Oddbean

The goal is to make such things impossible or irrelevant. Know the system and work around it. One way is for every user to have their own proxy relay that directly connects peer to peer to other users' proxy relays, running on sovereign personal servers in the cloud (or in your home + an IP-cloaking remote proxy). This is what we're enabling at https://vaporware.network Another way is going extremely hard on mesh networks and various darknet protocols. There is always a way. Privacy will win. https://image.nostr.build/46346879bfd9b53fdd4f85597a8c4585d95486caee746e255bc16241fa20aad9.jpg You can read the rest at blog.vaporware.network

I like the idea of individual relays being as easy and cheap and commonly used as possible. For the reasons you mention (censorship by centralized relays becomes effectively irrelevant if we can just route around them), and also bc if we want web of trust to work, we’re going to want something like a personal relay to keep track of our WoT. Why? because calculating WoT on the fly is not going to work for anything beyond the simplest and most basic of implementations. A personal relay would be perfect for this. #WoT

Yes exactly. You're getting it. I had worked on a WoT style project on another decentralized network - one in which each user has a full VM (which can do your WoT computation, as you rightfully point out as a need). You may be interested in reading a summary of it as it stood at the time: https://gist.github.com/vcavallo/e008ed60968e9b5c08a9650c712f63bd (I plan on rewriting this for a Nostr context) The language is specific to the platform (Urbit) but you can glean the general ideas. We're working on bringing the "personal algorithm" and WoT idea to fruition generally, for all decentralized social platforms.

Awesome. I will def take a look! 😃 I am in the process of demonstrating what I’m calling the influence score as a method of nostr content stratification and discovery. You can see it in action at https://brainSToRm.ninja where I use the influence score to stratify wiki articles on nostrapedia. One of the advantages I hope to convey of the influence score is that it can see past 2 hops on one’s social graph, unlike most WoT score implementations that give a score of zero for anyone more than 2 hops away. And a second advantage of the influence score is that it is not a popularity contest: your score can be high even if you do not have a high follower count. I haven’t incorporated context yet, but it’s pretty straightforward to make influence scores that are context-specific, and that’s coming up soon on my roadmap. I would LOVE to run a personal relay that maintains my influence score calculations for me, that would be 🔥 🔥 🔥 🔥 🔥

I had a reply to this concern elsewhere in the thread: nostr:nevent1qqsvfnf87s5gsg4unc502whwz6wjm7vmz02peghwupnhr0m34a7w4rgpr3mhxue69uh5ummnw3ezucmvda6kgtnkd9hxuete9eu8j7szyqh04fc4hw6xm4d7dd7634msqfndz9n5hyfms9u2mk6u9e3anpenzqcyqqqqqqgy8tj68

Good suggestions! 🤙 Seems still a hard work for me. 1. If you want this to work on scale you need this on a button click. 2. Somehow we have to make it sustainable. But hosting these cost money. So at the end, you have to compete with "free" solutions. Somehow the hard part for me seems that as a content creator e.g., it does not only depend on you.

Yes, at the click of a button. We've demostrated that already elsewhere - it works nicely. And yea, hosting. But, to make a long story very short: the market will handle that. Our computational model quite well suited to something like a bidding-based resource hosting market (or you could always pay a more traditional VPS service and run there, again at the click of a button. Or host it off an old laptop or pi in your closet. Or all of the above if you're a fan of lots of fail over redundancy)