I'm that weird person who when making a test post in a DM to myself, I still enter the alt-text on the image... Something to do with not trusting that I selected the correct switch to make it a DM...
I should now be running the media proxy via a separate domain, which is advisable anyway. The local media were already done via a separate domain years ago so they can be served by cdn.
Weirdly the setting for using a separate domain for the remote media proxy applies immediately, to old posts and new. While the setting for local media applies to new posts only; old posts keep having their old url in their html content blob. Further weirdly, sometimes old posts with local media get captured by the remote proxy as well.
Preview media, whether remote or local, always goes via the remote media proxy, bypassing your cdn.
I was first thinking about writing about a blogpost about how to make these settings, but I’m still too confused myself.
I still need to figure out how to do similar things on Friendica. On Pleroma, PeerTube and WordPress I can host the site myself and offload local media serving to CDN, but on Friendica there is no such easy setting...
Somewhere between 2 and 5 March 2021 a change happened that posts before that date don’t have their local media captured by the remote media proxy, while posts made after that date do, until between Oct 2021 and Jan 2022 it started to use the proper CDN url. All of them are on S3, so that isn’t the difference. What I don’t like is that some of these urls seem hardcoded in the stored html content of the posts; same gripe I have with WordPress.
Disabling link previews on Pleroma after the thing the other day, made me realise how little I miss it. When doing the same on Friendica just to check, I notice that here it's a user setting, not an admin setting. The link preview, that is; the oEmbed setting is in the admin interface and uses an allow list.
ChatGPT text can sound very knowledgeable until the topic is something you know well. Like tic-tac-toe.
Once I heard that ChatGPT can play tic-tac-toe I played several games against it and it confidently lost every single one.
twitch streamer Doug-doug recently had ChatGPT play Pajama Sam (a point and click for children) by verbally explaining the situation to it, and it was able to pass the tic tac toe minigame without too much failure — which clearly has an intentionally flawed opponent algorithm because it’s for children lol
I'm getting way too good at knowing how long to put a cold cup of tea in the microwave to make it properly hot again. Sign of how often I'm forgetting to drink it.
Upgraded to the new Friendica 2023.05 that just came out, since I wasn't expecting it. And releases that come out when you're not expecting them are best not missed. #Friendica
diet tip: before you eat anything, think about being an alien's human pet, and think about what the alien veterinarian would say about the alien feeding it to you
This series has characters named 'Raven' and 'Octavia of the Sky People', so you'd think it would really be my thing. But I'm into season 3 of The 100 now (it will leave Netflix at the end of the week, so I'm not going to make it through all seven seasons) and it's so-so. The Good: it has Alycia Debnam-Carey in it and some strong female characters. And a properly created language. The Bad: it's all very militaristic and the very talented Genevieve Buechner is completely underused. But what gets me the most is the way main characters can commit massacres and that the story steps over this rather easily.
"Rather than moderating content, Liam was asked to adopt fake online personas—known as “virtuals”—in order to chat to customers, most of them men looking for relationships or casual sex. Using detailed profiles of customers and well-crafted virtuals, Liam was expected to lure people into paying, message by message, for conversations with fictional characters"
Hired as customer service reps, these freelancers were instead tasked with luring in the lonely and lovestruck through a network of dating and hookup sites.
So, a lot of dating sites are actually secretly just dating simulators packed with microtransactions? 😬
This line was particularly disturbing: "She spoke to elderly users in care homes, and others under protective conservatorships who asked virtuals to wait until they were given their next allowance."
Horrible to think how they are preying on lonely people, especially vulnerable ones.
Presumably these "dating" sites and apps will start using AI instead of paying people even meagre wages. That's bad enough.
What happens if putative social networks start using AI to fill them with fake people? How easy will it be to tell when we are actually engaging with human beings online? Will it become impossible, in the low-context short length world of social media?
Are people going to be put off using computers by computers pretending to be people?
Every once in a while I mention people who have promoted themselves for the purposes of finding love. However when trying to find an example I can never quite find one. Its something I tried to do …
@FediThing Sadly online dating is the dominate way people meet in the west. On top of this is a huge up tick in loneliness, personally being exasperated by this type of thing.
The Japanese government was paying bars to run single nights because of the impact on the society,
I think that says it all right? Another reason why I keep thinking about new types of dating like @evan talked about at Mozfest in 2017
@FediThing COVID and digital stimulation? Judging by my experience and books like sherry turkle's alone together I hadn't even thought about the potential for long term manipulation and blackmail 😥 Wow! ☹️ @evan
website of a swedish furniture company: you can easily repair small scratches by painting them! me: ok, so what's the colour code of the paint I should use? website: we do not give out our colour codes since they are a trade secret.
Odd, why would Paramount be using a Let's Encrypt certificate and why does my phone say it's not a recognised Let's Encrypt certificate? While it perfectly accepts others...
I remember the time when these birds were so scared of humans, it was rare to see one and then it was from far away. Within my lifetime they evolved to not mind human beings at all.
