On thursday I started a new job, at adverplex cogolabs. It's a computational advertising startup, in kendall square. I've been there four days now, and I like it a lot. The work is interesting, the people are nice, and there are cookies. Leaving bbn was hard, and I'm going to miss the people there a lot. Gaming on fridays and bridge tuesdays with swatfolk I'll miss especially (though I may go back some). No more ontonotes (which is now available free from the ldc to non ldc members). No more C++. Still python, and now, ruby. I am happy to no longer be working for a defense contractor.

One of many changes is working in a cubicle instead of an office. At bbn, everyone had offices, which I understand is unusual. Switching to a cube should be downgrade, but that's not how I've been finding it. I like the slightly social environment of cubes. Mostly it's very quiet, everyone working on their own. But if you need to talk to someone, they're right there and you can see with a quick glance if they're busy. You have a background sense of what's going on in the group which I didn't have at bbn. For management people I suspect offices make a lot of sense because you have a dedicated space for meetings where you won't bother neighbors. And I suspect bbn employees would raise hell if it were decided that people would no longer have offices, but I'm happier without one.

Another difference is that I'm now paid partly in stock options. I have options for a (small) fraction of the company that vest linearly over time. This makes them somewhat analagous to salary; I am continuously getting options for a greater fraction of the company. After talking to friends about how to value options, I've determined that my best guess at their expected value has them making up a little over a third of my pay. There's quite a large chance that they will be worth nothing. There's some chance they they will be worth some, and a small chance they will be worth a whole lot. So chances are I would have made more money staying at bbn. Averaging expected earnings over all imagined outcomes, though, I should earn more now. In my normal life I am risk averse. Julia and I would like to be able to pay rent next month, it would be better if she did not have to drop out of grad school, we'd like to buy a house in the medium term future. We have much less use, for ourselves, for the large amounts of unlikely money that risky investments (stock options) can make. For charity, though, we ought to be risk neutral. We want to do as much good as we can, averaged over all possible futures. So I'm going to stop giving away 30% of my salary and instead give away any profits from the stock options. This is a medium increase in expected donation value and a large increase in risk, but if I'm correct in valuing risk as unimportant in giving, it's what makes sense.