I have just found out that proteins (at least in some cases) are folded into their functional conformation (i.e. their functional folding), through thermal fluctuations, and that this conformation minimizes free energy. i.e. that the reason that a protein has a certain conformation, is because this conformation minimizes free energy.

I am surprised by this, because if all proteins of a certain type (sequence of amino acids) have the same conformation, this actually minimizes entropy. I would have expected that if we would just leave the protein folding to random fluctuation, it would tend towards a reasonably high-entropy distribution of conformation, i.e. lots of different conformations for individual protein molecules of the same type.

What explains this?