Fucking Christ the @protocol is the most obtuse crock of shit I've ever looked at. It is complex solely for the sake of being complex and still suffers from *all* of the same problems as Mastodon.
Your server goes down? Sorry, all of your followers are lost. Account portability is no better than Mastodon. 'DIDs' serve literally no purpose. And none of the API code that Bluesky uses in their own app validates ANY of the crypto they're doing on the server. NONE OF IT.
Not only that, but instead of just... making a simple Rest API spec or something simple, they REINVENTED THE FUCKING WHEEL and made 'XRPC' and 'Lexicon', AKA a shittier, less flexible version of OpenAPI and JSON Schema (respectively) that works with absolutely NO existing tooling.
The actual protocol itself is poorly confused and incredibly awfully designed. Because of the useless bullshit crypto they're putting into it, it requires you to write a server that's strongly consistent with other servers. THAT IS EXTRAORDINARILY HARD TO DO!! And BLUESKY'S OWN SERVERS DON'T HANDLE THE EDGE CASES!!!
It uses pull-based federation instead of push-based like Mastodon. You have to write a separate 'indexer' that has your 'feed' on it. That requires a LOT more resources.
And because things must be strongly consistent, AND because any user can REWRITE THEIR OWN FUCKING HISTORY AT ANY TIME, you have to, as an indexer, account for lots of different edge cases where your recorded history diverges.
The protocol for federation is built to make federation as difficult and painful as possible. It is built so that Bluesky, the private company that makes the protocol, is the only 'indexer', the only one with a whole view of the network.
WHY DOES A FEDERATION PROTOCOL NEED USER AUTH API METHODS IN IT???? WHY DOES A FEDERATION PROTOCOL NEED A CONCEPT OF 'INVITE CODES'??????
Oh wait, BECAUSE IT ISN'T A FEDERATION PROTOCOL!!! It's literally JUST the API for Bluesky. That's it.
It's quite literally impossible to use this in more flexible environments. You just can't. You cannot build anything on this because it is so poorly designed and isn't generic enough.
And I went into this with an open mind. I was like "I'll just make a simple alternative to the BlueSky server in Elixir". But it CAN'T be a simple implementation like ActivityPub can be, because it is extraordinarily complex and requires you to make guarantees about your storage and how your application works.
It turns out using Git, which is almost always used with a centralized 'remote', to do federation, which needs to be weakly consistent, IS A BAD IDEA!!!!!
What this comes back to is... who cares about any of the crypto bullshit? Having a private key and signing everything with it proves nothing because that private key must have a reputation.
You can verify a domain on Mastodon. You can point a domain to a Mastodon server. You can do that with Pleroma. You can make your own alternative to Mastodon that works exactly like how Bluesky works with domains, but it would take a 4th of the time because ActivityPub is simple to implement!!!!
The ONE thing that this protocol brings to the table is the idea of strong consistency in federation. The only issue is, it makes that strong consistency so resource intensive and so hard to implement that it decreases community servers' ability and ease of federation!!!!
And also, NOBODY CARES ABOUT STRONG CONSISTENCY IN SOCIAL NETWORKS!!! Social networks are built on the idea that we all have a different view of things. We care about seeing stuff from our friends, not seeing EVERYTHING.
The 'account portability' piece is bullshit! The way 'account portability' works is by having two separate keys, one for signing and one as a 'recovery' key. You're supposed to be able to use the 'recovery' key to rewrite history if your account gets hacked or some shit.
WE HAVE THE ABILITY TO DO THAT AS SERVER ADMINS!!! MASTODON HAS THIS ALREADY!!
Additionally, if a Bluesky server goes down, their way of keeping access to your data is by STORING ALL OF IT ON YOUR DEVICE!!!!
Imagine if I had to store the 50k+ tweets I've made on Twitter on my device, and upload ALL of them to a new server whenever a community server went down. Imagine being a server admin having to deal with people uploading tons of JSON data and media a whole bunch at a time. And if you're implementing the protocol correctly, EACH JSON BLOB REQUIRES VERIFYING THE SIGNATURE!! So you'd have to do 50K SIGNATURE VERIFICATIONS! WHICH IS CPU INTENSIVE!!! AND SLOWS DOWN THE SERVER!!!
Additionally, the domain name is NOT your identifier. If you have a custom domain, that is NOT your identifier. Instead you have a 'DID:PLC', which is a kind of 'DID' (invented by, not a surprise, CRYPTO PEOPLE).
There is NOTHING FUNDAMENTALLY USEFUL ABOUT THIS IN FEDERATION. Because this DID is never made visible to a user, it is not human readable (it's a hash), and it doesn't do anything!!!
Am I understanding it right, then that they are building an individualist kind of portability as compared to Mastodon's more collectivist kind? They want to make each user a kind of sovereign citizen who owns their own profile, while Mastodon is predicated on you being part of something, a community on a server?
