Since some people have asked me, here's a quick thread on initial experience comparison between Bluesky and the fediverse. It's important to note that this comparison is somewhat absurd by default because Bluesky is still tiny & unproven, but there's interest so I'm happy to share what I've observed. Also, I don't get into protocol differences because honestly who cares. So:

The fediverse (including Mastodon) is better at: having been proven to scale, having multiple places running different services, offering more choice of apps, giving a choice of legal jurisdictions and business/non-profit models, and is still growing far faster, especially globally.

Bluesky is better at: Onboarding & signup experience, having fewer choices to make, quality of the default app, familiarity for people used to Twitter's design, discovery of other users within the existing service.

Here are things we don't know: What Bluesky's trust, safety & abuse policies will look like at scale, compared to fediverse services. Mastodon and other fediverse services have been uneven at times, but have matured over years; Bluesky is still tiny and hasn't been tested at all. We don't know if Bluesky can scale, either technically or socially.
There are also platform features that people might legitimately have different goals for. Bluesky makes content & people searchable by default, which Mastodon only does for hashtagged content. This aids discovery, but can also enable kinds of abuse. Fediverse partisans argue this constraint helps build healthy community, but many newer users find it a frustrating omission. Both platforms let you use your own domain, but it's much harder to do so on the fediverse right now.
People's perception of the overall fediverse is (probably unfairly) shaped by Mastodon, though many other services make different choices around key features that others might find meaningful. Understanding federation is a huge barrier for people who just want to (re)connect with friends or institutions. People's perception of Bluesky is shaped by Jack Dorsey's role in its initial conception; we don't yet know whether that's fair, though he doesn't seem very involved at all.
My feeling, which I expect will encounter some resistance for a while, is that if Bluesky succeeds (for whatever definition of "success"), it's just part of the fediverse. Email has thrived for a long time with multiple protocols making it run behind the scenes; there's no reason that the new wave of social networks built on the open web can't share that trait.

@anildash My strong suspicion is that Bluesky (or any similar corporate product that preaches federation) will use federation for their initial growth spurt, and then find a way to pull up the drawbridge once private equity or the Saudis decide that lock-in sounds more profitable.

I don't know why I would be so cynical, except for my lying eyes and the entire history of Facebook, Twitter, Google, YouTube, RSS, XMPP, OpenID, and even Open Graph meta tags.

@jwz @anildash not sure why RSS is lumped with these other platforms. It still works quite well. Until elon turned it off for twitter, it was one of the best ways to get some value from twitter.
@mglo Facebook used to both provide and consume RSS until it didn't. Twitter used to provide RSS until it didn't. YouTube used to provide RSS all over the place but now only in vestigial uselessness. Is that enough for you?
@jwz that’s not a knock on RSS, its a knock on the big platforms, one of many such knocks.
@mglo ..yes? Did you even read what I said?
@jwz I must have misunderstood your intent, it seemed you were lumping RSS, OpenID,etc with the platforms.
@mglo no, jwz is complaining about the embrace, extend, extinguish approach of platforms to open standards.
@jwz that said, I agree with your original suspicion about bluesky and federation.
As the flipboard guy McCue(?) said on the vergecast, ActivityPub is like 2-way RSS.
@mglo @jwz @anildash Hi, Nitter instances still do provide RSS feeds for Twitter, for example: https://nitter.net/ppdelft/rss

@mglo

Probably because of Google Reader which was incredibly popular until Google pulled the plug. This had the effect of pushing most RSS users into using proprietary feeds instead, and the popularity of RSS never recovered.

@jwz @anildash Facebook set up shop as a walled garden, with real identities that were nonetheless β€œsafe” because it was college millennials only, and then they ripped down the walls as soon as growth and profit demanded. In a saner and smarter world that would have been the death knell for both Facebook and Zuckerberg’s career right there.
@jwz @anildash And Slack once supported XMPP and IRC gateways. Once you've got the wall up around your garden, you close all the gates that don't have a tollbooth in front of them.
@jwz @anildash how did they locked OpenGraph meta tags? I even run one service which uses them to generating thumbnails (in similar way how FB does that).
@severak A few years back Facebook/Instagram and Twitter got into a pissing match where they intentionally broke link previews of the others' service.
@jwz @anildash I have the exact same concern about Signal.
@jwz Don't forget about Slack using IRC to then kill IRC compatibility.

@onpaperwings @jwz @nicomen True. There’s also a major flaw in @anildash’s comparison of social networks with email: Email clients had to support RFC 821, 822 and 918 (and their successors) in order to operate at all. Interoperability was a given β€” a requirement β€” to even compete.

