Again, great work. This has been going around in my mind for a day or so and here’s some observations:
And 130 isn’t all that “special“ once you find yourself being a white (+6IQ) college graduate (+5IQ) atheist (+4IQ) who’s ”NT” on Myers-Briggs (+5IQ).
If IQ correlates with all that, why are we filtering by all that AND IQ as if they were orthogonal?
Also, you may want to try filtering by age groups. Younger people tend to be more atheist and have time on computers and the internet. And if we get to them before they join a hedge fund, then they won’t end up ‘winning so hard they won’t have time to do anything meaningful’.
Moreover, you may want to add some gray to your filters. If 5% of lesswrongers self-reported to be outside the 130+ IQ bracket, and maybe sampling bias makes that 10%, then perhaps this should be reflected in the filter.
Another one that I just thought about, what about people who read and understand -some- of the sequences, and get some value out of them but don’t end up signing up and participating? Are they a waste of resources? As far as I know from experiences, people who participate in a community tend to be the peak of the iceberg (1-10%?), and the effects of this this self-selection should not be ignored.
Finally, to add to the uncertainty, an anecdote: Some people tend to be open to convincing if they find the right arguments. I self-identified as a christian when I started reading the sequences and came out an atheist, and I think there’s quite a few of us if I recall correctly from the rationalist origin story thread. Given that, I think the religion filter needs a fair bit of gray added to it.
Overall I think this estimate has swung in the opposite direction. I would be very surprised if the sequences had reached 50% of everyone who would benefit from them. At the very least, this fact would imply something counter-intuitive about the habits of our target audience. (e.g. that they tend to cluster and so LW has saturated word-of-mouth or news outlets where the target audience gathers, without even trying, but further trying would be unproductive.)
Again, great work. This has been going around in my mind for a day or so and here’s some observations:
If IQ correlates with all that, why are we filtering by all that AND IQ as if they were orthogonal?
Also, you may want to try filtering by age groups. Younger people tend to be more atheist and have time on computers and the internet. And if we get to them before they join a hedge fund, then they won’t end up ‘winning so hard they won’t have time to do anything meaningful’.
Moreover, you may want to add some gray to your filters. If 5% of lesswrongers self-reported to be outside the 130+ IQ bracket, and maybe sampling bias makes that 10%, then perhaps this should be reflected in the filter.
Another one that I just thought about, what about people who read and understand -some- of the sequences, and get some value out of them but don’t end up signing up and participating? Are they a waste of resources? As far as I know from experiences, people who participate in a community tend to be the peak of the iceberg (1-10%?), and the effects of this this self-selection should not be ignored.
Finally, to add to the uncertainty, an anecdote: Some people tend to be open to convincing if they find the right arguments. I self-identified as a christian when I started reading the sequences and came out an atheist, and I think there’s quite a few of us if I recall correctly from the rationalist origin story thread. Given that, I think the religion filter needs a fair bit of gray added to it.
Overall I think this estimate has swung in the opposite direction. I would be very surprised if the sequences had reached 50% of everyone who would benefit from them. At the very least, this fact would imply something counter-intuitive about the habits of our target audience. (e.g. that they tend to cluster and so LW has saturated word-of-mouth or news outlets where the target audience gathers, without even trying, but further trying would be unproductive.)