No, your data is stored in a server that can be a community or an individual person. But that doesn't store your feeds, like Mastodon does. So like your feed on here exists on your server, and it's built by people sending messages to and from your server.
On Bluesky, there is a single, more 'global' feed called an indexer. That builds your feed.
It is just extraordinarily complicated and not immediately obvious how to build one. ActivityPub allows you to build a 'feed' however you want in your app, but Bluesky requires a certain structure and way of pulling information in. The way you get all the posts is extraordinarily resource intensive and complicated to write correctly (because it needs to be consistent), which will make it so that there are few community-based alternatives.
it's a little more complicated than that-- your data doesn't live on "servers" bound to instance-like clusters of moderation policies and federation decisions, it lives on personal data stores (closer to the Solid model) which are commodity interoperable dumb lockers where bits go, but because they're self-verifying and signed, they can be consumed en-masse by multiple servers over time. your data never moved, access rights to that data just gets authZd and de-authZd
I had the opposite reaction. There is a huge potential in the long term benefit of having a large infrastructure of globally unique identifiers mapped to portable human names. This opens up the potential for strong public key verification and real authenticity to e2e communication.
DID potentially allows for better portability since you only are changing the human name, despite the problem of your archive. Yes, it is overkill for a social network, but it allows the AT protocol to be used for so much more. Imagine it combined with WhatsApp's new key transparency.
Yes, the AT protocol completely dodges the recovery problem, but so what? The recovery problem is hard and unsolved.
That might be true if they’d actually implemented the DID feature they said was absolutely essential, required for truly nomadic user accounts and the reason why they couldn’t just adopt ActivityPub. They never got proper distributed DIDs to work and instead have a “placeholder” implementation. Pretty sure it’s a central directory operated by Bluesky. I presume they issue new ones to people with invite codes?
This is correct, DID:PLC is basically DNS for DIDs. Then they layer DNS (domain names) on top of that. That is now two ways of doing effectively the same thing.
This is a pattern with the entire thing. Instead of using an OpenAPI spec, they invent XRPC (objectively useless). Instead of using JSON schemas, they invent Lexicon.
Each layer does barely anything and does not exist for a good reason. The sole function is increasing confusion and making it harder to implement.
It looks like it was really important for everyone’s data to be in an append only public ledger so it could be easily scraped/indexed/searched, but nobody seems to be complaining about the privacy implications beyond vaguely wondering when they’ll get DMs (never) and realising their blocklist is public.
It's not exactly like this. An 'indexer' is basically just like Google. The issue is that sometimes entire trees of conversation might disappear at random because it's literally a cryptographic chain of links that can break at any time. It's incredibly inconvenient and mastodon's / ActivityPub's way of dealing with that is much easier. If you delete a post, all servers are notified of the delete. That's it.
I don’t have an invite, so I haven’t tried playing with any of the sample code, just reading the docs. The PDS merkle tree is just repurposed blockchain stuff, right? So it’s easy to add records, but not remove them, or are you saying anyone can just delete the tree any time and start again? Or just with a recovery key? It also looks like a hypothetical AT/ActivityPub gateway would have to construct a PDS for every AP user passing through.
It's not actually, it's sorta copied from IPFS. Just about the only thing IPFS is good for is for making it easier to pirate stuff, which is completely good by me. They tried to make it into something more "Web3" but it doesn't work that well. The largest use of it right now is LibraryGenesis, which makes it easy and fast to download any book you want. You can also contribute storage to the pool (Anna's Archive is looking for volunteers).
Ah, ok. Still looks like they’re prioritising syncing big blobs of data around, rather than sending individual messages. Maybe their secret plans for monetising this rely on it. All feels backwards - a very specific data store and commands to manipulate it, rather than a message passing protocol and letting implementations store that state internally however you like.
Granted, yes, the XRPC and Lexicon stuff is a weird and unnecessary choice. But the crypto stuff is not overly complex. It is not all that useful now, true, but I see it is laying the groundwork for the long term goal of greater independence from the server, strong authenticity, and all kinds of applications that don't even look like social networks at all. Yes, most of the ideological drive behind the decentralization from blockchain people is fueled by libertarian fever dreams about destroying society and only having individuals, but we do need stronger guarantees around identity and stronger controls over our data, both things that the AT protocol is thinking about more deeply than the very limited-in-scope ActivityPub
The crypto stuff *is* overly complex because human beings just don't care that much about cryptographically verifying that you have access to a private key. Keybase tried solving that problem, it was unprofitable and failed.
The AT protocol *is not a protocol*. It is API documentation for how to get your backend server to work with the Bluesky apps. If it was a protocol for federation, it wouldn't include 'invite codes', hard-coded authentication requirements, or any of the other shit.
keybase tried to solve an important unsolved problem in the worst way possible and rightly failed. But saying no one cares about verifiable binding of human names is to say no one cares about certificate authorities or TLS
The original TBD pitch deck for Web5: An extra decentralized web platform (2022). Draw inspiration 790+ pitch deck examples from winning startups to craft your perfect presentation!
this is why they try to reinvent the wheel, because they won't allow anyone in their circles who will tell them otherwise. I have seen this time and time again.
