It’s possible to construct a relatively simple algorithm to distinguish superstimulatory / acrasiatic media from novel, educational or insightful content. Such an algorithm need not make use of probabilistic classifiers or machine-learning techniques that rely on my own personal tastes. The distinction can be made based on testable, objective properties of the material.
(~20%)
(This is a bit esoteric. I am starting to think up aggressive tactics to curb my time-wasteful internet habits, and was idly fantasising about a browser plugin that would tell me whether the link I was about to follow was entertaining glurge or potentially valuable. In wondering how that would work, I started thinking about how I classify it. My first thought would be that it’s a subjective judgement call, and a naive acid-test that distinguished the two was tantamount to magic. After thinking about it for a little longer, I’ve started to develop some modestly-weighted fuzzy intuitions that there is some objective property I use to classify them, and that this may map faithfully onto how other people classify them.)
Upvoted because I can’t think of any sense in which it’s possible to reliably separate akrastic from non-akrastic media without a pretty good model of the reader. Wikipedia’s a huge time sink, for example, yet it’s a huge time sink because it consists of lots of educational but low-salience bits; that article on orogeny might be extremely useful if I’m trying to write a terrain generation algorithm, but I’ll probably only have to do that at most once in my life.
On the other hand, it’s probably possible to come up with an algorithm that reliably distinguishes some time-wasting content. Coming up with a set of criteria for image galleries, for example, would go a long way and seems doable.
Wikipedia was one example of a challengingly messy corpus, though I do think there’s a sharp division between articles that make you know more stuff and articles that don’t). I personally wouldn’t consider the orogeny article akrasiatic.
It is possible I’m working from a quite specific definition of akrasia in this case.
Something that disqualifies things that don’t contain keywords from the thing you’re currently working on might function as a very crude version of this.
Come up with 20 independent algorithms like that, use bayes theorem to combine the results, and you should come pretty close.
Irrationality Game
It’s possible to construct a relatively simple algorithm to distinguish superstimulatory / acrasiatic media from novel, educational or insightful content. Such an algorithm need not make use of probabilistic classifiers or machine-learning techniques that rely on my own personal tastes. The distinction can be made based on testable, objective properties of the material. (~20%)
(This is a bit esoteric. I am starting to think up aggressive tactics to curb my time-wasteful internet habits, and was idly fantasising about a browser plugin that would tell me whether the link I was about to follow was entertaining glurge or potentially valuable. In wondering how that would work, I started thinking about how I classify it. My first thought would be that it’s a subjective judgement call, and a naive acid-test that distinguished the two was tantamount to magic. After thinking about it for a little longer, I’ve started to develop some modestly-weighted fuzzy intuitions that there is some objective property I use to classify them, and that this may map faithfully onto how other people classify them.)
Upvoted for this sentence.
Upvoted because I can’t think of any sense in which it’s possible to reliably separate akrastic from non-akrastic media without a pretty good model of the reader. Wikipedia’s a huge time sink, for example, yet it’s a huge time sink because it consists of lots of educational but low-salience bits; that article on orogeny might be extremely useful if I’m trying to write a terrain generation algorithm, but I’ll probably only have to do that at most once in my life.
On the other hand, it’s probably possible to come up with an algorithm that reliably distinguishes some time-wasting content. Coming up with a set of criteria for image galleries, for example, would go a long way and seems doable.
Wikipedia was one example of a challengingly messy corpus, though I do think there’s a sharp division between articles that make you know more stuff and articles that don’t). I personally wouldn’t consider the orogeny article akrasiatic.
It is possible I’m working from a quite specific definition of akrasia in this case.
(You need to put a backslash before any )’s in URLs, e.g. en.wikipedia.org/wiki/Blossom_(TV_series\)).)
(Done)
I would expect to have to do that either zero or at least two times.
I tend to be a one-evolving-draft sort of programmer. Fair point, though.
Upvoted for underconfidence.
Simple compared to what, and with what rates of false positives/negatives?
As in implementable in a couple of hundred lines of JavaScript. If I had good answers for the second question, I’d be a lot more sure than 20%.
Something that disqualifies things that don’t contain keywords from the thing you’re currently working on might function as a very crude version of this.
Come up with 20 independent algorithms like that, use bayes theorem to combine the results, and you should come pretty close.