Misc.

I had to declare email bankruptcy today. What did me in was not spam alone, but a combination of spam and high-volume mailing lists that might as well have been spam. I've unsubscribed from numerous email lists and I'm returning to my email being just for communicating with me, personally.

I recently changed the website over to use FreeSans. I don't know how it looks on Windows or Mac, but it looks really nice on Linux.

Apparently, I was just ahead of the curve. Appropriate for a blog started in 2000, I suppose. (via dangerousmeta!, which still amuses me with its name. It's the exclamation mark that really sells it.

Following up in a long-standing interest in online community formation, I notice that all of the supposed improvements on Slashdot like Digg and Reddit have long since decayed into democratically-moderated mediocrity, whereas Slashdot over the same time period was and remains a reasonable reflection of the community founders. Slashdot: 3, competition: 0.

Amazon MP3 Downloads: Yeah, I'm late to the party, but: Non-DRM'ed MP3s downloaded through the web browser (with optional program assist for Windows and MacOSX), for iTunes' price or better. All I can say is: What took so long? (Linux support is coming.) The selection will remain relatively poor until the other labels come onboard, but in the meantime I'll take what I can get. Encoding is stellar and the metadata in the MP3 is beautiful. If anything can dethrone iTunes, this is it. Amazon has continued to impress; they've done anything but rest on their laurels.

Spontaneous brain activity detected causing unforced errors. I think I know what this feels like; I suspected something like this existed. See also: utterly random muscle jerks. One thing I find amusing is the blog reactions; you can't leap to the conclusion this is a bug in human cognition. It may be, but you can't leap to that conclusion. The same random processes that cause occasional muscle jerks may also cause flashes of brilliance, or motivate an otherwise-actualized human to do something despite being content. The same goes for that execrable "conservative vs. liberal" study (the execrable part being the "vs." angle, not the study) after twenty flashes of the letter W, it is quite arguably rational to assume the next will also be a W. (Bayesian analysis would give it a very high probability under all reasonable priors, for instance.) Responding more quickly may reflect a difference in the learning rate, but the mathematical meaning of that term and the English meaning differ. Learning rates aren't "right" or "wrong", they only can be judged in the context of a given task, and drawing conclusions based on one rather abstract study is a travesty of science. It's an interesting data point, which is far from a proof of any meaningful statement, let alone the conclusions drawn. Cognition is complicated; if our naive ideas about how it worked were sufficient to explain it, we'd have built a working brain by now. Things that are self-obviously bugs are often features, things that appear to be features are often bugs, and most things are a little of both at once.