They like to keep repeating that they are not a blockchain and not built on a blockchain.
But, yeah, at any point in time where a design decision had to be made, they choose the option that was most closely tied to the cryptochain bros while retaining the ability to deny any *overt* blockchain connections.
DID is one of the most awful W3C specs (and if you like bad specs, look into DIDcomm). And the hero image on ipld.io/ speaks for itself.
The data model of the content-addressable web. It allows us to treat all hash-linked data structures as subsets of a unified information space, unifying all data models that link data with hashes as instances of IPLD.
Not my field, but does it have be visible? It’s my understanding DIDs solve the problem of encrypting the ownership of data, so anything a user publishes (and most sub-pieces of it) are always linked to their unique user dID? Is there something inherently bad about that concept in your view? Couldn’t it solve a lot of different problems for a renewed and more open web?
(I have no dog in any related hunts. Just looking for clarity).
Yeah I don't think it's a bad thing, but the more I've used Mastodon the more I've questioned whether I really need that? There's some posts like this that I'm proud of and want to keep archived, but there's others where I could do without. And I don't know if I'd want to import the good ones into a new server or not. Kinda on the fence.
There's a VERY real problem with mastodon right now because all of the frontend is JavaScript-based. And all of the archiving tools that are generally used right now do not work well with that, except for a few. And so large parts of the Fediverse are just... not being archived.
A lot of people on the fediverse are here because they don't want to be archived (personally I don't mind as long as there's enough context for the rubbish I post that I look no more stupid than anyone else). @SirLich
That has a strong feeling of #SSB (#scuttlebutt) to it. Except even more CPU intensive and somehow requiring more than the infinite storage assumed by SSB...
I made a passive joke the other day that the enumeration of blocks reeked of "every object has all information about itself" in the same way crypto does and joked that @ was just a blockchain but I didn't expect to be mostly right in this way, fuck
Well that explains the claim I saw the other day that you’d be able to migrate all your content to another server if your original one up and vanished. Turns out the answer is even more absurd than I imagined.
A lot of my tweets are media tweets, that’s specifically what I’m worried about. I post a lot of memes. If I had to store them locally on my phone that would be significantly more than just 12 mb
There's one unfortunate exception to this: Hashtags are commonly used in nowadays social networks, but are often used for discovery of posts from beyond your immediate social bubble (recent events, campaigns, topics, conferences, ...). An almost complete view of posts is desirable in that case but are very hard to achieve in fully federated systems.
I would guess so yes, but there's no guarantee that an indexer has a global view of the entire network.
I can't think of a good way to describe what they're doing here. In ActivityPub, you react to a stream of notifications from servers, and you build your own view of what's going on through that.
With @protocol, you get a ping saying "hey there's something new!" and then you have to go and crawl the entire site and diff it and figure out what's changed.
That is extraordinarily hard and extraordinarily annoying and extraordinarily resource intensive to implement, especially because huge segments of users' history can just... stop existing at any point.
ActivityPub just sends you the events (this is a new reply! this reply was deleted!) whereas @protocol requires you to investigate the server and figure that out on your own.
Yes it can get a little more granular but that's still how it works.
With hashtags, I don't think you need things to be strongly consistent. Nobody on mastodon has a view of the whole network and we're doing fine. I see more important news on here than I do on Twitter, specifically more local news and more pro-worker news.
When you have a big global strongly consistent stream like that, it makes it very easy for lots of eyeballs to get on clickbait, which increases capitalistic and sensationalist articles.
Being on a micro instance with just 2 users, I quickly grew the desire of seeing more than the posts I already had in my timeline anyways when searching for a hashtag. Possibly just weakly consistent, but as complete as possible.
This motivated me to propose a DHT-based architecture for that. And you're correct, that's horribly complex. I never managed to implement it so far. :/
Ohhhh yeah true, you do definitely have to be on a bit of a larger instance. I think about 500-1000 users is the sweet spot. Relays also help solve that problem, and we need more of them (but higher quality).
to be fair, that's simply a function of a decentralized system.
You *could* build a centralized indexer for AP that follows the local streams of all the AP servers out there, and then you could ask that centralized server for the firehose.
But then you're dependent on that centralized server.
About the best hybrid you could ask for is a cooperative event bus between servers that have chosen to federate. And it's almost what you get with your global event stream.
this is an interesting observation that clarifies in my mind why AT seems to "smell funny" to me.
"Strong consistency" isn't a feature, it's a flaw.
Not only does nobody actually care to "have all the datas", not only is it very inefficient, it severely compromises moderation and runs counter to building healthy vibrant online communities.
Partial consistency that is controlled by people is essential to safe and sustainable decentralized social networking.
