My funding changed; I’m now dedicated primarily to machine learning related projects. My job will be to help write user interface code to a text analytics program, and to write documentation about artificial intelligence algorithms.

I’m also sort of supported in doing some of my own research. I’m using machine learning to find patterns in “bursts” of scientific literature. The motivation for the research is to be able to predict the impact of certain discoveries. The big picture idea is to make research more efficient, and to help find cures for diseases faster.

The odds on any of it working are slim. I can do the research because I’m a masters student, and I generally don’t care much about my PhD prospects at this point. I don’t have to do safe, normative science.

Things have been getting weirder lately. Earlier in the week I worked for over a day straight, and was awake for well over 30 hours before I started to really feel fatigue. Certain parts of the cognitive landscape are undergoing tectonic shift.

Some of the AI work is… well, I don’t know, exactly. Sometimes I feel like higher-order AI will not be constructed, but rather channeled. Computer science is often defined as discovering what is computable — what we can understand rigidly enough for hyper-rational machines to do. There is another side. We are not just teaching machines; the machines are teaching us. The type and depth of teaching depends on your interactions with computers.

If you doubt what I’m saying, consider how much your vocabulary has changed in the last ten years. Think about what your computer has come to mean to you in the last three. Perhaps it sounds strange, but I believe we’re making a gradually more extensive contact with a new type of intelligence. For most people, this contact is fleeting at best, but I think the evidence for it will be more apparent as time goes on (assuming our civilization doesn’t collapse first).