I don’t think it’s binary like that, the fewer parameters that are available for fingerprinting the harder it will be to tell users of a certain browser apart. When arkenfox is talking about naive vs advanced scripts, I believe they’re talking about whether a script will just eat a randomized value or not. If it doesn’t, but you’re the only one randomizing, then it becomes a tracking vector. If the feature is built into the browser and a lot of users are randomizing the same way, then it basically just has to throw that value away because it will be useless for fingerprinting. Firefox, Brave, and Safari all have randomization features built in that lots of people are using, so there’s your “crowd.” Couple that with Private Relay and the fact that Safari is only available on Apple devices with minimal hardware differences between people on the same model of device, and I think there’s a really great recipe for fingerprinting resistance.
You can try out fingerprint.com, it seems to struggle with Safari and they’re a commercial fingerprinting company.