MIT Everyware
TeachingMachines writes "David Diamond has written a very readable article at Wired News titled MIT Everyware that follows up on MIT's OpenCourseWare initiative (previous story). It turns out that one of the most popular courses has been '6.170 Laboratory in Software Engineering, Fall 2001.' Diamond notes that '[u]ltimately, MIT officials know, OpenCourseWare's success depends on the emergence of online communities to support individual courses.'"
Algorithms is pretty much finished. No one is coming up with anything better than heapsort and the list/tree/hash algorithms are pretty much done as far as research goes.
Crypto is done. The basics are already hashed out (pun!) all that's left is individual implementations of it.
AI is a dead end as is proved over and over.
What's left?