There is one thing Bluesky does correctly which Mastodon does not. Once you verify your domain, it becomes your human-readable handle. Turns out that this matters to people.
On Mastodon (specifically - I can't speak of all of fedi) even the verified domain isn't searchable - never mind it being part of your account handle. This plus erratic presentation makes it next to useless in practice.
I do not understand a lot of what you just posted but this part I do, and had a suspicion this was going to be proprietary on some level. I envision them selling instances like turnkey franchises.
Protocols are usually simple(r) to implement, and they usually layer on top of each other for a good reason.
The @protocol is none of that. It is difficult to implement for unjustifiable reasons and solves problems nobody has.
A key aspect of it (that whole "you can choose your own algorithm!" thing), is so difficult and resource intensive to implement that few people will operate those 'algorithms'. BlueSky will be the biggest one of those, and that is by design.
Thank you. I appreciate the education. One of the reasons I am here. From my side of the universe, Jack has extremely questionable judgement, is not the benevolent character people have made him out to be, and is not someone I trust philosophically. Your analysis of the technology aligns with my sociological one. He is as big a problem as elon. His motivations are questionable at best and destructive at worst.
ActivityPub beat BlueSky to the punch in becoming in a dominant, entrenched decentralized social media protocol that's rapidly gaining adoption. Bluesky doesn't just have to match ActivityPub to win, it has to be demonstrably, significantly better. It's not clear sprinkling some crypto on Bluesky will make it more delicious than ActivityPub.
The issue is that ActivityPub is *the* standard. It's just the standard for federation on the internet. There is quite literally nothing in this protocol that is so essential that the BlueSky team couldn't have built *on top* of ActivityPub.
You want to add signatures on top of ActivityPub? You can do that while still holding to the standard. You want to add DIDs or some other form of identification? You can also do that while sticking to the standard.
that would require Jack et al to be smart enough to copy 90s era Microsoft's Embrace Extend Extinguish business strategy instead of the SV "let's use technobabble and inspirational language to get lots of investment dollars" strategy modern startups go for all too often.
I think the fundamental reason why we keep seeing more and more bullshit protocols and projects pop up like this is one fundamental mindset: a refusal to attribute the problems of the modern internet solely to capitalism.
The fact is that our protocols we have today generally work pretty great. The choice of protocols hasn't made the internet what it is today. It's the fact that unrestrained, late-stage neoliberal capitalism is reaching its logical conclusion: corporate control of commons.
Obviously mastodon isn't perfect, I don't think it is. There's definitely a general 'NIMBY' vibe here of "this is how things are and we refuse to change", and we have work to do to fix that.
But I think what's key is to keep an anticapitalist mindset. We can make things easier for users without allowing in what makes social media so fucking awful: capitalism.
As a capitalist political moderate, I agree 100% with you on this. We have already seen the disastrous consequences of "engagement-optimized" ad-supported social media systems, which are inevitable when they are being fully controlled by single for-profit companies.
Thats quite the rant. I agree on some points, for instance the NIMBY-ism here, as soon as I get a chance to work on it, I'll quite happily build a client with both quoting & search, because I didnt ditch Elon getting to enforce his views on how I should use shortform social media to come here to get @Gargron to enforce his on me either.
The tone policing, the "we were here first," and just the general snark I've experienced are the reasons I don't hang out here.
I have an account on an Australian server but I live in Chicago like you. Why did I do this? Cuz of the messaging that I can join any server I want and make friends. Well I want to learn more about Australia. One of the first posts directed my way was "you don't actually belong here." Great!
Be very careful before attributing the evils of social media to capitalism.
There's no doubt that commercial interests do not help, data mining and skewing of the algorithm for the benefit of paying companies or specific agendas worsens things.
However, a substantial part of the ills of many communities needs to be laid squarely at self centred humans.
There was absolutely nothing stopping people carrying on using Livejournal or other longer form less self centred platforms. That's not what the community as a whole wanted, however, they willingly embraced the new model and many people still want a popularity driven algorithmic timeline on Mastodon.
@davidgerard I don't think that actually is the case. Its a fundamental misapplication of mistrust. The cryptocurrency world plays "hide the trust" not "eliminate the trust" all the time. And you can build systems where the bulk of things are untrusted but are fine to the user.
EG, your iPhone's cryptographic module doesn't even trust the base phone OS, which is why ApplePay is so great for payments as the security is very very top notch.
We know we can't change capitalism. Doing that would require changing the hearts and minds of more than 50% of people in our countries, and capital is damned good at beating us in that game.
Protocols are just the only way we can think of to try to carve out some control.
I think, like other adversarial environments, there are genuine steps we can take to harden protocols against the capitalism you identify (CRDT seem a good thing, for example). I don’t think they can be expected to make them invulnerable though as it’s not just a technical problem. And ultimately, we probably want better snob tech solutions; I hope we find them.
I will say, you lost me here. Like, I'll follow if you're talking tech concerns but the moment you start attempting to link it to economic/political ideology you've lost the thread.