Bluesky doesn’t interoperate. They had a quick glance at the one existing open standard that exists β€” #ActivityPub β€” concluded β€œNIH”, and created their own, proprietary protocol. Their use of the word β€œopen” to describe the AT Protocol is meaningless until the spec is submitted to a standardisation organisation.

The current state of social network protocols is nothing like email. It’s even worse than the initial years of the web where browser vendors innovated furiously on top of HTTP and HTML and we had to endure the 20 years of IE winning and then dying before getting back on the standardisation and interoperability track.

Having lived through it, I can’t say that history is something I wish to repeat.

@bitbear I think @anildash had it right. Each RFC was based on β€œrough consensus and working code.” Multiple actual independent implementations. Look at the IETF site and you will see a huge number of iterations of SMTP and then ESMTP, MIME. You cited POP, but like SMTP, that was all a negotiation between completing implementations. The key was that everybody sat down at the table and worked it out. Mastodon/Bluesky/… could do the same if they make it a protocol play.
@bitbear @anildash p.s., point taken on http/html ... even protocol plays can go rapidly awry. Likewise instant messaging where even repeated efforts never got people to play nicely in the sandbox. I just like to be optimistic and have faith we might work it all out this time.
@carlmalamud @anildash I appreciate your optimism. I’m taking a more cynical approach and feel Bluesky has got a major job of convincing to do before I can place any amount of trust in them.
@bitbear @carlmalamud I think that's the right stance. I'm playing around with it and exploring because, I mean hell, I'm on all the other sites and they're even worse. But it's good to hold them to the standards of the open web, not the standards of the hyper-growth platforms run by fascists.

@anildash @bitbear @carlmalamud

Profitmaking interests will invariably attempt to herd everyone into a monopoly-controlled pen. How to prevent this?

@maria @bitbear @carlmalamud basically, multiple layers of architecture designed towards accountability. DNS did it for almost 50 years before being captured; HTML isn’t really captured after 30+ years, and email protocols only got locked down due to spam.

@anildash @maria @bitbear @carlmalamud HTTP/2 and HTTP/3 aren't quite captured, but they're approaching it rapidly.

So the file format (HTML) isn't captured, but the protocol (HTTP) is rapidly headed that way.

@codefolio @anildash @maria @bitbear I always felt we tried to shoehorn too much into port 80 (and then 443). My colleague @_mtr and I advocated a more general application layer framing solution with the BEEP protocol but it didn’t achieve liftoff. https://www.rfc-editor.org/rfc/rfc3080
RFC 3080: The Blocks Extensible Exchange Protocol Core

@maria
> Profitmaking interests will invariably attempt to herd everyone into a monopoly-controlled pen. How to prevent this?

Antitrust regulation. Or other anti-monopoly regulations, like what the EU is doing with the Digital Markets Act. But ideally in the form of an international treaty, akin to the treaties that attempt to harmonise the copyright and patent laws across jurisdictions.

@anildash @bitbear @carlmalamud

@strypey @anildash @bitbear @carlmalamud

Users should be brought in to protect our own turf. There's been insufficient attention paid, so far, to this resource (??)

Platforms for global sharing and discussion are natural monopolies; we need novel means of protecting against profiteering

@maria
> Platforms for global sharing and discussion are natural monopolies

Platforms, yes. Protocols, no. The fediverse is not a monopoly because it's based on the ActivityPub protocol. You can choose any software that speaks AP, and any host (or host your own), and still follow and interact with people on Mastodon.

@anildash @bitbear @carlmalamud

@strypey @anildash @bitbear @carlmalamud

"protocol" is not a magic word that can put an end to network effects (or stop people needing a single place to gather to share ideas and news of global events, or stop profiteers from attempting to seize control of that place)

(????)

@maria
> "protocol" is not a magic word that can put an end to network effects

It's not magic but it does exactly that. ActivityPub allows thelife.boats (your server) and mastodon.nzoss.nz (my server) and an unlimited number of other servers to form one federated universe (or fediverse'), which functions as...

> a single place to gather to share ideas and news

So the network effect doesn't trap people on any one server.

