most people optimize way too hard for legibility. they try to get all the most prestigious degrees or jobs or awards or papers or whatever. these people should at some point realize that it Doesn’t Matter, nobody actually cares, at some point the credentials are only shiny badges that don’t actually do anything. all the things that matter are gated behind actually being good at things. they should stop chasing the badges and instead actually go and do something they believe in.
but also some people fail in the opposite direction. they never do anything that is legible at all, and even if they are actually good, they end up indistinguishable from cranks. these people should go and do like one really impressive thing in some domain that they even vaguely believe in. publish a really cool paper, get one very prestigious job and hold it down for a year, get a million subscribers on some social media platform, make a lot of money, whatever.
Unfortunately I think it takes more like a career than a year to do one clearly legibly impressive thing. If you’re brilliant you can do something marginally legibly impressive in a year, but it’s going to be pretty subjective. So more like ten years to do clearly impressive stuff and even then the value of your illegible work will be highly debatable.
This is a big issue in alignment. Doing work that’s verifiably progress gives much more immediate rewards than actually trying to solve the very complex overall problem. So it’s hard to blame people for mostly searching under the steetlight in alignment.
Presumably varies a lot by domain. (Or equivalently, maybe in some domains the “do something you believe in” advice will lead you to rationally do legible stuff like accumulate credentials, because other paths will be transparently ineffective.)
Although if you have very short timelines(or even a moderate probability thereof) it might not make sense at this point to invest effort in legible things that aren’t directly on your subjectively most promising path to impact.
they try to get all the most prestigious degrees or jobs or awards or papers or whatever. these people should at some point realize that it Doesn’t Matter, nobody actually cares,
Someone like Richard Ngo[1] would make the point that it is bureaucratic processes (or artifacts of cultural inertia?) which force people to obtain enough credentials for certain positions, the more credentials the better the positions which become available.
Who compiled a curriculum on political philosophy, which I consider to be biased towards right wing-like derision of the managerial state and other institutions. However, I am not sure of how popular the bias is among LW.
Pushing back on “most people“ - it’s “most people around these parts.” Most of humanity is doing low skill work.
As for the remaining elites—I don’t know, I think chasing legibility is likely the dominant strategy? People in these spaces like to think that it’s not that important for success, but having looked at the profiles of dozens of rats and EAs, I’m convinced it is. The mainstream list is “T1 grad or PhD”; but if you add:
-Started a post-grad education
-Started a highly successful/multi-employee business
-Skilled programmer
-Incredible, legible autodidacty (eg Yudkowsky)
you cover ~95% of them, and the remainder are mostly liberal arts college grads.
I think it’s likely that “Doesn’t Matter” is actually just “Doesn’t Matter Anymore.” Credentials seem to actually be a good signal of competency; they just don’t scale well to the furthest heights of human accomplishment, akin to how common IQ measurements break down around 130 or so.
thoughts on legibility
most people optimize way too hard for legibility. they try to get all the most prestigious degrees or jobs or awards or papers or whatever. these people should at some point realize that it Doesn’t Matter, nobody actually cares, at some point the credentials are only shiny badges that don’t actually do anything. all the things that matter are gated behind actually being good at things. they should stop chasing the badges and instead actually go and do something they believe in.
but also some people fail in the opposite direction. they never do anything that is legible at all, and even if they are actually good, they end up indistinguishable from cranks. these people should go and do like one really impressive thing in some domain that they even vaguely believe in. publish a really cool paper, get one very prestigious job and hold it down for a year, get a million subscribers on some social media platform, make a lot of money, whatever.
Unfortunately I think it takes more like a career than a year to do one clearly legibly impressive thing. If you’re brilliant you can do something marginally legibly impressive in a year, but it’s going to be pretty subjective. So more like ten years to do clearly impressive stuff and even then the value of your illegible work will be highly debatable.
This is a big issue in alignment. Doing work that’s verifiably progress gives much more immediate rewards than actually trying to solve the very complex overall problem. So it’s hard to blame people for mostly searching under the steetlight in alignment.
Presumably varies a lot by domain. (Or equivalently, maybe in some domains the “do something you believe in” advice will lead you to rationally do legible stuff like accumulate credentials, because other paths will be transparently ineffective.)
Although if you have very short timelines(or even a moderate probability thereof) it might not make sense at this point to invest effort in legible things that aren’t directly on your subjectively most promising path to impact.
Someone like Richard Ngo[1] would make the point that it is bureaucratic processes (or artifacts of cultural inertia?) which force people to obtain enough credentials for certain positions, the more credentials the better the positions which become available.
Who compiled a curriculum on political philosophy, which I consider to be biased towards right wing-like derision of the managerial state and other institutions. However, I am not sure of how popular the bias is among LW.
Pushing back on “most people“ - it’s “most people around these parts.” Most of humanity is doing low skill work.
As for the remaining elites—I don’t know, I think chasing legibility is likely the dominant strategy? People in these spaces like to think that it’s not that important for success, but having looked at the profiles of dozens of rats and EAs, I’m convinced it is. The mainstream list is “T1 grad or PhD”; but if you add:
-Started a post-grad education
-Started a highly successful/multi-employee business
-Skilled programmer
-Incredible, legible autodidacty (eg Yudkowsky)
you cover ~95% of them, and the remainder are mostly liberal arts college grads.
I think it’s likely that “Doesn’t Matter” is actually just “Doesn’t Matter Anymore.” Credentials seem to actually be a good signal of competency; they just don’t scale well to the furthest heights of human accomplishment, akin to how common IQ measurements break down around 130 or so.