Abstract. Wearing the hat of a topologist, I will argue that despite the (justified) great interest in categorification, the good old "Kontsevich Integral" is even more interesting and highly under-studied. The gist: the Kontsevich Integral behaves well under cool operations that make a lot of 3-dimensional sense.
Handout: image: AKTSummary.png, inkscape source: AKTSummary.zip.
Also, a summary of a previous talk I gave on the same subject is at http://www.math.toronto.edu/~drorbn/Talks/Istanbul-0606/AlgebraicKnotTheorySummary.html.