,
11 tweets,
5 min read
Read on Twitter
A 🧵 reflecting on @KevinSimler’s interactive essay. It’s fun and easy—read it first!
His post has a sudden plot twist toward the end, and takes on a darker tone.
[SPOILERS of sorts coming]
His post has a sudden plot twist toward the end, and takes on a darker tone.
[SPOILERS of sorts coming]
Kevin’s post interactively illustrates and applies percolation theory. I’m envious and admiring; I’ve been looking for an application of percolation theory for 30 years. He found one right where I’m working now, in understanding scientific progress!
meaningness.com/metablog/how-t…
meaningness.com/metablog/how-t…
The last in Kevin’s series of interactive illustrations shows the effect adding a fraction of “careerists” to scientific knowledge networks. Once you get to a critical percent, they shut down progress. The model isn’t quantitative, but usefully conveys the intuition.
Careerism is a natural consequence of the incentives created by what Richard Feynman called “cargo cult science.”
meaningness.com/metablog/upgra…
meaningness.com/metablog/upgra…
“Careerism” has the implicit implication that science is a GOOD career, and people do it for that reason.
Unfortunately, it’s a terrible career, and one reason we have so much bad science is that many of the people who could best do it choose not to, because they figure that out
Unfortunately, it’s a terrible career, and one reason we have so much bad science is that many of the people who could best do it choose not to, because they figure that out
Conversely, most people currently working in science shouldn’t be. They either can’t do it, or don’t really care to. (Maybe they could or would given better incentives.)
As Kevin observes, going-through-the-motions science has crowded out the real thing.
meaningness.com/metablog/upgra…
As Kevin observes, going-through-the-motions science has crowded out the real thing.
meaningness.com/metablog/upgra…
Making science a not-terrible career has to be part of fixing incentives. But then more people will be attracted who want good careers. So there has to be a filter. Which is what brought us to the current impasse:
Most of what scientists do is artificial activities required to get yourself through endless steps of career filtering, put in place with good intentions to ensure that only good science is funded, but which instead almost completely replaced actual science.
This is so bad that some institutions have started choosing who to fund literally at random. At least that wastes much less time, so if someone competent randomly gets funded, they can do science instead of writing grant applications. (h/t @michael_nielsen)
@michael_nielsen Also, funding by peer review results in group-think and whole scientific fields floating off in a self-perpetuating irreality bubble for decades. Randomness will fund mavericks, mostly crackpots, but some may blow up established dysfunctional disciplines.
We can do better than random, I hope!
We know something about what creates innovative “scenes.”
And we know something about how institutions can reflect on, and upgrade, their operating principles.
meaningness.com/metablog/upgra…
We know something about what creates innovative “scenes.”
And we know something about how institutions can reflect on, and upgrade, their operating principles.
meaningness.com/metablog/upgra…
