OpenAI announced results on local smoothing, the Exponential Time Hypothesis and quantum field theory, alongside external academic examination. Its public ETH manuscript describes counting solutions to fixed-width formulas in subexponential time and polynomial space, explicitly noting that this does not give subexponential algorithms for all NP problems or prove P=NP. OpenAI reports Lean formalization of the mathematics and ETH results, but independent examination and practical algorithmic improvements remain separate tasks. The physics group has deferred release while preparing its own exposition. Selective early access also creates information asymmetries; these announcements should not be presented as final acceptance of every claimed breakthrough.