Also I don't care if I'm spreading FUD or if I'm wrong on some of this stuff. I spent an insane amount of time reading the docs and looking at implementation code, moreso than most other people.
If I'm getting anything wrong, it's the fault of the Bluesky authors for not having an understandable protocol and for not bothering to document it correctly.
Thanks for this thread. I'd been curious if their federation stuff had any value. I'm surprised to learn it's crypto/web3 stuff, even though I was aware Jack & others involved are enthusiastic about that stuff I hadn't heard it described as a web3 project or anything.
the wild thing is that ActivityPub doesn't even have a solid test server that you can run against - we all just wave our hands and fix interop bugs and just call it a day.
i read it earlier. some interesting points. but in my experience it's not "crypto idiot" thinking. they have a plan, and explored lots of tradeoffs. there were reasons for the decisions they made. they've been pretty open about them.
Thanks for reading! The entire point is that those reasons are circularly justified and are not valid.
The sole thing you need to know is that instead of just doing an OpenAPI spec, which is the standard for documenting HTTP APIs, they invented their own that has no language support and can be expressed as a subset of OpenAPI + JSON Schema!
That alone should be enough to tell you that these people don't know what they're doing. But even if you go back to basic stuff like the DID:PLCs, it's literally just shittier DNS. And the crypto is 'necessary' because of that shittier DNS that does nothing and serves no purpose. And there is not a single thing in their FAQ about ActivityPub that is accurate or true.
"I'm going to rant about this thing that's not finished yet, I don't care if I'm wrong or if I'm just spreading lies, I'm going to do it anyway because engaging in good faith is overrated"
I think the thing to take out of all of this is that, objectively, the functionality the Bluesky authors wanted could've been built on ActivityPub. And that would've allowed it to interact with Mastodon and the rest of the Fediverse.
But they specifically built this in a way so that it doesn't do that. Meaning that I can't follow my friends on Bluesky, and they can't follow me. There is no good reason for them to have done that.
The reason this matters is that it poses a very real threat to the existence of a Fediverse. The Fediverse allows you to follow and communicate with people *across social networks*. And this protocol completely removes the ability to do that.
If this gains popularity, it has the potential to become the norm, and that would not be a good thing. Because that interoperability aspect would be lost.
Ouch. I saw a post from the middle of this thread, opened the thread to read, and .... yikes. 😱 Sounds like a pretty vicious indictment of BlueSky over all, at least on the coding side.
You did mention "crypto" in a few posts, and I'm wondering if that's just "encryption" or if they actually built cryptocurrency/blockchain BS into the AT Protocol? I may be a bit dense about that, but I could use a clearer sense of it. (Others probably could usefully know this too.)
That is just encryption (signing), but it builds on top of protocols created by the crypto (cryptocurrency) community, and the founder and owner of Bluesky is a crypto person. There are a number of objectively useless and overcomplicated parts of the spec, which is a hallmark of more scammy cryptocurrency 'protocols' that promise to change the world.
I dont think this is a useful way to talk about this. Mastodon has a higher learning curve and I feel like there are definitely ways we can and should reduce it.
Criticism of Mastodon should be looked at seriously, especially when it's from black people, as this space is overwhelmingly white. That criticism may be frustrating to engage with, but I think we could be more welcoming on this.
You’ve made an excellent analysis here! I recommend you to blog this standalone as a blog article. Would be better to link to on Twitter and … Bluesky, I guess? The message that the protocol is designed around a main instance etc. is very important.
Planning on doing this eventually! Only problem is I'm really bad at writing articles and really great at making Twitter threads. Moonbeam (GPT-4 long-form writing app) has definitely made it MUCH easier for me to get feedback on an article and get ideas for how to write things better though.
Also I feel like with writing I have to be more 'airtight' and make sure I have things perfect, and frankly it's extremely irritating to try and figure out what's going on here lol.
Also you're not going to be able to show this to anyone and convince them not to use Bluesky. Folks are using Bluesky because it's easier for them to use and understand, and because Mastodon has garnered a bad reputation. This is a wake up call to us: if we do not make the app easier and more accessible, people will switch to something that will leave the internet worse off.
This will not convince normal users to avoid Bluesky but this is a valuable info for tech folks, journalists and media CEOs who might invest into the network because they assume it’s as open as the fediverse.
DIDs serve a perfect purpose — the entire goddamn thing is a trojan horse for a cryptocurrency grift. like, look at the authors of the DID spec, how fawningly they talk about "distributed ledger" nonsense.
Not competent to comment on any of this, but have seen a lot of vaporware tech schemes over the years, and do have a question.
Can you tell if Bluesky is actually utilizing the @protocol for the single instance platform they in operation? Could this all be inoperable spaghetti code and marketing puffery with a stripped down Twitter clone operating in the background? Just curious...
bluesky reeks of a techbro playground where everyone thinks engineering alone will magically solve everything that's wrong with social networks, I think that's why they're trying so hard with all the crypto crap
Cuz there's literally NOTHING that #bluesky does better that any other #ActivityPub software - not only but including #Mastodon - already does better, more scalable, more reliable and more efficient by a long shot.
