Oddbean new post about login | logout I believe in the long run public relays won't exist anymore, they're too easy targets for spammers and sybils and other evil agents. So what relay models could exist? 1. closed relays with manual whitelisting; 2. paid relays; 3. invite-tree relays like wss://pyramid.fiatjaf.com; 4. wot relays; 5. pow relays like wss://labour.fiatjaf.com; 6. <add new ideas here please>
PV fiatjaf 🤙🍀
I still don't understand the idea of relays
They host all the notes and other stuff.
Relays provide the decentralization in Nostr. Instead of posting to a single server that is controlled ultimately by a single CEO, Nostr allows you to post to multiple relays (servers). So if one goes down or decides to remove your post, there are still other places for people to obtain your content. The apps eg Primal, Amethyst etc communicate with all the relays to find the content of the people you follow in order to show it to you.
hodl relays? how about relays that provide verification with bitcoin transactions? let's say there is a fresh address that only i know, i tell the relay: i will send 0.01 btc to address x. the relay will constantly check this address, as long as 0.01 btc is there i will be registered with the relay. in fact my npub and bitcoin address will be matched and i will be rewarded with a relay for holding bitcoin. so only those who have a bitcoin address with 0.01 btc in it will be able to use this relay.
This is interesting but too cumbersome, better to just pay the relay a fee and everybody is happier.
Some public relays will always exist.
Personal relay, private write and public read. Keep your note always online. Such as #cfrelay , base on cloudflare's workers. You can run your personal relay with cloudflare's free plan. https://github.com/haorendashu/cfrelay
Without public relays, followers have to know your personal relay first to follow you.
History of telecom tells a different story. By the time current relays fade, advancements will bring Relay 2.0 or better. We still have public TV channels and free newspapers in the Middle East and Africa. Sure, they'll be obsolete someday. But how long have we been saying that?
paid relay that accepts paypal ?
Actually brilliant. What's the PayPal fee like nowadays on like $3/mo?
6) Relays that unlock functions over time, for example that allow only plain notes without images or links for new users, and let user pay to upgrade immediately;
I was thinking about this one, and it seems to be a very good option that must be explored. One drawback is that it may confuse new users if their posts with links or images just fail.
Since making relevant PoW for each note is annoying and somewhat broken, another option that came to my mind was a relay that required PoW on pubkeys instead of on event ids. But since your pubkey is something you create only once this is a super inflexible scheme and annoying. But then I realized we could define a new event, say kind:13000, which would be just a replaceable PoW for a pubkey. Then we could have a relay that has difficulty adjustment for the amount of PoW they require in such event, and pubkeys can upgrade their PoW as the time passes. Or relays can require a certain amount of PoW, but from a recent event, to force users to redo their PoW every now and then, and other variants of this. This all in the context of "beginner" relays. Once you're a established Nostr accounts with followers and activity then other relays will start to open up for you more naturally without requiring PoW.
We already talked about something similar months ago, if I'm not wrong with @hodlbod. The idea was to invite people to join Nostr and do the PoW for them using a specific kind for delegated PoW as soon as they create the account, to allow them immediate access to a PoW-restricted relay. The difficulty adjustment and forcing a redo is an interesting option.
I think all of these ideas should be available, with the ability to combine them all. Delegated, replaceable PoW on pubkeys or event ids. This unlocks some interesting functionalities, like being able to set a higher bar for entering a community / flagging content, the difficulty being adjusted by moderators based on the amount of resources they have. If the community sees good quality content with low PoW, they can hive-boost the PoW to raise the content to the front of moderators' eyes. Good content could stand out this way. Higher PoW is also a nice way to kickstart an anonymous profile with no web of trust at all.
Yes fixing proof of work would be good
I predict option 6 will prevail above all else in the long term.
I don't think that will be the case. I know it's just an analogy, but there are free websites. Actually, most websites are free. They are able to stay free because they either require login (AUTH in case of relays), or the hosting provider does have anti-spam measures like chaptas and IP blacklisting.
I'll take your answer as a new item in the list: captcha-based relays. Basically all "web2" websites of today rely on some form of whitelisting or KYC, such as email verification (and requiring an email address from a provider that does its own KYC controls, not any rogue provider), mobile phone verification (same, it's KYC by proxy) or oAuth from an established provider that has done some of these. IP blacklisting is not effective to prevent spam account creation and captchas are often used together with other measures, or only in services that do not allow you to host your stuff for free on them so there is no big risk. But I think captchas could would well for relays in a less orwellian world.
If I where to pick one solution that I think will work for free relays, is the WoT approach. But not the current one (two hops), but a more sophisticated pagerank-ish style, that is provided by a third-party service provider. This is what I am building btw. The nostr version of chaptas, where instead of measuring the ability of identifying bicicles, we use the "spamminess" of the npub. If one doesn't want to provided an identifier, then he has to pay, as he's indistinguishable from a spammer.
