This is right, but the difference is actually not that great. Scientists and those that accept their conclusions on evolution (I am avoiding use of the bizarre term “evolutionists”) freely acknowledge that it’s way easier to harm a gene or protein through unguided mutation than to effect positive or creative change. We all agree on that. We also agree that sometimes these “damaging mutations” (also not a term we would use that much because the reality can be much more nuanced) can sometimes lead to fitness gains. The only real difference between Behe and the scientific community is that he claims that harming things is pretty much ALL that unguided mutations can do and that to achieve any gain-of-function (and he uses a single disulfide bridge as a regular example), something more than random unguided mutation is needed.
To answer an earlier question in this thread, there are two important pieces of suggestive evidence that the cholesterol-clearing function of ApoB is enhanced in polar bears: one is that they eat a diet very high in saturated fats and yet have low LDL levels and do not develop cardiovascular disease. Their cholesterol clearance must be working very well and the ApoB gene harbors the second-most highly selected variants in polar bears - highly suggestive. Secondly, the variants themselves are mutations that cluster disproportionately in the domain that we know is involved in cholesterol clearance. Again, highly suggestive.
Also, you ask about the thousands of variants and how many of them are “damaging.” That’s just not something we can know without exploring each variant one-by-one in the lab. The computer predictions are just not there yet. They predict the type of mutation (missense, etc.) and then attempt to say if it alters the domain structurally and call it damaging if it looks like it does. You see the problem there, right? While most mutations that cause structural change will weaken the function of the domain, some will bring enhanced or modified function. While more rare, these are more important than your average damaging mutation and so when they do appear, they often get selected for and accumulate in the population. Behe and other IDers are putting way too much stock in the PolyPhen predictions simply because a superficial look at the results support the position that they want to be true. It’s confirmation bias, pure and strong.
I also want to point out that the few times that Behe has acknowledged the creative new-function effect of genetic rearrangements, like occurred in the citrate+ mutations in the LTEE (which he acknowledged) or the tURF13 ion channel (which he hasn’t), he chides scientists for confusing WHAT happened with HOW it happened. So, even when we are lucky enough to capture one of these events in such a way as to figure out with relative certainty how they unfolded, Behe fires back, “yeah, but you can’t show it wasn’t guided.” Here’s a direct quote (although talking about feathers, not molecules): “It never ceases to amaze me that Darwinists like Coyne are unable to separate the question of what happened from the question of how it happened.”
Ultimately, he’s right. Down to the precise molecular events, the breaking of the DNA strands, the DNA polymerase grabbing the wrong nucleotide… we can’t know whether or not the hand of G-d was involved or not. If it was, that leads to some pretty uncomfortable conclusions as far as I am concerned, but if people want to believe that mutations and genetic rearrangements are all intended and executed by a supernatural force, that’s totally fine. But what we can’t do is pretend that this is a valid scientific approach.