ
19/ "Academic research today is not always distinguishable from industrial research" awkwardly spread across two pages.He means to say that academic research is too industrial but the opposite may be true as well. We've lost an ecosystem that spans the space of research well. https://t.co/xgZZYLBDno
20/ The combination of a drive for efficiency and the unmeasurable nature of research leads to terrible streetlight effects. https://t.co/jMSkMrYarn
21/ Problem: new ideas often come from new researchers who have the least job security. https://t.co/hLZjxsNxyg
22/ THIS was a cruxy part for me. You expect "high risk" things to fail. Expecting something to fail means that you will not put as much effort into helping it succeed *especially when you have little skin in the game* (as is the case for most funding orgs). https://t.co/h1pyDFkBE6
23/ "High risk" implies a known probability distribution weighted towards failure. In reality, we're talking about "high uncertainty" research where by definition the researcher knows the probability distribution better than anybody else.https://t.co/s9Ph8QF5Hm https://t.co/mpdlhZtzIi
24/ High efficiency research programs will not shift paradigms. https://t.co/xqNzT4GGqv
25/ So many suitcase handle words in research funding 😭 https://t.co/yb4q5Wpijg https://t.co/0hncOjP2lk
26/ Here's another fundamental tension - only a few people can do the sort of unfettered research Braben is advocating. Frankly, given free reign most people would just goof off. The system will then be inherently unfair. https://t.co/GsUA8TYcXO
27/ This rings uncomfortably true. You see "progress reports" of cool research all the time but somehow nothing seems to come of them. https://t.co/dr8C3nWBcg
28/ Nature's deadly enemy, compromise! https://t.co/cGkSLphRiw
29/ Braben's description of the UK's funding system might be an argument against the notion that 'breaking up the NSF' into multiple funding agencies would decorrelate funding and lead to good things. https://t.co/5Iz412TEcl
30/ Again, fundamental tensions. We want government spending to be fair and accountable. Venture research is inherently unfair and unaccountable. https://t.co/9db7Xd8jhY
31/ The book also digs into the university education system in general. It's a whole other can of worms, but the negative effects of the seemingly-good increase in students mirrors the effects of the seemingly good increase in PhDs. https://t.co/1nInybI2VU
32/ Another uncomfortable truth is that the people who lead breakthroughs might not be the most likable individuals. A fundamental tension between "no assholes" and "assholes may create new paradigms" https://t.co/Omw2TU6v0l
34/ Different innovation organizations' approach to what I would call "push" vs. "pull" recruiting and their reasoning behind it is worth paying attention to. By contrast, DARPA recruits PMs solely via "you don't call us, we call you" https://t.co/nAMiijCUFz
35/ "Take all the terrible things about how funding agencies do things and do the opposite" https://t.co/LK5kqmMvSn
35/ A festival ... of science! Gatherings that are both highly technical and cross-disciplinary are underrated. https://t.co/uttgwMLEbp
36/ The idea that know-how transfer is a difficult and active process is important.For much more depth on that subject, check out the idea of "manage the transfer, not the technology" in @SafiBahcall's "Loonshots." https://t.co/hvtGeuysPo
End/ If this was intriguing, I absolutely recommend the book. Stripe press recently released a beautiful hardcover:https://t.co/nlcpStEZD3And I recently did a podcast with Donald Braben where we dig deeper into several of these topics https://t.co/ETkHixtLmP