@anildash @bitbear @carlmalamud

@strypey @anildash @bitbear @carlmalamud

I'm aware, but as many have observed, a profit-making concern offering a slick enough experience can still get everyone back into the pen

@maria
> a profit-making concern offering a slick enough experience can still get everyone back into the pen

I'm uncomfortable with this framing. It seems to imply that we know better than most people what's best for them. AFAICT the main thing keeping people in one big pen is lock-in. Once there's regulatory protection (both legislation and enforcement) for people's freedom to take their social graph and data and walk, service choice is up to them.

@anildash @bitbear @carlmalamud

@maria
@anildash @bitbear @carlmalamud we can't keep corps out, and TBH we probably shouldn't if we want the Federation to be the societal norm. Our energy there should be making them abide by our rules, and the few early adopters seem to get that. Also our energy should be getting our standards bulletproof something like FB can't ride roughshod. Both are doable if we fund both the ActivityPub and [federation sw] maintainers since they're in like every small instance admin

@olavf @anildash @bitbear @carlmalamud

I'd love to figure out how to get the word out better, through our publications and elsewhere, so please get in touch with me wherever I can help!

@carlmalamud @anildash I cited SMTP, IMF and POP. β€œMultiple implementations and working code” is based on some innovation on top of an interoperable foundation, leading to standardisation. Bluesky is doing the opposite.

If Bluesky was interested in sitting down at the table to negotiate an interoperable and improved protocol, what’s stopping them from joining W3C and submitting their ideas there?

To me it seems pretty clear that any interoperability with ActivityPub is actively avoided by the Bluesky developers.

Bluesky is also not even competing at the same playing field without submitting their spec to IETF, W3C or similar. How much are you willing to wager that they ever will?

https://github.com/bluesky-social/atproto/issues/255

Already a decentralized federated protocol Β· Issue #255 Β· bluesky-social/atproto

Activitypub already exists. Why not just work on that? Why is this needed?

GitHub

@jwz @anildash My problem with this take is BlueSky is more designed to not let this happen than Mastodon is.

We're essentially talking right now on XMPP 2.0, completely designed with no lessons learned.

@jwz @anildash Hell, let's look at email, which is a very decentralized service, that a lot of vendors jumped into, and now...

Google and Microsoft largely determine who is allowed to send emails to people.

@ubergeek @jwz @anildash

That isn't my experience, probably because email was the first tool chain I dabbled into on the internet and for consumer users 'retail' brands dominate the landscape, but they don't own it.

There are plenty of viable alternatives.

@simon_lucy Anyone using email an an example of a federation success is someone who has not tried to actually operate a mail server in the last 20 years. Ask me how I know.

@jwz
No I wouldn't do anything like that. It isn't federated now and I dislike the comparison that some make to 'explain' federation of social media to it.

I don't think, even in the good old days, when running mail servers was necessary it was federated.

@simon_lucy @jwz it still is federated. We just have a couple of companies that control who can federate, is all.
@jwz I've been a mail admin (FOSS on-prem) for a large university within the last 10 years, and my own personal-friends-and-family mail server has been running continually since '08 (sporadically and unprofessionally before that in my youth). It's work for sure, but it doesn't look too much harder than, say, running an active multi-user Fediverse instance competently.
@Slyphic Oh, come on.
@jwz When's the last time you ran either dude?

@Slyphic I run my own mail server for my business which serves not only local users but also needs bulk delivery to *actually succeed* to hundreds of thousands of outside addresses, and I have been doing so for two decades, when most in my position have just given up and turned over essential business operations to Google.

So, yes, I do know what I'm talking about, and you can go fuck yourself.

@jwz
I'm not arguing there hasn't been a move to centralization of email, or that it's work to run one, but to say it's not viable nowadays while running one is not a coherent argument. You aren't a special wizard, there's tens of thousands of people as capable as either of us.

But since you want to slap epeens on the table, we pruned old inactive accounts to stay under a million active end-of-yearly. So like orders of magnitudes bigger than yours.

Ditto.

@jwz @anildash

Don't forget Slack closing their IRC bridge once they were big enough to not need to draw on the compatibility argument.

@jwz
@anildash

Right now it's at stage 0 of the enshittification cycle.

@jwz @anildash The proponents of BS often make the argument, I guess to appear "unbiased", that we should give BS the benefit of the doubt. I keep asking, why? There's a clear history we all lived through. Seems more unreasonable to wait until history repeats itself on a new service.
@jwz @anildash yep seen it happen. Conversations about how to avoid "vampire attacks" and how to "lock in value" start to pop in and* boom fuck decentralization and freedom
@jwz @anildash I agree with most of this, but... RSS?
@jwz @anildash Harder to do that if half your friends are on the other side of the moat. That part is on us though.