Hacker News new | past | comments | ask | show | jobs | submit login
Bluesky Frustrations (anderegg.ca)
43 points by GavinAnderegg on May 9, 2023 | hide | past | favorite | 69 comments



The biggest frustration I have with Bluesky is the velvet rope and its consequences on community building.

I've been on the wait list since Feb 22, but for whatever reason I don't have access. That's just sour grapes, I know, but it also means that others who don't run in the right circles also don't have access. It means the community is intentionally exclusive and cliquey.

I don't know what it is about tech that everyone wants to become the people who abused them in high school, but it's a bad trend and we should be better. If invites aren't going to be chronological they should at least be random. Invite people who are different intentionally. Otherwise we end up with another clubhouse-esque VC circle jerk.


I can't say I blame them for not wanting to suddenly have to firefight 1 million users on an immature platform.

Aware that your issue is the approach and the fact that invites are the main way of scaling this rather than just issuing random waitlists, but there is an issue about networks which I assume they are needing - i.e. users already on there will want their friends on there with them, as it is a social network.

The issue about creating networks of people that know each other and interact is probably pretty key to building a real social network - see how Facebook rolled out by college campus.


I also don't blame them for being invite-only and have no intention of begging people for an invite.

But, the inability to check any post by people already there is absolutely unacceptable. Not being able to verify a screenshot of public figures is just a huge no-no in my book.


That's on the client, not the protocol. A few folks have built clients but for some reason hasn't made this a feature.

All posts are publicly accessible though -- here's Jake Tapper's posts for example

https://bsky.social/xrpc/com.atproto.repo.listRecords?repo=d...

EDIT: This client lets you look up any post/profile without logging in https://blue.amazingca.dev/?username=did%3Aplc%3Aupkgvxvhn4e...


They have self authenticating data structures no? Someone should be able to write a program that can verify any post.


Yes, only a web-frontend client is missing, see someone else who downloaded all posts https://news.ycombinator.com/item?id=35845504


I don't think they even have anything (note I am purely speculating) beyond a single ec2 instance at the moment. it's going to take a lot of time to scale the necessary infra to support the demand


So decentralized it can't even run on a second server yet...


I will also say, looking at the quality of the app in its current state, they just do not have a strong enough engineering team. They've been working on this for years and the UI/UX is extremely rough.


I get and agree w/ your general point. But at least for Bluesky watching folks in my feeds talk about getting access it's been a pretty diverse group. Much, much, much more diverse than Clubhouse. And by that I mean across industries, economic groups, etc and not just "diverse within Technology"


Watching folks in my feeds talk about getting access to Clubhouse back in the day and not getting in was a pretty diverse group too ... I think I might still be on the waitlist at clubhouse - by the time my number came up I had less than zero interest in the whole thing.


Granted both our posts were anecdata but what I didn't add was that for me Clubhouse was the opposite. The folks I knew joining Clubhouse were very much cut from the same cloth.

I already knew I didn't want to use it as I have no interest in using a voice/audio based app. But the membership didn't help any either.


I signed up for the beta on February 21st and got my invite last week at the top of May as they started letting folks in. From a graph I saw, they only had a couple hundred users til about April, then the May batch brought in another 50K (of which I was one). It's at 65K total now I think.

I have a negligible following on social media and don't know anyone over there so I don't think they're being cliquey with the invites.

To someone else's point, they're doing a lot of firefighting. When I got in, your followers list did an infinite loop, "hellthreads" were causing errors every time you viewed a reply, and content moderation was (and still is) constantly breaking (NSFW stuff keeps breaking through on their placeholder "what's hot" algorithm). I was able to register @salesforce.herokuapp.com as my handle, etc.

I think it's a beta in the truest sense of the word and you'll probably get an invite soon.


It was quite telling how easy it was for Musk to buy Twitter. Most founders of successful startups fight tooth and nail to hold on to their company but Twitter rolled over so fast the deal was done before Musk had a chance to change his mind.