It's like you're trying to fragment nostr on purpose. Glowie shit. We need peer to peer
like #keet ? 😁
Probably not. First red flag is when I looked it up I saw something about "end-to-end encryption" which means honeypot, like nostr direct messaging. If nostr becomes P2P under its current leadership, I presume the DM feature will remain a honeypot, but the public posting features would be better. "End to end encryption" wouldn't be one of the first search results because people would be talking more about the social media posting aspect. If nostr becomes P2P because of people like me taking over the leadership, the DM system would also stop being a honeypot and start giving users real advice about privacy and security instead of gaslighting users about "end-to-end encryption." Second red flag is it calls itself an app instead of a protocol.
“people like me taking over the leadership” really stood out in your reply 😀 Looking forward to where you lead us
That's really nice to hear. I don't function well enough without Digit to be a big part of it if she doesn't show up, but there are others like me who care about these things too. You could be one yourself.
6. Collateral relays - auth if your account has collateral that can be swiped if you violate TOS 7. Knowledge test relays - closed relay with whitelisting based on passing some proof of knowledge.
I wonder if there’s an advertising model, where somehow a relay is free with the trade off that some posts show up in your feed. Maybe a “proof of follow” that as long as you follow an aggregation advertiser npub, you have access. Wait, that doesn’t work. Every time I think of a solution to this, there’s always a trivial client side solution to block the ads. Even now. 🤷♂️ maybe that’s a good sign.
- Whitelisting via invite code - Open by default relays that ban based on negative interaction or reports - Pay as you go vs subscription relays I do think we need some better affordances for user feedback from relay (or use the ones we already have better). It's also hard to use weird relays in tandem with outbox because you don't want to have to bug the user about 30 different relay policies.
Relays with signups.
Pay as you go with ecash relays that gossip to one another, also paying with ecash
As hard as it might be--wouldn't there be a way to create toolsets to stop the spammers? I'd rather keep everything open, create good tools to help manage nefarious actors, and allow anyone to run a node easily... (Pipe dream perhaps...but if Nostr becomes centralized we risk losing the free speech aspect)
Free(mium) relays existence is nostr killer feature. They must exist to give voice to everybody. The event persistence will inevitably be the same as on the "disappering messages" feature.
Single purpose, embedded relays on hardware. Looking into that now with @Cobrador 's Tollgates. (Wifi routers that pay each other to build a network) A user gets whitelisted to talk to the embedded relay to handle the ecash payment over nostr, then a (nostr) management app publishes a session event that another router-specific (nostr) app picks up to open/close the gate for the user.
Throttling malicious traffic through cost right at the source might be helpful. Large relays can impose a cost as well, but it's difficult for them to distinguish between spam and a request that contains e-cash/has value without wasting some resources on spam. Perhaps a combination of rate limiting through PoW and hardware accelerated filters could help to reduce costs for large relays of they still face this problem.
nostr:nprofile1qqsthdwa5rs42euhnuz5xsrmmssr84hshwes7uj392vpeldj7z0zw3cppemhxue69uhkummn9ekx7mp0qy2hwumn8ghj7un9d3shjtnyv9kh2uewd9hj7qg3waehxw309ahx7um5wgh8w6twv5hsef7u3d, I think I finally understand what you have in mind. Large TollGate = Large Relay => traffic is already paid for by downstream TollGates If you are sharing a Toll Number Authority (TNA) with your counterparty, you might want to share that relay with them as well. Relays that bridge between two networks are just multi WAN TollGates 😀
have you explained the candycane-coded text somewhere?
Se fechar relay público, sou a primeira a deixar o #Nostr. Nem um pouco na pegada de correr atrás de convitinho ou pagar por relay pra postar bobagens... nostr:note1qqqqyvnz7s9e837dz8c7hdktzp2nuaxw69tq654quz9v7x5ddtcs87tvxn
ad supported relays are inevitable
6. Trainable Relays The data alone is probably valuable enough for OpenAI or any other AI system to provide free relays in exchange for AI training.
Can someone please stop fiatjaf urge to destroy his own creation? nostr:nevent1qqsqqqqzxf30gzunclx3ru0tkm93q4f7wn8dz4sd22swpzk0r2xk4ugppemhxue69uhkummn9ekx7mp0qgsrhuxx8l9ex335q7he0f09aej04zpazpl0ne2cgukyawd24mayt8grqsqqqqqpca3f02
You can do that on the current paid relays easily by just typing their public key on the relay homepage and paying. But an option to automatically invite someone to some relay without previously knowing the pubkey (because they haven't generated one yet) is a very important thing to have, @hodlbod has it as a proposal and implementation, but I never read because it was too big.
nostr:nevent1qqsqqqqzxf30gzunclx3ru0tkm93q4f7wn8dz4sd22swpzk0r2xk4ugppemhxue69uhkummn9ekx7mp0qgsrhuxx8l9ex335q7he0f09aej04zpazpl0ne2cgukyawd24mayt8grqsqqqqqpca3f02
nostr:nevent1qqsqqqqzxf30gzunclx3ru0tkm93q4f7wn8dz4sd22swpzk0r2xk4ugppemhxue69uhkummn9ekx7mp0qgsrhuxx8l9ex335q7he0f09aej04zpazpl0ne2cgukyawd24mayt8grqsqqqqqpca3f02
