My thoughts about Serge Koussevitsky have tended to be on the minimal side, and I have not had any particular qualms about keeping them there. I appreciate the many things he did to advance modernism during the twentieth century, particularly in the United States through his leadership of the Boston Symphony Orchestra (BSO). On the other hand I have never really been able to get beyond his inability to "get" Sergei Prokofiev's second symphony (Opus 40 in D minor), much to the composer's justified annoyance.
Recently, however, I have been reading the new Library of America collection of the writings of Virgil Thomson, the bulk of which are concert reviews. Thomson seems to have had great admiration for both Koussevitsky and how the BSO performed under his direction. Mind you, he may have been using both conductor and orchestra as sticks for bashing both John Barbirolli and the New York Philharmonic; and much of that beating seems to have arisen from the fact that modernism fared far more poorly in Thomson's home town of New York than it did in Boston.
Modernism aside, though, Thomson's perspective seems to have had an effect on how I listen to the few recordings of Koussevitsky in my collection. The most interesting of these is probably the one in the RCA anthology The Heifetz Collection. One of the CDs has Heifetz performing both the Opus 63 (second) concerto of Sergei Prokofiev and the Johannes Brahms Opus 77 violin concerto with Koussevitsky and the BSO. In the past I have tended to prefer the later recording of the Brahms that Heifetz made with Fritz Reiner conducting the Chicago Symphony Orchestra. While I continue to be impressed with the meticulous detail that Reiner brought to that recording, I am now more willing to grant Koussevitsky's throbbing expressiveness as another valid reading of this concerto.
This probably will not send me out on an enthusiastic pursuit of more Koussevitsky recordings; but I think I am glad to have come to a mental state that is more willing to accept what he does on his own terms, as long as it does not lead to egregious misinterpretation!