While there may not be any direct evidence to cite, one can infer from some of the available information.
The first Block 5 booster, was basically B1046. (Which flew 4 times, and was lost on purpose in the Dragon In Flight abort test). They are more or less sequential to production order I believe so the numbers on the cores offer information.
The current fleet leaders are:
- B1051 - 11 flights
- B1049 - 10 flights
- B1058 - 10 flights
- B1060 - 10 flights
Right off the bat, the Block 5 boosters were doing a pretty good job on reuse. 1046 getting to 4 flights, and would had more, but for the Abort being a non-recoverable use case.
I think that B1049 hitting 10 flights, on the 4th build of the booster is a pretty good sign that the work done to get to Block 5, really nailed most of the big issues. Similarly with 1051 making it to 11 flights on the 6th booster of the type built...
No doubt there are subtle software changes happening, but hardware wise it seems like they got it right pretty early. And then to maintain human rating from NASA, they likely have to be really careful with any hardware changes. But there has not been much publicly discussed in this realm.
The focus with Falcon 9 seems to have been on getting better at the time between flights possible from a pad. (Pad refurb, TEL issues). Then they spent a lot of effort on Fairing recovery. (Those super cool boats with the huge nets, that they gave up on). Eventually they figured out that they can land them safe enough on the water to recover. No need for the big nets.
The flight rate has clearly gone up and up. From 9 flights in 2016 to 21 with the Block 5s starting ins 2018. To 31 in 2021, and plans for 50+ in 2022.
Thus it seems that while the core boosters may not be evolving the same way they did earlier in the program (Merlin 1C->1D->1DFT) the program itself is growing in ways that matter to its success.
Also Starship is the future anyway. Long live Starship. (It is going to be so awesome to see that fly!)