The problem is with the product space not the product. There is no fix for all these issues - only compromises. Bluesky will not be any different - people will be toxic, unhappy with moderation, unhappy without moderation, want blue checkmarks, think blue checkmarks are BS - blah de blah blah blah.


Hey you seem like a cool person, I'll have an invite soon if you want me to send you one, my email is in my profile here :)


> People are having a great time! But I suspect this is because the service is currently small, simple, and centralized. Once the decentralized systems are in place, there’s a good chance it’ll be more confusing.

Precisely! This is what has been ticking me off about the Bluesky love-in so far; it's as if people are so desperate to call something "new Twitter" that they forgot the systemic factors around why Twitter turned bad.

Today, BSK is exclusive. Getting an invite gives one the ability to peer into what's happening at the cool table. It's invite-only, so most people there will be cool and not try to clickbait or build clout.

Problem is, Facebook was exclusive once too. It didn't last. Twitter also was once just a quirky little online space. Then the Arab Spring happened and now everybody writes with an assumed gravitas, as if their 280 chars are going to be featured in a CNN story.


> the systemic factors around why Twitter turned bad.

Being controllable by one guy who does dumb stuff on a whim?

That's a tough, tough nut to crack, I think.

On one hand you can just make something totally open, and then it turns into a "hive of scum and villainy" and normal people don't want anything to do with it (something like: https://www.upworthy.com/bartender-explains-why-he-swiftly-k... )

Or you can have strong moderation and ownership. When that goes well, like HN, it's about as good as it gets in terms of being a mostly pleasant community to participate in. But that could all go away, if, say, YCombinator sold it to the Saudis or the Kardashians or whatever...

I can't for the life of me figure out why someone who was doing some genuinely cool stuff like rockets would want to get involved with something where even when it's working pretty well, people are going to complain about it and think it's unfair.


>Being controllable by one guy who does dumb stuff on a whim?

The entire twitterverse has been complaining loudly about how terrible twitter is for years before Musk considered buying it. The problems are inherent to the product-space, not the owner.

Musk simply came in and screwed with the brand. His ownership didn't really affect the toxicity of the platform as far as I can tell (although he has definitely contributed to it with his personal usage, e.g. dogecoin shilling). If anything, he actually took heat off the overall toxicity by shifting attention to his own actions.


> His ownership didn't really affect the toxicity of the platform as far as I can tell

This is not at all my experience. The new thing with the 'blue check' replies makes it much, much worse.

He let a bunch of right wing extremists back on, as well as various and sundry other bozos.

Now he's given Tucker Carlson, the guy Fox fired for being too racist, a platform again.


I signed up for it months ago and got an invite code. It’s therefore not invite only.


If you need an invite code to join then it is very much "invite only"

I also signed up months ago and yet am still waiting for an invite.


Eh, not if you invite yourself.


“I have a sandwich, therefore no one is starving”


I'm happy with Mastodon and want the fediverse to succeed. I don't want one single project controlling the fediverse. I have a feeling Bluesky is trying to Microsoft (embrace, extend, extinguish) the fediverse. Why? Because there is no clear way to make money off people in the fediverse. You can't rent seek like the centralized services do. My sense is Bluesky is embracing federation the same way Microsoft embraced open source.


Given that Bluesky does not use ActivityPub, it's hard to make a case that it's embracing, extending, and extinguishing it, given that there's no embrace in the first place.


It's embracing just enough things that look vaguely ActivityPub-like if you squint and is getting a lot of marketing that the parts it didn't directly embrace it extended/chose "better" alternatives, such that from a PR perspective at least does seem to using a similar playbook, even if not that exact playbook.


There's already a very simple way to describe this kind of relationship: competing. They're competing protocols and products.


That's a given, the concern is still valid that the asymmetric nature of the competition may be "unfair". One is an adopted open source standard of a small but diverse ecosystem and the other is a well-funded, seemingly more popular, more closed than open with a lot of vested interest in crashing the open source ecosystem.

