On Friday I made my first visit to the Mathematical Institute's shiny new Andrew Wiles building (on the old Radcliffe Infirmary site). The occasion was a talk by Thomas Stocker, co-chair of the IPCC AR5 Working Group 1, presenting an overview of their recently published report (that's the Intergovernmental Panel on Climate Change's The Physical Science Basis).
There was nothing novel in the talk, but it was interesting to see Stocker's perspective on the report. And it was the keynote to a climate science conference and the audience was almost entirely scientists, so he didn't need to waste time on boring background. The questions (about model uncertainty and the risks of ensemble "herding") were also challenging, and I had a quick browse of the conference posters afterwards.
The Andrew Wiles building is nice and shiny - the Penrose tiling at the front entrance is lovely - and has a decent looking cafeteria. And the Mathematical Institute runs a lot of interesting courses... will have to see if I can find something fun to audit.