Why 8/12 victory cards? And how did you get to that number?
My guess is that you started playing 3/4 player and wanted a number that was evenly divisible by that group- 12 is the obvious number. And then you shrank it to 8 to make 2p games similar.
Related question: why not 12 of each kingdom card then too? Making them evenly divisible seems fair using the same logic.
Originally there were 12 of every kingdom card and victory card (and I gradually printed more Copper / Silver / Gold / Curses, not knowing how much would be enough). Most of my games initially try to work with 3-5 players, and then I support 2 or 6+ if that works out. In a 5-player game where everyone wants a particular card, you may just end up with one of them. They get $5 on turns 3-4 and you don't, you know. I wanted enough copies of a card that I could expect to get a couple copies if I wanted them. So that was what mattered for a lower limit. And then the upper limit was, I can only print so many cards. I didn't know at the time that the number of cards would be an issue for publication, but man, I didn't want giant stacks of things we weren't buying. So 12 seemed reasonable and I went with 12. Yes, being divisible by 3 and 4 was nice too.
The original game ending condition was any empty pile. Normally it would be a victory pile though. When I learned that the number of cards was an issue - will people buy a box of just 500 cards, no incredibly valuable board or anything - I looked at ways to cut down. One was, lower the action card piles to 10 cards, but change the end condition to any victory pile. You had to leave a buffer you see - if I bought the Remodels down to one left, whoever's winning could buy that to lock in the win. So I have to leave two Remodels. With Remodel not ending the game, having only 10 Remodels was like having 12 had been before. We were getting use out of that last Remodel that never did anything but end the game, plus the Remodel you had to leave as a buffer. But the victory piles were still the end condition so they stayed 12. Then when I changed the end condition to "no provinces or 3 empty piles," I kept the non-Province VP piles at 12, because I felt like, having 12 of a kingdom victory card made it easier to go for that strategy. I wanted those cards to be competitive and having more cards was part of that. Now, Estate for sure did not need 12 and could have just not been a pile. If I had needed to cut cards, it was on the list. Since I didn't, it was 12 because the other VP piles were.
For 2 players you could just have a longer game, but it seemed good to pare it down, so it's 8. For more players you need more Provinces and so I add 3 per extra player to keep it a multiple of the number of players. Possibly 4 per player (so 16 for 4 players) would have been better; my thinking at the time was, more players means a longer game, so maybe it's not so bad to only have 3 per player for 4+. Speed it back up a little.
Curses ended up as 10 per opponent to make it possible to balance Witch over different numbers of players. It's probably 10 because it's a round number; it seemed like enough pain. And then Copper/Silver/Gold just tried to be enough to reasonably handle expansion cards that I already knew were coming.