Sam's criticism was of the atproto tech underneath Bluesky, and in that space, I think that the complaints are reasonable and legitimate.
But going as far as "There's nothing it does better than Mastodon" just isn't true. Bluesky (the app, not the protocol) does one thing exceedingly well: It looks and acts exactly like Twitter.
I don't see that continuing to be true once they add federation, but as a user, that is a huge difference.
You can say all you want that Bluesky offers nothing, but the reality is that it offers me an experience I like better because it is one I am used to. I understand how threads work; I can see reply counts; I can see like and retweet counts; I can post quote tweets; when I click on a post, the replies to that post are displayed without clicking again.
Because Bluesky is a single instance, the experience is much better for me than Mastodon.
This is definitely true and valid. Mastodon's web app is okay but needs a lot of improvement. Mobile apps definitely fix a lot of complaints for me, but the basic web app needs a LOT more work.
I took a look at the frontend code and it's going to be difficult to rehabilitate imo. Old school react, in JS not TS, react-redux containers with immutable.js structures. It's not bad per se but it's fallen out of fashion and would be a lot of work to update. Might be easier to just make a new one.
It does have the property that it's straightforward to write alternative clients on top of the Mastodon API (not ActivityPub); elk.zone is a decent take on something more twitter-like. It runs into the limits of the Mastodon API, especially around large threads like this one, which just take ages to load, but it's a lot closer to what you're familiar with if you're a Twitter user.
Yeah I think Mastodon definitely needs to rethink some of their API decisions specifically for dealing with larger threads like this one, or for more viral posts. With the amount of notifications I’m receiving it’s incredibly hard for me to stay up-to-date with what’s going on in the thread here, but it’s not so bad if I’m using a third-party client like @MonaApp.
Notification grouping is a key benefit that I am happy to see in any client; Elk does some of it (though less than I would prefer), and Ivory does it too. It's a huge pain in the ass when skimming the notifications tab. (The separate mentions tab does sort of mitigate that element, but "notification-based dopamine" is a valuable part of using social media for me.
There is only one proper form for the account identity in a distributed system: a public key, and a data object signed by said key, which points to servers and other things as required. Then that identity can be moved and updated by the owner of the key and by nobody else. The key ID can be stored in a Kademlia to make it globally findable.
Something I wonder about, given your analysis. When public institutions think about where to host their own social media, they will want as much control over the process as possible. Does this mean they and the technical people who are advising them will feel more comfortable using activity pub
can't say I understand much of this, but I will say this, as a user, Bluesky is fun and easy to use.
Mastodon on the other hand. Making people copy and paste URLs to follow other users is just really bad design. Mastodon feels fragmented. It may serve a niche but it's never going to take over from Twitter. But I think Bluesky will, the AT Protocol that is.
I literally criticize Mastodon further up in the thread. I would very likely not use Mastodon if I weren't already running a server. The critique here is interoperability and the lack of an actual protocol (this is literally just a set of API methods).
Copy / pasting URLs to follow is a fixable design decision, I don't agree with it, but there's nothing inherent about ActivityPub that requires it.
Yeah, you can definitely tell it's cryptocurrency developers. It's like the "whitepaper" for some new token where they're not doing anything novel or interesting so they just build a mountain of complexity and indirection. It's all supposed to make it look like there is some "technology" there and not just bullshit.
Come on! Why should everything be about race?! Based on what you say #Mastodon is overwhelmingly white? And even if it was, it wouldn't be because there is a gate that only allows whites in! Also, there are people other than "white" and "black" too. Criticism of brown people matters more too? What about Asians? I'm asking cuz I wanna know how important or irrelevant my criticism is with your logic.
Bluesky sells themselves as a federated social network. There is absolutely no reason why they could not have built their protocol to be compatible with ActivityPub.
No, they're not. If they're going to sell themselves as a federated social network, they need to have valid reasons for not implementing *the* standard for federation. And they simply do not. Their reasons are bullshit and they demonstrate repeatedly that they do not understand ActivityPub.
this is the most salient thing — I've been on both of these platforms for as long as they've existed, and was involved in the predecessor work that led to them, in many cases, and simply put — there are lots of valid reasons why most of my network from Black twitter is either on, or considering, Bluesky, and isn't on Mastodon. And "their protocol sucks differently" isn't one of the concerns.
I mean maybe, maybe not. If we’re talking about the protocol and building a community server, it absolutely is a big fucking pain in the ass and I don’t think many will bother. But I don’t even think people realize Bluesky is federated and will care.
I don’t think a lot of developers will want to work with the API either because it’s also a pain in the ass to work with. Because there’s no code generation for any language except TS.
the fediverse might well stay the non/anti-commercial/capitalist community and I'm fine with that. What happened to twitter will happen to bluesky, and people that got lured will end up back on the fediverse later.
