Oddbean new post about login | logout Either Nostr evolves to have every client being a relay, or it will fail. The risk is the same as that of LSPs. Think adversarially and think far into the future. Imagine Nostr running at Telegram scale with 1B users, not the tight knit community is now.
We also need sideloadable clients for all mobile, in the event app stores try to censor
Why would it fail? Government shutting down relays??
It doesn’t even have to. Relays can self sensor if there’s a threat. Just look at Phoenix pulling out of the US.
*cough* run your own relay *cough*
Yes, this should be made as easy as running your own #bitcoin node. Not everyone is running their own Bitcoin node yet, but maybe it can help each other. Integrate it optional and easy in all Bitcoin nodes like @umbrel did. The other way around would be nice as well: if there will be a simple plug and play #nostr relay option (just like @umbrel, Start 9 do for Bitcoin) adding an AppStore on that Nostr relay software and make a Bitcoin node an optional plugin of your #nostr relay would be very benificial for both worlds.
Surprised BTCPay server hasn't integrated a nostr relay option with it's docker setup yet.
They have…
Use: opt-add-nostr-relay See https://github.com/Kukks/NNostr
This is a feature of nostr. Relays can do anything and clients can easily switch
Although Nostr is designed to be decentralized, if a few relays become more popular or widely used, they can create central points of failure. This could happen due to better performance, reliability, or trustworthiness of certain relays If these popular relays are attacked, compromised, or taken offline, it could significantly impact the network's usability, even though the protocol is decentralized. So Samon's idea makes sense.
At some point Nostr will be the web.
@jb55 🥸
p2p is obviously ideal but fails in practice due to the nodes being offline problem and notes not loading. nostr is a really good store-and-forward network though with efficient set reconciliation (negentropy), so it has a lot of opportunities to take advantage of p2p. Just seems unlikely p2p will replace relays entirely due to its unreliability. Outbox model is already tricky, p2p would be techniques that extend outbox to online peers. most likely for DM purposes though, not really microblogging content where many people are consuming at once. It may seem simple to just say “just switch to p2p and everything will be better”. But i don’t think it would actually work for microblogging
Nostr currently works great for confidently blasting out an ill informed opinion about how nostr fails if you don't listen to me. Maybe be the change you want to see in the world and 1 popular client making it's users relays as well would be an improvement. If you're right, everyone that doesn't use your client fails.
Good point
Then every smartphone should be a relay because if not all bitcoiners are running a node (I’m running both Core and Knots), I don’t think every Nostr users will run a relay at home 🥲
A similar approach to what Phoenix and Breez do with Lightning?
18k daily users compared to maybe 1B one day, - even 1M - is a huge step. We better be prepared and scale. #Nostr #Bitcoin #UX https://stats.nostr.band/
I want to be able to connect to one relay (my own, self hosted) which connects to all other relays. Abit like a Bitcoin full node. No idea how that would work or if it's possible but that would make Nostr truely decentralized.
What you want to do only makes sense when you don't understand what the protocol is doing. Relays aren't single point servers. They're more like download mirrors, except you download the file from all of them simultaneously, verify that the file is the same for all, and just keep one of the copies. There is a cost to adding a relay, in that there's more copies to download and discard and more to upload, but the advantage is that a single uncensored relay can overcome millions of censored ones. And since your relay list is public info by default (including read/write status) anyone who wants to hear from you can just check the same relays you use. Maintaining a copy of all of your own notes through a self-hosted relay is a great idea, but only using that relay would cause significant problems with zero benefits. You don't even gain enhanced privacy.
I agree, I do not understand 🤣 My initial thought is we don't connect to millions of relays we connect to a very small handfull. If someone is censored from the small handful of relays I'm connected to then I assume I can't see their notes. Or am I way off?
You're correct on that. If every relay you connect to censors Bob, then you won't see anything from him. Bob and his followers have a lot of options other networks don't have to check this is happening. For one, Bob can spin up a new npub, DM a lot of his followers to tell them the censorship is happening, signed so they know for sure. Some will verify, and once the word is out new relays get spun up and old dies. Bob's regular buddies can track it too. And security focused groups can regularly audit relays, and easily sound alarms. So long as some people are vigilant against abuse, and the rest are willing to change their relay setting when the alarm is raised, bad relays are easily abandoned because none of the data they host is exclusive or under their control.
I think his is correct. Everyone is a node/relay by default. Meshtastic meshnet radio works like this. What about xmpp, bittorrent as base p2p network ? Proven for decades to be able to move tons of data. Bitrrorrent for years was more than 1/2 of all internet traffic. Still huge
Relays are already running on Android, why not have them reachable over Tor? https://i.nostr.build/fIcvFrKXSbjRkAQR.jpg https://github.com/greenart7c3/Citrine
Not every user will be equipped to build, maintain or even run their own client/relay safely. These users will require either third party clients/relays as a service, or third party software they are not equipped to manage on their own. That does not translate in a true p2p network. there is still a chance fewer such trustworthy projects will support most users. Where are they hosted? who maintains them? true p2p is not feasible. there must be something in between. on the other hand #nostr already allows for p2p networking when honest alternatives aren't available. p2p alone cannot scale.
