RAD Lab

(This is Mike Jordan). It's interesting to note that while I use my upstairs office for thinking time (generally in the mornings), I'm finding that meetings are often more fun and productive when held in the RAD Lab than when held in my office. There's something about being behind a glass window where other people are walking by and can peer in that enhances the level of energy of a meeting. It also of course leads to other unscheduled meetings before and after the scheduled meeting. I'm coming down to the RAD Lab 3-4 times a day, typically once in the early afternoon and 2-3 times later in the afternoon. People are grabbing me to try out ideas or to ask questions. I'm having a much higher level of interactions than I've had in Soda Hall before.

I've had several times discussions on the white board next to our cubicle, and someone would walk by and ask us what it was about, and we would end up in a discussion with three or four people on the subject. It also happened as I was simply thinking by myself on the board. I find these white boards very effective, only went to the closed rooms a couple times.

It's nice to have everyone together, to see people talking amongst themselves, and getting to know everyone by being in one location. I am thoroughly enjoying that part of it.

Have gotten great quick answers from machine learning students on “what is the canonical reference for X” or “what is the name of the technique that accomplishes X”

I have access to, and talk more with, the ML people. I love the space. It is significantly better than before.

(This is Dave Patterson.) Randy, Zhangxi, and I talked about using FPGAs for invisible layer. Gebis, Armando, and I talked about chip reliability vs. powering off. For reasons that seem mystical to me, meetings in the RAD Lab meeting rooms are much more effective than meetings were in my office. I feel like my productivity has made a noticeable jump in the last month in quality and quantity. Wish we could get someone whose expertise was on such issues to study us, to explain what these phenomenon is, so that we could reproduce it!