Full power low weight GAs are a dangerous manoeuvre. We don't practice them often and they often go wrong. LOSA data showed that one in 10 result in an undesired aircraft state, but I think the rate is probably close to 1 in 5. peekay4 mentions the B757 incident at Oslo which was so very nearly a crash. The crew were subject to the somatogravic illusion which is human limitation. We have not evolved to assess prolonged acceleration and can easily misinterpret it as pitch up. In the high energy GA, the AP is your friend.
In the Thompson case, it was sensible to have the AP engaged. The question that needs to be asked is why the AT didn't command the power required. Intervention from the PF would have resolved the issue, but in the high workload situation of the GA, it would have been best if it had responded as it should.