I feel a frustration with the section that begins “this is not fair treatment of things people actually said, but an unfair treatment that illustrates a point.” On one hand, it’s not unreasonable to do that.
But, those three cases seem like areas where it sounds like it would have been totally practical to spend an hour touching base with each person in question, clarifying what they meant and getting their permission to give an account that could be straightforwardly described as a “fair.”
This seems like a) reasonable effort to invest in a fairly Big Written Work, and b) I have this impression that a nontrivial amount of high level discussion between Eliezer and others has some talking-past-each-other going on (sometimes primarily in one direction, sometimes in others).
So getting the nuances right (and putting in some extra effort to more obviously be getting the nuances right) seems important, for a book that Eliezer is writing targeting people in his filter bubble, arguing that they are commonly making some particularly important mistakes.
And maybe this means the anecdotes don’t end up clearly illustrating the point as well… and maybe the point is still a good point, but if it turned out that the people that seemed to be Eliezer to be missing the point, were either not missing it, or missing some different point, that’d be important.
And, meanwhile, if the original point being made turned out to be exactly right, you wouldn’t need to disclaim it with a “this may not be a fair account”, which would have been more compelling to me.
I wonder if it would have been as frustrating if he had instead opened with “The following are very loosely based on real conversations I’ve had, with many of the details changed or omitted.” That’s something many writers do and get away with, for the very reason that sometimes you want to show that someone actually thinks what you’re claiming people think, but you don’t actually want to be adversarial to the people involved. Maybe it’s not the fairest to the specific arguments, but the alternative could quite possibly turn out worse, or cause a fairly large derail from the main point of the essay, when you start focusing on the individual details of each argument instead of whatever primary pattern you’re trying to tie them together with.
I think Eliezer may have been too modest(!) in describing the treatment as unfair. I think I recognize Startup Founder 1, and that looks very much like a conversation I’d expect the two of them to have.
I expect that Eliezer had more evidence than he conveys for the hypothesis that Startup Founder 1 was engaging in blind empiricism. But I have doubts about whether Eliezer was wise to reject hypotheses about why Startup Founder 1′s advice might be right. Here are some mistakes that can be avoided by the “release early, release often” attitude:
creating overly elaborate hypotheses, rather than looking for hypotheses that can be tested cheaply.
being overconfident in one’s ability to model users.
identifying with one’s announced plans in a way that leads to them becoming a substitute for implementing the plans; I suspect this sometimes creates an aversion to exposing the plans to possible falsification.
I imagine that Startup Founder 1 suspected that Eliezer was making at least one of these mistakes, but couldn’t articulate strong evidence for those suspicions.
I feel a frustration with the section that begins “this is not fair treatment of things people actually said, but an unfair treatment that illustrates a point.” On one hand, it’s not unreasonable to do that.
But, those three cases seem like areas where it sounds like it would have been totally practical to spend an hour touching base with each person in question, clarifying what they meant and getting their permission to give an account that could be straightforwardly described as a “fair.”
This seems like a) reasonable effort to invest in a fairly Big Written Work, and b) I have this impression that a nontrivial amount of high level discussion between Eliezer and others has some talking-past-each-other going on (sometimes primarily in one direction, sometimes in others).
So getting the nuances right (and putting in some extra effort to more obviously be getting the nuances right) seems important, for a book that Eliezer is writing targeting people in his filter bubble, arguing that they are commonly making some particularly important mistakes.
And maybe this means the anecdotes don’t end up clearly illustrating the point as well… and maybe the point is still a good point, but if it turned out that the people that seemed to be Eliezer to be missing the point, were either not missing it, or missing some different point, that’d be important.
And, meanwhile, if the original point being made turned out to be exactly right, you wouldn’t need to disclaim it with a “this may not be a fair account”, which would have been more compelling to me.
I wonder if it would have been as frustrating if he had instead opened with “The following are very loosely based on real conversations I’ve had, with many of the details changed or omitted.” That’s something many writers do and get away with, for the very reason that sometimes you want to show that someone actually thinks what you’re claiming people think, but you don’t actually want to be adversarial to the people involved. Maybe it’s not the fairest to the specific arguments, but the alternative could quite possibly turn out worse, or cause a fairly large derail from the main point of the essay, when you start focusing on the individual details of each argument instead of whatever primary pattern you’re trying to tie them together with.
I think Eliezer may have been too modest(!) in describing the treatment as unfair. I think I recognize Startup Founder 1, and that looks very much like a conversation I’d expect the two of them to have.
I expect that Eliezer had more evidence than he conveys for the hypothesis that Startup Founder 1 was engaging in blind empiricism. But I have doubts about whether Eliezer was wise to reject hypotheses about why Startup Founder 1′s advice might be right. Here are some mistakes that can be avoided by the “release early, release often” attitude:
creating overly elaborate hypotheses, rather than looking for hypotheses that can be tested cheaply.
being overconfident in one’s ability to model users.
identifying with one’s announced plans in a way that leads to them becoming a substitute for implementing the plans; I suspect this sometimes creates an aversion to exposing the plans to possible falsification.
I imagine that Startup Founder 1 suspected that Eliezer was making at least one of these mistakes, but couldn’t articulate strong evidence for those suspicions.