Woke up mysteriously early — around 7. Couldn't go back to sleep. This turned out to be fortunate, for I ran into wjl on the bus and he reminded me that neelk's thesis proposal was going on at 9:30, which I then went to. Neat stuff, and seems rather pleasantly practical. Saw gustavolacerda's talk at noon about inferring causal models. I certainly don't know enough statistics to fully grok it, but it's also interesting work.