Hacker News Reader

Comment by mg | Hacker News Reader

I have not yet tried Bluesky, but paradoxically, I think the "success" that currently everybody wants to try Bluesky might be an indicator that it will fail.

If there is so much demand, why does nobody offer a free and open client? This might be a sign that it is hard to write and/or run one.

I tried Nostr and it was extremely easy. Because it is so easy to write and run a client, there are plenty already.

View on HN

People have released multiple open-source alt clients (mostly in beta states, because it hasn't been very long). I've played with the API myself and it's really quite easy to use, especially with an SDK (there's an official one in typescript and community ones are sprouting up in almost every major language). I'm in a discord for people doing stuff with the API that has almost 1k members

I tried Nostr on a web client. It was pretty easy. Just going to the website was enough. I have not seen such a thing for Bluesky yet.

Is there a client online that one can simply use by visiting a website?

You could definitely build a read-only one; only account creation and the things that go with it are invite-only, the rest is already public and distributed. I don't know of a client like you describe existing yet, but I could use the API endpoints I've already worked with to build one if I wanted to take the time.

Edit: here's one https://skypulse.dvy.io/

The invite-only aspect is part of the protocol?

There's an underlying protocol (AT protocol), and then there's the Bluesky app that's built on top of it (and has its own public API, etc). Only the Bluesky layer requires authentication (and is therefore gated behind closed signups for now). My understanding is it's mainly gated because they're a tiny team experiencing explosive growth, so they have to make sure it doesn't take down the whole system (they recently had to migrate DBs because users doubled in 24 hours)

High-level explanation of the protocol: https://news.ycombinator.com/item?id=35822267

But why are there no other public webclients that let you create an identity and communicate via the AT protocol?

There's a layer of data and interaction that exists at the application level right now. I don't know exactly where the line is drawn between the two, but I think eg. likes might be at the bsky level right now (while posts themselves are at the AT Protocol level)

I believe the team plans to move as many things as possible down to the AT level over time

Because currently nobody communicates via AT. There's nobody to talk to.

Yes, staging.bsky.app was the web client all along. They finally acknowledged it’s really a prod site, so now they’re moving it to bsky.app

The vast majority of people moving to Bluesky do not care about decentralization or open clients or any of that. Like, at all.

https://twitter.com/ryxcommar/status/1651704456088518656

Which is why it might succeed: regular people want to use it, not just enthusiasts

Regular people don't want to deal with invite codes.

My experience begs to differ

I remember a time when people sold invites for GMail and the list of buyers wasn't short at all.

That's a service, not a social network which aims to beat the network effect.

Also gmail can talk with other email users.

The argument I was replying to was about people disliking invitation codes. If that's practical or not for a social network is beyond this comment.

Social media is a service.

I agree, the vast majority of people moving to Bluesky don't understand their history and now appear to be doomed to repeat it.

> If there is so much demand, why does nobody release a free and open client? This might be a sign that it is hard to write one.

It's invitation-only, with a total of ~50K users, many of whom are non-technical celebrities. That's why.

Here's a living list of clients and other projects:

https://github.com/bluesky-social/atproto-ecosystem

Reply on news.ycombinator.com