A quantum-cryptography paper on efficient unclonable encryption includes an unusually direct authorship note: GPT-5.6 Sol Ultra found the proof during an extended conversation with Seyoon Ragavan and drafted a preliminary version of the paper. The author states that he is fully accountable for correctness.
The paper claims the first efficient, plain-model, one-time information-theoretically secure unclonable encryption scheme for one classical bit without the earlier efficiency or security-loss caveats. Its proof uses the commutation and anticommutation structure of the Pauli group and a standalone linear-algebraic lemma.
The defensible story is a new research workflow. The model participated in a long expert dialogue, generated a proof candidate, and accelerated drafting; the researcher selected, checked, revised, and signed the result. That is stronger evidence than a one-shot answer, but it remains a human-accountable paper until independent experts validate the proof.