LLVM and GHC are, at least in my book, currently the most exciting developments in the open-source world. I haven't been this excited to track a changelog since the Linux 2.0 series.
I'm glad this article was posted, because I had missed another article it references regarding Don Stewart's work using genetic algorithms to evolve good optimizations...which was far more exciting than the HN discussion would indicate:
A sample from that article: "Which LLVM passes to use, and which order to run them, with what analysis in between, is a huge search problem, though. There are roughly 200 optimization and analysis flags, and these, mostly, can be run in any order, any number of times. If we want to run say, twenty of the passes (about what -O3 does), that’s what, about 10^46 arrangements!"