Especially when it feels like the "more closed than open" is presenting itself as "open source" in what feels like a bait-and-switch to piggy-back on the goodwill of the actual open source ecology with a goal towards moving people (back) to more closed, corporate controlled tools.

("bait-and-switch" was the word I think I was looking for here.)


> more closed than open

Literally 100% of this is open source. (Oh wait I think maybe one of the mobile apps isn't yet, but like, it's react native, IIRC, so it shareds most of it with the web app, even if not 100% of it is.) Yes it is true that there's no open governance yet. That is a problem. But let's talk about problems in a straightforward way, grounded in actual details, than in FUDdy ways.

"They are competing and I would have preferred if they weren't" is much easier to have an actual conversation about than "they're trying to EEE mastadon", which is inflammatory, inaccurate, and does not promote mutual understanding.


I think you are nearly intentionally misreading the conversation then. "They are competing in a way that seems like EEE for Mastodon because it feels like an open source bait-and-switch" can be a very useful, actual conversation. The conversation premise starts from a presumption "it's not EEE, but it feels close, let's discuss why/why not because it can help dispell fear, uncertainty, and doubt or sand-bag the FUD with reasons". FUD on its own isn't the problem, it's what you do with the FUD. People telling you that they feel FUD about a project can be a useful gut check and a place to examine what missing information or what missing evidence to go look for.

Personally, I believe that driving this towards meta-conversation is less useful to "promoting" mutual understanding than trying to use an analogy towards EEE as a way to describe feelings about a project. Analogies are useful constructs and dismissing analogies as "inflammatory and inaccurate", especially analogies involving feelings and opinions, is its own "inflammatory and inaccurate" move.


That's fair.


Bluesky is going to be successful because of its current state of being invite-only and sponsored by Dorsey. It's essentially the new blue checkmark: an indication that you're part of the group of "always-online Twitter powerusers". Once Bluesky opens up people will join because they either want to be part of that group or the people that you'd follow on pre-Elon Twitter are part of that group. Technology has little to do with it


Yep, since Twitter's UI is extremely easy to copy, all you need is the users and the culture. Granted that's usually the hardest part, but Elon is giving everyone tons of reasons to move away.


Most people are frustrated because they want to be on it but don't get an invitation (me included).

I'm surprised how these old tricks still work very well.

If Bluesky can catch the journalist as Twitter did there is a good chance it can replace Twitter. Even though I saw some news outlet set up their own Mastodon servers.


"These were hashtags, posts, or users the service thought I should see – and there was usually something in there that left me feeling worse."

I'd gotten into the habit of catching up on social media in bed before getting up in the morning. Then I noticed that I was getting up pissed off most mornings, so stopped the habit as an experiment. It was the difference between being my naturally cheerful self in the shower, and some species of low level rage monster.

I've spent a life as a voracious newshound and infovore but have come to see info-gluttony as somewhat analagous to calorific gluttony. The sources of calories really do matter. A more hygienic media bubble is a better place for me to live.


I'm with you here. I have a personal rule that the first information into my head each day comes from the Bible, which for all the criticism it gets is ultimately a story of hope in the face of many disasters. It's been a huge boon to my mental health.


So absolutely little of what is out there is actionable or needed; just ignoring it and focusing on your day and your people is hard (it's like avoiding snack food) but it is so worth it.


It looks like Bluesky is already in the lead of Twitter alternatives with fast growth and attracting normal users (not techies) and demand in less than a year after launching their app.

Users don't need to 'Choose a server' and aren't confused on signing up and know how to use basic features like search and can find the users that are who they say they are. For BlueSky, two months with this growth in adoption is very early days.

But we will see what happens when the invite system is lifted and if the users are retained enough on BlueSky and what the users on Twitter will decide to move on to if Twitter gets 'worse'. But it seems that BlueSky so far is winning amongst the rest of the other failed alternatives that have tried for years to challenge Twitter.


