You run a service that serves public text updates. Why am I reading a blog post and not a series of bluesky posts?

I know that different mediums serve different purposes, but it’s tough to see why this wouldn’t be bluesky-first content.

Please don't give such terrible advice.

Twitter is not the place to write/share much of anything. The other two, why aren't they for long form posts?

That doesn't make it worse or anything. There's a lot of arguments supporting a text limit per post. But there's no reason why you should be posting this kind of long form content in a thread when you have a website you can link to (which is what their bsky acct did).

Given that, and given that no one needs exposure to real bluesky posts more than bluesky right now, why make it longform at all?

There's maybe a political post here or there but I don't really see discourse or anything on bsky. And either way you should be able to add a labeller to your account that filters out discourse and rage-bait.

People are over Web3 protocols. We have torrents, email, git, and other decentralized technology that was invented to solve specific problems - this idea of inventing a social media protocol is cool, but it comes with all this baggage. It's a for-profit company - wait it's not a for-profit company - it's decentralized - wait it's not decentralized.

This dance never happened with a Napster or a Popcorn Time - it was very clear what the decentralized utility was, and it was really a means to an end of getting to the content. But Bluesky doesn't have any content I want, especially not badly enough to jump through those hoops, or use like a B-version of the real thing. It's like building a new restaurant to compete with the best steakhouse in town but you focus on the trays instead of the food.

"You can change providers!" "You mean sign up on another website? Like when I moved here from Twitter?"

We got up to around 15 engineers since then. We decided it wouldn’t increase our throughput enough to hire aggressively (org building alone is a major distraction) and decided it wiser to preserve our burn rate, which we did.

Inevitably we get compared to Threads, and that’s just how the market works, but they had 60 people on the team when they launched, and they did so by running off of Instagram’s infra and T&S. Watching them handle 100 million signups in their first week? That’s the rubric that the market is grading on today. So, that’s what that moment looked like for us. And again, I think we captured it well. We managed to build a really strong initial userbase with it. So the challenge for us now is to create new moments, which is what the product work is about.

Nice to see you're still patting yourself on the back lol I think you took wayyy too long to release your product and it's overall not an impressive looking social media app, certainly doesn't look like the work of

> 15 engineers

I do think Bluesky may have chose the wrong launch strategy by using the old and tired invite-only method. But hard to say for sure. If they had fully opened to the public last year or the year before, they might have gotten a large influx of users, who would then complain that it doesn't have all the bells and whistles of Twitter, which had 10+ years to develop them. So who knows. At least with invites, they made sure to get people who really wanted to be there, even if it was a smaller crowd.

Hard to say indeed. I'm not sure how a social network bootstrap method can get "old and tired" either - humans don't change how they socialise that quickly!

Facebook achieved remarkable things by starting in the Ivy League community then branching out, it really stands as a fascinating example of how curating the early community is critical in the bootstrapping process for a social network. I'm happy to bet that there is a 0.01 to 1% of the population who actually matters for getting groups of people to use something, and if a social network manages to find them then it will succeed, invites or not.

It could be that there is too much dross out there for free for all signups early on. To me it looks like a core community should be established first.

But they do get wise to things like false scarcity of digital goods or artificial gatekeeping, and it puts them off. When a strategy is no longer novel or original, its loses its effect.

Yes that early core community can really set the tone for a long time, and may have been worth it. On the other hand, an even bigger risk to a social network is not getting enough community to cause positive feedback effects and organic hockey-stick user growth. Failure to launch. Especially with a glut of social media competitors crowding the landscape.

Product took forever and when it finally came out it wasn't impressive. They should have staffed up some badass devs & designers who ship or even considered a top tier agency who would have made something truly stunning. Test the market at each step. Decentralization deserved better!

I think a good comparison is OpenAI (including the whole "are they open vs closed" part that Bluesky had to deal with). OpenAI has done an excellent job of: 1) Releasing good products on a good cadence 2) Keeping us waiting in suspense for the next big thing.

Decentralization has a very high excitement potential - I think Bluesky just dropped the ball. Didn't prioritize product.

