I think the parent comment is saying that it's impossible to arrive at a specific folding endpoint because that state is dependent on continuously changing environmental variables.

Take a look at the configs for Amber (molecular dynamics simulation -- https://ambermd.org). QC might help map the space of inputs that would converge, but it probably couldn't identify a hypothetical 'done folding' state for any given protein.