I think it was Bazzell who used to use local hotel addresses for things that required home addresses for sign ups, but that was a technique to avoid handing out information and not poisoning the well.
Disinformation is a well established tool in OPSEC for targeted searches but I think people overestimate how much it helps in passive surveillance that happens inside systems.
False data can create noise, but it’s not going to break the identity links that actually matter. Your phone number, primary email, recovery methods, financial accounts, and devices are still going to tie everything together.
I’d spend 95% of my effort reducing unnecessary exposure and removing high risk connections instead. If you want to add noise, treat it as a small supporting tactic and not the strategy itself.