I wonder how well the "find users who are who they say they are" is going to work once the test server starts federating. They've got this whole platform now with users who never got the brief on federation and its implications, so it'll be interesting to see what will happen once they unleash federated identities on the unsuspecting masses.

BlueSky is already worse because you can't just link to public skeets, you need to be in the special club to see the stuff people post. Sure, you could use some kind of external service that parses the underlying data, but what user is going to use that?

I'll happily stick to Mastodon until the people I follow all disappear to nostr or BlueSky. I might give running a BS server a go, but with the utter lack of documentation there is now I'm guessing that'll be in a few years at the earliest.


> Sure, you could use some kind of external service that parses the underlying data, but what user is going to use that?

I already do this for Twitter, given it fucks up Discord embedding, and also use a similar tool for Bluesky, called psky. All you have to do is turn the "b" in "bsky" upside down, and change it to "psky". Not hard to use. I've heard some Bluesky clients also give out "share" URLs that are world-readable.

That said I agree that posts should just be public (and my understanding is that the current state is temporary).


I hope the Bluesky PBLLC will submit their "AT protocol" to the ietf as an rfc. we already have an existing protocol though


Maybe they’re not submitting it because they’re worried the IETF will say “we already have an existing protocol” and use the standards carrot as a way to force them to make compromises.


they can submit an I-D if it were so desired:

https://www.rfc-editor.org/pubprocess/#:~:text=Anyone%20can%....


> But we will see what happens when the invite system is lifted

The devs have said that they don't intend to lifting the invite system, because having a tree of who invited who is a powerful tool for detecting things like spam.

It is not part of the protocol proper, and so once federation is turned on, it will be lifted in that sense, but it's not yet known how widespread the invite system will be used throughout the federation.


Serious question. The issue of public block lists comes up quite a bit. I’m not sure why this is a problem?

This could be me. I’ve always used social networks with an assumption that everything I did was public.


https://mosquitocapital.substack.com/p/bluesky-is-not-ready

> 2) Blocking: Your list of blocked users is public. Anyone can see it through the API. Someone harassing you? They, and all of their buddies, will know as soon as you block them. People can build profiles of who's ripe for harassment just by searching for large blocklists.


>Someone harassing you? They, and all of their buddies, will know as soon as you block them.

But Twitter has a huge "This person has blocked you" page? This doesn't sound like a massive upset to the status quo.


But this is not a "discoverable" thing, and only the person who was blocked can see it. I (or more contextually appropriate, a bad actor/harasser) can't trivially check if someone else has been blocked by another user on Twitter.


Imagine you were a touchy person with a big audience. Also imagine you had a tool where you could find all the people who blocked you, and you noticed someone important on that list. You could make that person's life (at best) quite annoying if you wanted to know why they blocked you.


It's a harassment vector that has even been the topic of a lot of controversy in Mastodon already. Some instances got harassed for their public instance block lists. It wasn't even personal blocks, but instance-wide and instance-to-instance blocks (and then in-fighting). A lot of instances like the transparency of public instance blocks, especially as that can be a helpful tool for evaluating an instance's admin policies from the outside if you are looking for an instance to join. However, the same helpful tool also gives trolls ammunition when they are looking for "fun" fights to pick and suggests who might be "weak" to their attack.


It could be seen as defamation. If the block list is titled “nazis” or “antisemetics”, two of the worst labels you can be assigned in modern western society, then you can be publicly described that way with no context or recourse.


I wonder if Bluesky will be Clubhouse all over again: it was cool in the early days when invites are hard to get, but the novelty wears off quickly as it grows? Guess I'll find out once I manage to get an invite :D


Mastodon will fail simply because nobody is paid to work on the project and things the core team are not good at (like creating a good UI/UX) will never happen. And that's going to continue to be the perpetual hurdle to user adoption or anyone taking it seriously outside of tech circles.


Mastodon can only “fail” at being what it’s not trying to be: a Twitter replacement.

