I just sent the final proofs of the #ActivityPub book to the editors at O'Reilly.

Evan Prodromou

@everton137 Absolutely. You saw that we have had 2M new users since August on the Fediverse, right?

FediDB, Fediverse Network Statistics

FediDB is a cutting-edge service providing detailed statistics and insights into the Fediverse network.

(fedidb.org)

We're seeing a lot of growth here.

bryan newbold

@evan @everton137 bummed that you have such a negative public take on atproto and Bluesky and see it as antagonistic and "dangerous".

I have a huge respect for ActivityPub and the folks involved, including you. I think we are all working towards the same goals, there is no final word on protocols/architecture/governance/regulation in this space, and that there is a lot of room for collaboration.

bryan newbold

@evan @everton137 even if folks don't trust Jay and the Bluesky board (who are the relevant individuals and stakeholders-holders for the future of the org), we are working in the open and everybody can read and learn from our experiments, successes, and failures. We totally endorse inter-operation efforts like bridgy fed, and collaboration on overlapping needs like identity systems, OAuth, and E2EE.

Evan Prodromou

@bnewbold @everton137

Hey, Bryan. I like that BlueSky, the social network, is an active part of the Fediverse through BridgyFed.

Jérôme

@evan @everton137 our active users are going down however.

Evan Prodromou

@bnewbold @everton137 I moderated some of the language in this thread.

Steve Bate

@everton137 I recommend doing your own research. Some of the points being made about BSky don't appear accurate to me. I see a lot of community involvement in developing alternative UIs and other components of the stack (like tagging). There has been talk about having parts of ATProto standardized at the IETF, but they're waiting until the architecture is mature and proven (unlike AP). As a "losing" protocol, they gained 4MM new users just in the last month. 2x the MAU of the Fediverse.

3️⃣ days without TF incident

@evan @everton137 thanks for sharing. There's a rumour about that BS development of ATproto was to create a way to essentially divest business of moderation safety etc and enable a more "free speech" social media. Do you think that has any weight?

Evan Prodromou

@arichtman @everton137 no.

3️⃣ days without TF incident

@evan @everton137 ok, thanks

Evan Prodromou

@arichtman @everton137 That was a little brief.

The point of ATProto is to be different and incompatible. That's the business model.

Any benefits that come out of that primary goal are secondary. Yes, their moderation system is more distributed. I think there are precedents of doing client-side moderation, like email spam filters, that we could use for the Fediverse.

But the features aren't the point. The point is a protocol that the company owns and controls.

Matt Keeter

@evan I'm surprised to see this level of hostility from you!

I really enjoyed your "Big Fedi, Small Fedi" post, and from reading their architecture docs, Bluesky is genuinely designed to address some of the "big fedi" challenges that you pointed out.

If you believe that Bluesky is inherently bad because it completes with ActivityPub, that's fine, but the level of FUD in this thread is disappointing.

Evan Prodromou

@mjk I care a lot about the social web. I do not want to see its growth stunted by competition at the protocol level at this crucial time.

If you found something untrue in what I said above, please let me know. I'm happy to correct it.

damon

@everton137 you can take a look at these https://atproto.com/guides/faq#why-not-use-activitypub https://fed.brid.gy/docs#compare

Evan Prodromou

@mjk Here are a couple of links to give an idea of how I feel about BS and other proprietary social networking protocols.

Evan Prodromou (@[email protected])

'Prodromou, however, has strong words for any organization looking to enter social media with a new decentralized social media protocol. “I’m not interested in any protocol besides ActivityPub,” he says. “Anyone working on brand new protocols in 2023 should stop immediately. They are going to do more harm than good."' This may be surprising for people who have known me for a long time. I've generally been supportive of trying new protocols and tools.

Prodromou.pub (prodromou.pub)

http://evanp.me/2023/10/06/activitypub-the-socialcg-and-the-social-web/

Steve Bate

@jerome @evan @everton137 Yeah, those Fedidb numbers are difficult to reconcile... it shows 1.5MM new users are since July, but only 100K increase in MAU? You'd expect at least the newly joined users to be included in MAU. ‍️

julian

@[email protected] not necessarily... setting aside the fact that the numbers themselves might not be right (because numbers are hard, people!!), new user counts without a corresponding MAU increase can be explained by large new instances joining the fediverse.

e.g. when Flipboard opened up their entire network to the fediverse, that would show up as a huge number of newly added users, but a great majority of them wouldn't be active due to churn.

Steve Bate

@julian Given the "monthly" granularity, I'd *expect* that a user joining would count as activity in that month period even if there's churn in the next month. The August data is showing about 400K new users and about 40K active users (90% of new users are not considered active in the month they join?). Seems strange to me, but I agree nobody knows if these numbers are accurate or meaningful. I see reason for skepticism.

julian

@[email protected] of course! That's perfectly reasonable.

If FediDB captured my forum's nodeinfo for the first time, they'd see this:

usage:
    users:
        total: 25698
        activeMonth: 197
        activeHalfyear: 952

Depending on how you consider the "joined" action to be counted, FediDB could consider this as 25.6k joined users, but only an MAU increase of ~200.

Steve Bate

@julian Yeah, that's seems wrong if I'm understanding you. Given "joined" is not explicitly reported, it's probably taking the previous month's total and the current month and computing the different in total users (where a positive number is considered "joined"). They should wait at least one month to report those numbers to avoid artificially inflating the stats.