Elon buying Twitter was the factor that changed our plans and led us to build an app, as it became clear a path to adoption within Twitter would never pan out. We were not set up as a product team given the legacy of our origins, but did the best we could last spring, shipping an app to hundreds of thousands of users with just one app dev, pfraze. OpenAI has been around for nine years, and while what they've done is truly impressive, launching projects with a fully-staffed engineering team is very different from building a team as you grow a userbase.

Hindsight is 20/20 when it comes to market timing, but our goal from the start has been to deliver a decentralized social experience with good UX, and deliver on the previously theoretical concepts of algorithmic choice and community moderation in a global public square. Because of this, our product focus has been driven by showing what's possible on an open protocol, and only after doing things like custom feeds and third-party labelers have we circled back on some more basic features users have been requesting like DMs and video. We still believe decentralization, and the innovation it enables for social, has high potential, and are always looking at how we can better surface it in the app.

What shenanigans?

When he fired a whole load of staff there were cries of Twitter’s demise. The opposite has happened: many new features added (more than the previous 10+ years?) and site stability is solid. So how is the competition “wide open”. I use Mastodon (since 2017) but I’m not kidding myself about the small fringe userbase (I like it this way BTW).

And Dorsey has just binned bsky and championed Twitter.

A small contingent detest Musk so much that it obscures observable reality.

The site will continue to be one of the most popular in the world largely because people don't want to "redo Twitter" on a new website. It really doesn't matter what Musk does.

I think one of your biggest competitive advantages is just a sane decent core team, who attract similar people from their network. A sane and decent core community is a refreshing change from Twitter's increasingly crazy and distasteful one, and a federated protocol is inherently more trustworthy than centralized ones like Twitter or Threads. I hope the nexus of those two things helps you take off in 2024. Good luck!

Calling X and NoStr "freedom tech". Whatever that means.

It it were still relevant, they would have quoted DAU.

Around 200,000 folks liked a post today.

Even as they talk now about addressing the stuff that's missing from Bluesky like DMs and videos (which, no wonder it hasn't taken off when those features are absent), it still sounds like even when those features launch, they'll still be subpar. DMs not part of the AT protocol? Did they just not think of DMs when they designed the protocol? And only 90 second videos? Really? In my opinion, Bluesky isn't going to take off if this is the roadmap.

Since Twitter/X stopped making some posts public and make it so hard to view profiles you'd think Bluesky would be all over this.

We had a small earthquake not too long ago in SF and I had no Twitter to check - that instant live Internet community is just gone, and it's not just that. Everything used to come out so fast on Twitter, it was a really exciting place to be. It's of course totally ruined now and has been for years, but it was once an awesome thing.

Not sure why Bluesky isn't all over Hollywood and politics, in a presidential election year no less. They're just not doing this very well.

Ugh, that would turn it into just another online cesspool.

The things bluesky wants to do well really just don't map well onto ActivityPub.

Things like:

- Custom feed providers - Custom indexers/firehoses - Labellers (i.e. content filtering/flagging services) - Proper, cross-instance moderation services - Separating identity from the instance

> DMs not part of the AT protocol?

Currently the AT protocol is basically all public. If you want DMs you need to work out an E2EE layer that can still tolerate moderation and a propagation system that doesn't leak details of who is messaging who. That's all very hard to do right. Even signal spent years trying to get this right and they have pretty centralized infrastructure. I'd rather no DMs at first with the expectation that they'll be done properly when they actually ship.

> And only 90 second videos? Really?

Mastodon sets a video limit of 99MB. If you wanted to match the quality of a 1080p 30fps youtube video for example, that 99MB would buy you about 82 seconds. So the 90 second limit really isn't that unreasonable. And if you want a longer form, external video embeds have been supported in the app since the beta.

mstdn's weakness is the activitypub protocol, it has many limitations, but understandably because it's designed in good o' GNU Social days. On the other hand nostr had like gazillion of NIPs

ATProto fits the right spot, designed by the people who knows what they are doing, and addresses all the problems.

user bases will remain separate as not everyone reads every network. but a new app on its own isnt that compelling. even ones that specialize in a niche like be real will have trouble finding large scale relevancy

I've seen some Nostr users bridged as well