Great thread, a little harsh but I can see the reasoning even on the parts I disagree with. I would push back on a few points: 1.) DIDs were not invented by crypto people and do not require blockchains at all. DIDs are a translation layer and data model for decentralized PKI systems (including, say, PGP, git, radicle, etc), that's it. They facilitate interop between key-based identity systems, everything else is means not ends. FWIW I'm one of those DID people.
I think I know this feeling from my time as math tutor. You get something that looks almost right but something is off, a small detail, and then… you scratch it and somehow the whole wall falls off. It’s like realizing the people who wrote it got everything backwards, probably think (metaphorically) the sun is the moon, the moon Pluto and earth an ice cream cone. Everything they got right was just an accident. Took 2 hours though, probably more in your case. Thanks!
Rereading this thread months later. Thank you for the research and the thorough writeup. This thread is now a useful resource; a friend cited it in a thread today and I'm glad to reread it.
the wost part is that nobody that's flocking to bluesky is paying any attention to this or worse, just straight up doesn't care.
Also, it always seems that anyone who gets into crypto develops a complexity addiction that makes video game villains seem straightforward in their plotting. So, of course, bluesky was always going to have that problem of being needless complex behind the scenes. Plus, the slowdowns from resource hogging activity are likely going to be its downfall.
This thread reads like everything that goes through my head every time I have tried (and bounced off) to dig into the ATProto specs. So much of it seems to be intended to make it impossible to actually federate with them - either because you need to implement *everything*, you need to spend thousands in hardware to host/process *everything*, or just because the spec heavily relies on crypto tech for no obvious reason. When you want to use it to build something, it's infuriating!
Watching Queen Charlotte since it's a Bridgerton story after all. But after being disappointed with series two and already knowing none of my favourite characters are going to be in it, it's going to be difficult to get into it.
So far I think the character of young Queen Charlotte is very interesting, but I dislike that she doesn't match at all, writing-wise, with the older Queen Charlotte we already knew. At least young and older Lady Danbury match very well, allowing the actresses to really sync their performances. A lot of it is Lady Danbury's story, and King George's as well.
Skipping adverts as a subscribed user on Amazon Prime Video takes about as long as watching them, but I still skip them out of principle.
Ok, so you can skip them and they are at the moment only adverts for their own programs, but I think as a paying subscriber to their streaming service I should only get the stream I'm clicking the start button on.
Image description: comic by heyokyay with two figures talking. A: I used to always postpone things I had to do until the following day... B: Sounds very familiar. Usually it's a mistake. A: These days, I postpone things until the next year. It's much less stressful!
There's this shop I need to go to and it has different closing times on different days of the week. The only constant is that when I look it up, it closed five minutes ago.
If you have to deploy and test several µservices all at once, what you effectively have is a very inconvenient overcomplex monolith. If they're not independent, they're not µservices.
In trying to find out why a new version of a thing uses so much more resources, I've come across an interesting new data point. When I turn it off, the warning that it's off doesn't appear and the things it was supposed to be doing, are still happening.
So this site documenting the AT protocol has a frequently asked questions section, but no way to ask questions. I'm suspicious.
That FAQ explains why they couldn't just use ActivityPub, by listing a couple of things missing in it. These things happen to be the exact things the Zot protocol does have, while Zot has existing implementations that do interoperate with ActivityPub. They couldn't have looked long. #AT #ATProtocol
This Saturday the interesting Chloë Grace Moretz film Greta is on BBC One: 23:40 UK time, 00:40 CEST. Of course I have it on DVD already, but if you don't you can catch it. #ChloëGraceMoretz #Greta
Reminded of the time I wrote a small useful utility that nicely ended with an exit(), which later got incorporated into a web application because it was so useful and which took me an embarrassing amount of time to debug why the server always 'crashed' after clicking my new button.
When changing output resolution on a chromecast, selecting ok in the menu does nothing no matter how many times you do it, until you also hit cancel, then it sticks. Just now found out that just doing cancel twice, no ok, also succeeds in changing the resolution.
I have to keep doing this kind of thing because my tv and chromecast have a dodgy handshake.
When I switch them on, the tv and chromecast will negotiate resolution, hdr settings and colour encoding and agree upon which to use based on the best available modes. Following this agreement my tv will do one thing and the chromecast another.
I feel long running sci-fi series, whether it's Star Trek or Doctor Who, are harmed when their stories start to cater to nostalgia. Just create new stuff, like you used too.
Checking whether I've done today's wordle already since it's the new one in a few minutes and I'm not going to lose my current streak yet again just because I forget a day. I'm at 65 on the English one and 29 on the Dutch one. Luckily I did apparently do it already. (I very often fail on the film frames one but it's fun to try that one too)
Gidi Kroon
in reply to Gidi • •Gidi
in reply to Gidi • • •