Oddbean

Ah I don’t think there’s much activity on the GitHub page. Everyone hangs out and stays updated in the Keet room. It’s active every day if you have questions or want to know anything. You’ve been invited to join Holesail.io as a Peer. To connect, paste this link into Keet.io, the unstoppable P2P chat app: pear://keet/yrbto5rbefzwycpcxjijbmrjfhrnidjtgt1tozgzh53chq7971nwp3d4pmiakkpksq6mjod8otqtj3jyut3ppxgg3idzcryqneekhrhhnxtucku8 See you there! Invite expires at 2024/11/5 15:16 UTC.

Yes, indeed. Pear should be a simple integration after #pubky. It almost drops out for free. Question is who will be first to innovate. Relays I think have evolved towards the email type model. I strongly believe that relays are still better than the status quo, but we are heavily reliant on damus, and the relay network is centralized. If you remember a few lawyers once went after bittorrent and sent lots of shutdown and litigation letters. Nostr would not survive such a wave. But bittorrent did survive it, and so its censorship resistance is proven and powers pubky. Once we have that then we could have pear, pretty much for free, but I think there's alot of pear skeptics. To give you an idea I integrated pubky into my software in about 2 hours this weekend, it's an easy task. The technical barrier is small, the psychological barrier is large.

it makes a lot more sense to integrate pears. Pkarr is just piggy backing on bittorrent dht mainline which is what dat did years ago and moved on to implement what keet uses today, because of many significant updates that can never be deployed on top of the bittorrent dht. hyperdht/hyperswarm is just the more modern capable version and what punky sells as brand new is pe3fectly possible on top of battlw tested hyperdht since forever already. also, pears isnt backed by by shitcoiners who grew out of ethereum/ipfs/iroh environment, which might not matter, but maybe it does?

iroh is a new york based for profit with a convertible note from IPFS or rather protocol labs and open to more funding in the future. it started as a fork of ipfs to implement another client and then adated to make more changes to it. the founders worked for a long time on the core of IPFS and now have their own take on it. There is more nuance to it and where ppl come from and how they relate to web3, but thats the gist

Everyone is so opinionated and aggressive about shit that is so unimportant sometimes 😂 I can’t imagine why it would be a problem to use either. Sure mainline is way bigger, but there’s also reasons to potentially have a new ecosystem and set of tools for certain things. But of course it’s still way smaller than the one with 25 years of history and network effect. Especially since they haven’t yet turned their attention to the tools with simple UX to replicate and distribute the DHT yet, because they’re still trying to prove out the applications to get developers interested. Resources are limited, everything is a trade off. I think we are forgetting that all of the above are VASTLY better than centralized, permissioned platforms, and that what we build on is largely a matter of preference and what exactly you want to do with it… imo.

I am not aggressive I just don't like people saying stupid stuff with such confidence, when I literally started Pkarr on Hyper DHT 🙃 If I know the limitations of a system that a supposedly fan, I am going to be annoyed by laziness :) But yes none of this matters much to be mean about it.

Agree, better than the status quo. However the devil is in the detail and there are nuances. If you want to maximize impact and do the most good for the most people, choosing where to invest your time and energy is critical. And it is a moving target. That is why scrutiny is so important. Having observed 1000+ projects over more than a decade 90% do not make it. You cant always predict which ones will, but you can see the ones that give themselves a good chance, and you normally have to get most things right. Worst thing for a dev is to invest time in a project that doesnt make it. But, especially with AI coming, we are on the brink of something big.

Yes, about right. Nostr was in fact one of those destined for the scrap heap. Even its creator had moved on to other things. You have no idea the work and team effort it took to get it to where it was noticed by jack. After that it was relatively easy. But still to level up nostr is at this point going to take enormous work, and luck.

I think betting on dat or hyper-stack and the pear runtime is a save bet. keet is state of the art and proves what is possible, which other stacks lack. The stack behind the pear runtime has also been worked towards since 2013 and survived a lot of tough times. Other projects are at the mercy of funders and if those leave ornchange their minds, people move on and things die quickly. That is different from the pear runtime which despite lack of funding for long periods of time, survived and is striving, because people give it a lot of love and effort and this is rare to find imho. It outlived many projects over the past 10+ years, projects with more hype and more funding. ...it might outlive nostr as well 🙂 And also because Keet can do a lot of things nostr cand and was never designed to do.

But I also promise you Hyperswarm won't grow ever until they start supporting upnp. I am sure more than 333 clients are using Hyperswarm, but only nodes that have clearnet open ports ever become routing nodes. That is a self inflected limitation, because when the Hole punching is so good, why would anyone open ports, and why would the DHT bother with upnp (I saw a very old issue asking about upnp and Mathias rightly said it is extra complexity not needed because hole punching is good enough). So even if millions of apps are using Hyperswarm, the DHT won't grow until this technical aspect changes. Either way I worked with the best tools available. Being only implemented in JS is another deal breaker, Pkarr wouldn't have had a single team interested in it if the only way to use it is to run a JS runtime. I am just trying to say... I did my dawn homework. and it took years to justify my decisions to myself and others. So yeah it bothers me a tiny bit when people know absolutely none of these informations but are still champions of the tech they clearly didn't wrestle with enough.

i am not sure what is being measured here or what "340" nodes means. Keet alone has plenty of more users and each of them is automatically participating in the DHT. And thats only keet. Dat ecosystem counds at least 25 more listed projects, all with their own user base, so that alone makes it way kore than 340. So hard to say how or what you measured 🙂

how can you be so wrong so confidently man? dude just learn the thing you are such a fan of, it is good to learn regardless, it is more empowering to understand, defined more empowering than to say nonsense online, people won't use your favorite tech because you tell them to, they will use it when it satisfies their need. You are more likely to satisfy their need when you learn.

You made me realize something incredible. Nostr's value proposition is that many clients can be built using the same rules. Leading to a diversity of clients doing different things. And so many different developers are funded to give different features. But actually not one client is innovating here, they are all innovating in a centralized way according to a centralized group-think. Something has gone wrong. But this is also an opportunity. Tweak the funding model to get back to innovation.