**Re: Reply Guy Exploit** I can't help but think the filtering needs to be done by the relay/s. That's the only way to keep it transparent. You also probably don't really want clients carrying the processing or code burden. I'd want to run my own relay but I can also see configurable relays as a paid service. Advertisers & scammers are probably the next people to use the RGE.
Slippery slope. Relying on smart servers led us into the master/slave relationship that is prevalent on the Web today.
Its censoring & it's needed. The decision is where should it occur. On the client or via the relays.
On Community Relays. A public square will always lead to master vs slave dynamics. Many interoperable private squares will not.
Every Community caters to different niche audience (which can be just one person, you, btw) and can have their own unique way of curating for that niche.
Trying to find general ways to handle spam client-side for every use case or niche leads to endless discussion and disagreements.
There are options on both sides. Clients could provide better NIP05 filtering, let me mute replies from accounts without a NIP05 and let me blocklist keypairs with NIP05s from certain domains. Paid relays already solve this but getting Users to adjust to them is a bigger ask. We’ve seen them work for 18+ months now but not everyone is willing to stump up money and put their attention behind a paywall.