The concept of information complexity, though introduced before Chakrabarti joined Dartmouth, is being brought to fruition as part of his work here. This concept has already proved to be very useful in a diverse range of problems in complexity theory and is expected to continue to yield new exciting results. Chakrabarti's own optimal results about high dimensional approximate nearest neighbour searching are a good example.
The work on approximation algorithms for unsplittable flow has influenced other researchers in their work extending the results, as discussed above.
The Dartmouth Theory Seminars brings together graduate students in large numbers and exposes them to new theoretical ideas in a way that no course at Dartmouth currently does.
The larger goal of any theoretical work remains the same as in eras past: to deepen humanity's understanding of the fundamental process of computing.