Great post: I like your style. The first observation to make is that individuals who make extraordinary contributions are often extremely eccentric, and also the quality of their pronouncements usually has high variance. So you’ve succeeded in increasing my probability estimate that you will say something very worthwhile, though maybe at the price of decreasing the (my) expected value of your average statement.
Yes, a draft version is done already, you can find it on arXiv if you are interested. I’m not sure I would say the argument of the book is that “compression is fundamental to epistemology”, it’s more along the lines of “the problem of building specialized lossless data compressors is a deep and interesting one; if we attack it we will probably find out a lot of interesting stuff along the way”.
Insightful. Will has endorsed “up the variance!” in as many words, but I hadn’t made the connection that explicitly maximizing variance like could be a strategy.
Great post: I like your style. The first observation to make is that individuals who make extraordinary contributions are often extremely eccentric, and also the quality of their pronouncements usually has high variance. So you’ve succeeded in increasing my probability estimate that you will say something very worthwhile, though maybe at the price of decreasing the (my) expected value of your average statement.
Presumably you’re the Burfoot who wrote or is writing a book on compression as fundamental to epistemology?
Yes, a draft version is done already, you can find it on arXiv if you are interested. I’m not sure I would say the argument of the book is that “compression is fundamental to epistemology”, it’s more along the lines of “the problem of building specialized lossless data compressors is a deep and interesting one; if we attack it we will probably find out a lot of interesting stuff along the way”.
Insightful. Will has endorsed “up the variance!” in as many words, but I hadn’t made the connection that explicitly maximizing variance like could be a strategy.