Re totems, amulets and "mojo:" A number of my customers who have good ears say that these old-stock parts do sound slightly different from modern silicon transistors of comparable gain. You, yourself, have noted that manufacturing techniques were different back when, so maybe the customers are hearing the result of that. If you do a "Technology Of The Big Muff" article, I'll be happy to provide a few devices for evaluation. Meanwhile, I'm glad that I can accomodate the DIY desire for "authentic" at a reasonable price.
It is true that semiconductor processes have changed a lot over the last several decades. We are perilously near the era when there will be only three main semiconductor processes on the planet, given the need for ever-higher standards of cleanliness, precision, and resolution, and the necessary costs of achieving that.
However, most transistor circuits, like especially the transistor big muff, are designed to hide the variations of transistors. I know a large number of people who swear that they can hear microscopic variations of circuits, including the oxygen content of the interconnecting wire. However, this kind of assertion has always fallen afoul of any scrupulously fair tests in the audio world. In the hifi world, the "golden ears" fared so badly on the first few blind tests that no self respecting Subjectivist will agree to take part in the tests any more. When the tests expose that they cannot in fact do any better than random guessing, they then become sure that the test is flawed, rather than believe they were prejudiced. I would expect much the same thing to happen in the musical world.
My boss placed a largish bet on this recently. He footed the bill for a test that was large and as fair as we could make it without actually isolating participants and doing double blind setups. The results were... interesting.
I would love to do something similar with different sets of magic mojo transistors versus 2N3904s in a Big Muff. But a fair test needs a relatively huge investment in time and effort. You first need a fair sample of listeners, and that needs to be good-ears guitar players. Then you need a setup where they must make selections of better, same or worse with no information other than what they can hear. Finally you have to record the testing so the inevitable deniers can be refuted. It's a huge amount of work. But serious fun...