P2P can scale. Bittorrent. Massive. Put the grunt work on the clients, let the relays/nodes just route and transfer the data.
torrent is not used by a broader audience or in any similar way to how social apps work. there might be some use cases for p2p client/relay architecture, but it won't cover all that #nostr can be used for.
Why can't torrents be used for social media ? We need intelligent clients using vast amounts of data moved around the network. The former is not the problem and torrents alread achieve the latter. Torrents also achieve storage redundancy giving resilience and state.
it's not about technological capabilities, it's about the technical skill required to maintain such a system on premise. most people do not have that skill. even torrent has evolved to use third party hosted services, think popcorn time, stramio. only by doing that they are able to reach broader audience
nonsense. with proper MLS encrypted group chats that don't make relays carry all the risk nostr could easily become larger and more secure than telegram.
This is same as saying that either Liquid evolves to have every user being a member of the federation, or it will fail. The risk is the same everywhere.
Not that worried, your social network graph is stored/cached on your phone (npubs you follow and your nsec) and can’t be destroyed. Most content on a social network should be ephemeral anyway. And you can backup your notes on your own node/relay and republish. Then there are the postbox events.
Agreed. Then there is the problem of a single private key. If that key is compromised, we are no longer in control over our identity. This is particularly problematic for companies that need to share their private key with several people. If Nostr were to have two private keys, one of them a master key with admin capabilities for the identity, the protocol would need a new architecture. As a result, whenever a Nostr spin-off architecture solves this problem, every company, corporation and security-minded user will migrate to the protocol with higher security. That's just how things work.
NIP-26 Delegated Event Signing https://github.com/nostr-protocol/nips/blob/master/26.md
That's a great NIP solution. The main issue I see is in cases where the root key is already compromised. It also requires a fully secure handling of the root key. If we implement NIP-26 and then create new root keys from scratch, we could safely move from our old Nostr identities to new ones.
Root key management is hard, no easy way around it. This is one of many solutions: https://primal.net/e/note1xw04kjsq2cnay0xua7l5cnqgu9qkfnun4npmw54tyy68v8kr2wuq4x6jq8
Decent idea. I was looking at this yesterday. It is possible to run an idle relay in 20MB of memory. Hope it happens, although the relay network is currently shrinking and centralizing around a few reliable ones nostr:note1qatxpzv7kaegvyyzr3l0y0q2s94ss5df72lfwtudhcj98xugrgsqeywe2z
You launched a closed-source bitcoin wallet in 2024. Just shut up and go back to Twitter.
Aqua's open-source: https://github.com/AquaWallet?ref=jan3.com
While individual relays within the #nostr network may experience failures, it's important to recognize that such occurrences do not compromise or invalidate the #nostr protocol itself. #nostr is fundamentally just a protocol, a set of rules and guidelines that dictate how data is transmitted, and it remains resilient regardless of the operational status of any specific relay. For those concerned that the failure of relays could lead to a catastrophic collapse of the network, rest assured that this is not a realistic outcome. The protocol is designed to endure and function despite issues at the relay level. That said, it's worth acknowledging that the growth of the #nostr ecosystem might not meet everyone's expectations. This is a separate concern, one that pertains to the broader adoption and expansion of the network, rather than the performance or reliability of the relays themselves. Even if the ecosystem's expansion is slower or less robust than anticipated, this does not inherently reflect a failure of the relays or the protocol itself. nostr:nevent1qqsqw4nq3x0twu5xzzppclhj8s9gz6cg2x5l905h97xmufznnwyp5gqppemhxue69uhkummn9ekx7mp0qgsvnvvlln2ru6jlyweayugxecv7ftfdlzd6zqca63sh7x6ercggjegrqsqqqqqp5jxhgh
Hmmmm... nostr:nevent1qqsqw4nq3x0twu5xzzppclhj8s9gz6cg2x5l905h97xmufznnwyp5gqpz4mhxue69uhhyetvv9ujuerpd46hxtnfduhsygxfkx0le4p7df0j8v7jwyrvux0y45kl3xapqvwagctlrdv3uyyfv5psgqqqqqqsu4ccgj
EVEN IF NOSTR CENTRALIZES ON A SINGLE RELAY, AND NO IMPROVEMENTS ARE MADE, NOSTR WOULD STILL BE A SIGNIFICANT IMPROVEMENT OVER THE EXISTING STATUS QUO. WE WILL DO MUCH BETTER THAN THAT THOUGH.
You should read about Reticulum Propagation nodes. This conjecture is not only inaccurate, it's as short sighted as the big blockers in 2017. Having a client be a public Relay is just not feasible on the hardware most clients operate on. and if every client is just a private relay no one is being served your content. Get a Mini PC, have a few TBs of memory, create a public relay and purge it on a schedule. Hosting your own server solutions will be a very important part of the future.
Keet
