the open source community produces great software, but does not push the boundaries of the computer science. comp sci is applied mathematics, it is not programming. big difference. some of the best computer scientists rarely program anything other than latex markup.

most of the innovation is trickle down tech from military, nasa, or tech companies.

lastly, while you can make this argument that comp sci is extremely inefficient (oh, i bask in your wisdom...) i suggest you read this article about a successful computer scientist's criticism on another flagging industry for perspective.

Artificial Intelligence - A Modern Approach by Russell and Norvig

The best book for catching up with the trends in AI systems over the past couple of decades.

pretty cool, off topic but it made me think of peter norvig's sudoku solver which is about 100 lines of python code and can solve even the hardest puzzles in <.02 seconds. btw, brute force approaches to diabolical sudoku puzzles are not physically possible, follow link to find out why.

python is fantastic for writing concise, functional, and simple solutions to many search problems.