It has already succeeded at being a decent social network for people for whom Twitter was too toxic.


>It has already succeeded at being a decent social network for people for whom Twitter was too toxic.

I don't see Mastodon being less toxic, the whole mastodon.lol (~70k users) drama over a video game was a perfect example of that.

On the other hand it also showed to me that single user self-hosting is the only way to go with Mastodon.


idk, I get way less abuse on the Fediverse than I ever did on Twitter. There is definitely drama out there, and the mastodon.lol debacle was well, ridiculous. (As Douglas Adams said, "People are a problem." -- that's not going away no matter what nifty new protocol you come up with, unless it exclusively routes to /dev/null.)

My point was more that there's no real profit motive to Mastodon, so it can't fail in that sense, which is generally what we're talking about when we're talking about web services.

If 75% of the new people on Mastodon eventually go to BlueSky, the people who remain will probably be perfectly happy to remain on Mastodon, and I don't know how one can call that a failure.


Didn't Mastodon recently announce search improvements and quote posts, which were some of the Twitter features lauded on Mastodon as "toxic"?


Yes, but with some differences (opt-in, being one) that hopefully will prevent some of the abuses.

Moderation is also way better on Mastodon, so even with those features, the impact should be lessened.


plus you're not sticking the ActivityPub protocol genie back inside its bottle.


> (like creating a good UI/UX) will never happen.

Ivory is a fantastic UI for Mastodon. Mastodon's only big hiccup really is the confusion around picking an instance. But that will go away once a few "normie" instances get a foothold and serve as a kind of "America Online" for the Fediverse.

Secondarily the fact that it doesn't tune for engagement means casual users will tend to bounce off of it (since there is nothing to sink them in and keep them coming back). But to many that's a feature rather than a bug. Akin to losing smokers by banning smoking in bars.


> But to many that's a feature rather than a bug.

I don't use Mastodon (nor Twitter, nor Bluesky, etc.), but if that sort of thing did appeal to me, Mastodon would be the only one that would be acceptable precisely because it doesn't tune for engagement.

"Tuning for engagement" and/or using the usual "stickiness" tricks is a huge red flag to me.


> Ivory is a fantastic UI for Mastodon.

for Mastodon being key here because they still haven't made it able to correctly talk to non-Mastodon instances (I'm convinced[1] this is because they've used `int64` for their status IDs despite the spec saying don't.)

[1] My Akkoma instance shows no posts - status IDs are opaque, non-`atoi`-able strings. My GotoSocial instance shows many posts but they are all the same post - status IDs all currently being with `01` which would explain the presence of posts and all being the same. In summary, stupid.


Mastodon has hugely improved my user experience by being open to third-party apps like Ivory, which is amazing (and much better than Bluesky.) For people who want algorithmic content, presumably third party services could handle that problem as well… except that the Mastodon community has been pretty opinionated about not allowing content search an aggregation at a project and instance level. This sort of from-the-top hostility makes that kind of extension into a risky project.


People are being paid though. There are several Patreons and the core project runs one.


It's peanuts compared to VC funded projects.


Considering that VC funded projects tend to attract "career oriented" developers that have no incentive other than money or prestige, i'd say that could be a good thing.


> Mastodon will fail simply because nobody is paid to work on the project and things the core team are not good at

That's not the reason, the reason it will fail is that although federation is a strength in some ways the other way is you're putting your faith in maybe just 1 person moderating and maintaining the instance you choose. If you build your audience at one address that place might disappear overnight if that persons life circumstances change and then (correct me if I'm wrong) you need that persons permission to move to another server.


Exactly, a lot of people expect VC funded output from Mastodon when Mastodon project only gets about $30k a month. Enough to pay for 1 developer, 1 designer, and not much else.

Tech doesn't get funding because it's good, it gets good because it's funded. Blusky will be the Windows to Mastodon's linux.


Person who never used a service has opinions about said service. Not much to see here.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: