All posts

The Texas Sharpshooter Fallacy « You Are Not So Smart

September 14, 2010
If you have a human brain, you do this all of the time. Picking out clusters of coincidence is a predictable malfunction of normal human logic.

We see patterns everywhere, our brain is hardwired to do it. Most of them are harmless or even beneficial. The problem is when people start assigning meaning to the patterns, as explanation of whatever they want to believe or to convince other people of (such as "vaccines cause autism", which is completely false).

Cfengine3 lexer for Pygments

September 8, 2010

I have written a Cfengine3 lexer for Pygments, the open source syntax-highlighter used by Gist and many other sites. It seems to work fine on all the cfengine policy files I have tested, but if you find anything that doesn't quite work as expected, please let me know. This is my first-ever Pygments lexer, so if you are an expert and can advise me on better ways of doing things, I'd very much appreciate the feedback too.

Google Scribe-generated prose (aka "Google Scribe channels Kanye West")

September 7, 2010

Google has launched a new labs project called Google Scribe, which essentially provides an auto-completion service. You type, it autocompletes. The following text was generated by typing a single space, and from then on simply accepting the default suggestions (pressing "Enter" every time the suggestion popup appeared). I had to stop at some point, but it seems like a good source of mindless entertainment. Also, the result varies depending on the initial seed word you provide, and also on whether you just accept blindly (Enter) or type spaces at certain places. These results also provide some insight into the training texts that were used ("I'ma let you finish...")