RE: https://boing.world/@pre/116206661645246636

The darling CEO steps down, and a scapegoat CEO steps up. They turn the heat up to 11 and take all the blame.

A new CEO eventually takes reins and promises to "fix things." They make a few popular changes. People praise the new ceo for walking things back to a 10, where the board wanted things all along.

Let's be clear what's at stake here, because this isn't just about a social media platform.

Bluesky PBC wants ATProtocol to be the *infrastructure* for the entire Social WΜ·eΜ·bΜ· Internet. That level of ambition deserves scrutiny from everyone from every angle.

This isn't an ActivityPub vs ATProtocol thing either. Frankly, imo, they both can go under or survive.

Will these projects improve the social web for the *people* is the only question that matters.

Objectively, too much power still rests in the Public Good Company's hands. It's like I'm watching troops amass the border while the General is on the radio claiming "this isn't an invasion." That's the vibes I get from Bluesky, currently.

And I'd rather be embarrassed and wrong, than politely watch another Meta superpower take hold of the web.

That said, I'm suspicious of BS while still optimistic about projects like Blacksky. But they're up against a lot...

Rudy, the founder of Blacksky was cool enough to share expenses.

What immediately sticks out to me is the AppView expense. Blacksky must index the entirety of Bluesky's posts.

That means ATProtocol is designed to favor the largest node. The small nodes must cary an uncontrollable, unpredictable expense. That's a pretty unattractive aspect of this decentralized ecosystem.

You could be breaking even one day, then overnight in the red after a big migration that you may or may not benefit from.

Bluesky is as-tech. They may lead with subscriptions first. But the ads are coming.

What does an ad network with a public firehose of user data look like?

No really, I'm asking. What does that look like because we've never seen that before.

For one, I think, the concept of privacy is completely dead. The Cambridge Analytica-style of psychological targeting is now democratized. So that's fun. We don't make a big enough deal about that firehose.

Anyway, I just made chocolate banana bread.

You see, social media is like chocolate banana bread. It-- lol jk.

No but for real. I'm BS skeptical to put it mildly. And I really hope I get to be embarrassed about that one day.

@fromjason
The last I checked, my own small-to- medium home instance cost ~100/mo
@ohmu on here right? Yea that's nothing compared to running something on ATP πŸ˜…

I think of ActivityPub and ATProto as optimized for different scales and full-network searchability. For something small to medium (up to hundreds of users, maybe low thousands), ActivityPub is in general a lot cheaper than an ATProto AppView (which is what gives you the full-network searchability) -- especially if you're using something like GoToSocial or snac. There's a bit of an apples to oranges comparison, though, because ActivityPub doesn't get you full-network searchability.

[If you're willing to give up on full-network searchability -- or rely on an external service like the microscosm -- you can have a partial appview, which makes the cost structure more ActivityPub-like, and I see a lot of that happening with formats other than micro-blogging. But that's not what Blacksky's trying to do.

On the other hand once you get up to tens of thousands of users, ActivityPub gets expensive too. And if you're actually treating moderation time as a cost (as opposed to viewing it as unpaid labor, which most AcitivityPub instances do), it's likely to outweigh the storage/compute/networking costs. Blacksky's invested a lot in moderation tech and processes, which gives them a big advantage over most larger fedi instances., Then again most larger fedi instances don't moderate well so it's another apples-to-oranges comparison.

@fromjason @ohmu

If I were building a large mostly-public network I would very probably use ATProto. It's also a much better developemnt environment.

On the other hand if I were doing something involving networked communities of tens or hundreds (maybe low thousands) of people, I'd go with ActivityPub for now. But Rudy's business model for Blacksky includes hosting other communities, and especially once they have local-only posts that could be intriguing.

@fromjason @ohmu

@jdp23 @fromjason
This feels like a stupid thing to ask but is not being able to edit posts baked hard into ATProto?

In principle, no; posts are explicitly not immutable.

In practice, though, maybe it is to at least some extent. https://pdsls.dev/ allows the abiilty to edit posts, and I sometimes make use of it to fix typos. But if I just edit (as opposed to recreate) the post, the results aren't visible through the app. My guess (and this is totally a guess) is that there isn't a firehose event for the edit -- or maybe there is but the event isn't getting propagated. So is this something that Bluesky could fix if and when they decide to implement post editing? No idea!

(And no it's not a stupid thing to ask!)

@ohmu @fromjason

PDSls

Browse the public data on atproto

@fromjason I dunno I’m on a social media break
@fromjason enshittification is the destination, the indispensable end-state to milk all possible profit.
You can’t stop enshittification from within: Step in front of it and be trampled.
Venture Capital rules: greed is God.
@wtrmt agreed. It baffles me when people defend VC funded projects with "let's wait and see" when every example we have proves there's only one outcome.

@fromjason
Bluesky users: "It'll be fine, right? It'll be fine... I think it will be fine. Won't it? Won't it? It'll be fine."

🀦

@fromjason this also reminds me of the 'glass floor' phenomenon where a new leader who would usually be faced by a glass ceiling gets promoted except they're given an impossible job they're described to fail.
@fromjason Yep. *SHE* won't... some other CEO will though. Slimy...