
A visible proof trail can support a serious candidate result without settling the physics or independently proving who contributed what.
The argument
Treat the construction, formal correspondence, physical adequacy and model contribution as separate claims. Confidence should rise only where the published checks actually reach.
What is new
Craig now says Astra suggested the finer-grid method and led the detailed proof. The October preprint, Lean code and supplementary checks make the result unusually inspectable, though not the precise contribution split.
Why it matters
A correct formal derivation can still leave physical criteria unresolved. Reflection positivity, practical simulation and the interacting continuum theory remain outside the settled claim.
LecXie’s practical interpretation
Publish a verification ledger: proof replay, statement correspondence, physical tests and versioned contribution records. Do not let success at one layer certify the others.
Evidence & limits
The technical artefacts are public, but the collaboration narrative comes from Craig and OpenAI. Named experts challenged the work; no independent validation or contribution reconstruction was available.
