Minyong Kim July 09 presentation is an excellent intro to the section conj (this is maybe the same as his Cambridge lecture).
See Toen AIM talk on Homotopy types of algebraic varieties. In the last section he discusses a nonabelian Abel-Jacobi map, but it looks different from the map usually appearing the section conjecture. What is the relation? Toen says that his conjecture can be viewed as a generalization to higher-dimensional varieties.
nLab page on Grothendieck section conjecture