We review recent progress in the study of timelike Compton scattering (TCS), the crossed process of deeply virtual Compton scattering. We emphasize the need to include NLO corrections to any phenomenological program to extract Generalized Parton Dist
ributions (GPDs) from near future experimental data. We point out that TCS at high energy should be available through a study of ultraperipheral collisions at RHIC and LHC, opening a window on quark and gluon GPDs at very small skewness.
We discuss the production of two pion pairs in photon collisions at high energies. We calculate the according matrix elements in k_T-factorization and discuss the possibility to reveal the existence of the perturbative Odderon by charge asymmetries.
