We don’t, no. That some protein sequence might constitute a local optimum, and mutant variants off it are selected against, doesn’t mean no other sequence could functionally do it’s job.
Even for the most ultra-conserved sequences known, we also know there are individuals among us (and other species with these genes) carrying mutant versions of them, apparently largely unaffected, and hence the sequences are capable of functionally substituting for the canonical sequence and support life. Those mutants might have ever so slightly lower fitness that manifest on much larger timescales (are visible to purifying selection over the course of many generations in a way that matters in evolution) in ways that don’t translate into any immediate quality of life defects.
I’m sorry but your continued attempt to derive contradictions or falsifications from highly contingent and generalized statements, all stand and fall on your own excessively dichotomous thinking, and total lack of comprehension of protein biophysics.
No, a local optimum does not imply a sequence is isolated(how narrow the peak of the hill is does not say anything about how broad the base is, or whether other functions overlap it in sequence space), it just implies it’s the best one among local variants at it’s current function. We had an entire thread devoted to this that I can’t be bothered repeating here.
Edit: fixed link