The behavior is an increase in execution time when adding or removing FCP built-in LUTs that scales quadratically with the number of clips being adjusted. Back on FCP 10.8.1 and Ventura 13.7.8 this was over 10x faster than FCP 12.2 on Tahoe 26.4.1. I don't know when it got worse. Here are some basic numbers:
Approx. Elapsed Time to Add/Remove Camera LUT
on Sony MXF clips, FCP 12.2, Tahoe 26-4-1, M1 Ultra
50 clips: 16.8 sec
100 clips: 52.6 sec
200 clips: 213 sec (138 sec on FCP 10.6.10, 145 sec on 10.8.1)
Approx. Elapsed Time to Add/Remove Camera LUT
on Sony MXF clips, FCP 10.8.1, Ventura 13.7.8, 2017 i7 iMac 27
50 clips: 6.4 sec
100 clips: 11.2 sec
200 clips: 31 sec
It is caused by a complex interaction between the FCP "Flexo" framework and the macOS "Foundation" framework. The causal factor in FCP has long existed, at least back to 10.6.10.
In software engineering terms, it is called a "structural O(N^2)" or sometimes a "a hidden or accidental O(N^2)," where execution time increases with the square of some term. That can be seen in the above table, where execution time to add or remove LUTs increases roughly with the square of the number of clips adjusted.
This is often not the result of a single inefficient algorithm but rather of how different layers interact. This is the case with FCP and macOS. There is a code path in the Flexo framework that has a quadratic cost, but it previously wasn't that bad for the end user. However, it was calling a function in the macOS Foundation framework, which itself became significantly slower at some point. That slower behavior did not significantly impact most apps. However, because the FCP code path handling LUT add/remove called it a lot, this showed up dramatically in FCP.
Even on FCP 10.6.10 on Ventura 13.7.8, the time to change LUTs rapidly escalated with number of clips changed, just starting from a lower number than FCP 12.2 on Tahoe 26.4.1.
As described by bayoyong2, the real-world impact on FCP is major. It basically means you cannot freely import lots of clips (which often get auto-assigned LUTs), then freely drop those LUTs in a single batch. It will cause an FCP "beachball" hang which varies with the square of the number of clips adjusted. It rapidly becomes impractical, plus some users would think FCP had locked up and would force quit or take other steps.
It's tempting to say "FCP should add a bypass feature so it only adds LUTs on import if you enable the feature." But FCP has long had automatic LUT feature, and adding another config option because the feature is basically broken isn't the best way. I can't imagine Apple ever doing that.
I have to study this more, than I will probably file two bugs, one on macOS for the Foundation framework and one on FCP for the Flexo framework's handling of multi-clip LUT add/remove.
I don't think this is a simple fix, even if done entirely in FCP, because it would seem to require a significant rewrite of some Flexo framework code. See below Instruments traces.
https://photos.smugmug.com/photos/i-TrM3xtx/0/LSHqv59T2dWMpshQ2c8bMgL9dbthhPjhMMsZdtpkJ/X5/i-TrM3xtx-X5.jpg