Yes, I guess it matters a lot whether this was quite close to the best they could do or the best they could do given specific resource constraints or whether they just didn’t bother to try doing better (eg to invest more tokens into readable papers)?