Suppose I have a functor f: C \to D of infinity categories and let p: E \to \Delta^1 be the associated cocartesian fibration. Suppose that p is bicartesian, so that f has an adjoint g: D \to C. I have seen the claim that the precomposition with a cartesian edge s_Z: g(z) \to z, z \in C = E_0 factors through the functor g. More precisely, precomposition with
s_Z defines a map map_{E_1}(z,w) \to map_E(g(z),w)). I have seen the claim that this is homotopic to:
map_{E_1}(z,w) \to map_{E_0}(g(z),g(w)) = map_E(g(z),g(w)) \to map_E(g(z),w) where the last map is postcomposition with s_w: g(w) \to w…