Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

> The social graph lock in problem is well documented and well understood.

i don't actually think it is. i don't know _anyone_ who uses just a single messaging app (and thereby a single protocol-level social graph). i have some mental map in my head: "if i want to reach friend A, i do it on Signal. friend B: Discord. friend C: SMS/tel/PSTN. friend D: Matrix". i think this is a pretty common experience these days: i'd hazard that my mix of 4 apps is on the _small_ side.

i admire Beeper, JMP.chat, and other groups trying to improve messaging via better abstractions. i think it'd be cool if they could maintain iMessage support, i also think it's not critical to their success. the pain points caused by that graph problem you point to is 1) maintaining that mental map and 2) coordinating large group chats. i don't see that the client-side/Beeper-style solution to this is notably worse if they support only 29 protocols instead of 30: for as long as my peers are reachable by more than one messaging app, the odds of bridging between them isn't radically different.




> The social graph lock in problem is well documented and well understood. > i don't actually think it is.

Nitpicking but I was saying that the general social graph lock in problem (also referred to as chicken/egg) is well documented.

> i don't actually think it is. i don't know _anyone_ who uses just a single messaging app (and thereby a single protocol-level social graph). i have some mental map in my head: "if i want to reach friend A, i do it on Signal. friend B: Discord. friend C: SMS/tel/PSTN. friend D: Matrix". i think this is a pretty common experience these days: i'd hazard that my mix of 4 apps is on the _small_ side.

Hi! Nice to meet you! I use only one messaging app for all of my friends! It's called texting. As far as I know, all of my friends do the same, with the only exception being a few Internet-only friends where we use Discord.

The "mental map" that you are describing is exactly what I want to avoid. I am thankful that I have not had to make one yet, and when people tell me to use over-the-top chat apps like Whatsapp, I can see that the map must be made.

Just because this is the norm, doesn't mean I'm going to do it, especially since we don't do it now. As much as the interoperability problem between RCS and iMessage is an incredibly annoying problem, I would take a single unified messaging experience over some crazy fragmented one with a zillion apps any day.

> 2) coordinating large group chats. > for as long as my peers are reachable by more than one messaging app, the odds of bridging between them isn't radically different.

A little confused by this, because Beeper and other unifying clients cannot in fact make groups which have participants on multiple platforms at all.

You said you need 4 messaging apps right now to communicate with everyone you communicate with. How many of those users also have all 4 of those messaging apps? Obviously it's not all of them, or you'd just use one messaging app. The fact that you need four implies that for a given selection of contacts, there is a chance that it is impossible to create that group chat, because there is no shared platform they are all on. Then you factor in that in some scenarios you need your contacts to include additional contacts, and perhaps your 4 messaging apps needs to grow to make it happen. And of course if you already made the group and you need to just add one more person then you might have to scrap and remake the group somewhere else. But then that group that already has some messages in it still exists, and people will keep texting it! Now you've split your group chats!

On top of this, I want to note that the mental map you have built is also prone to becoming stale. If one of your friends is on Signal and Whatsapp but prefers Whatsapp, but then uninstalls Whatsapp and forgets to tell you, then you very well may send a message to that person and have it never arrive. Of course they might bail out of both Whatsapp and Signal, and just go back to SMS. Now none of your messages will land- you didn't even think they were interested in SMS.

Sure, if they are a close friend its likely they'll let you know. Most people have 1-5 close friends. But most people also have far more contacts in their contact book, and some of those people they might only message a few times a year. That's not a mental map that can be maintained, or if it can, I don't want to.


> I use only one messaging app for all of my friends!

i admire the resolve. on the other hand i think that rules out iMessage playing much role in that long-term, right? like, they're just never going to play nicely with others, it's not easy for the broader developer base to integrate with much less improve, and so on. so you're back to SMS, and the baseline SMS experience now is pretty limiting and stalled (much as SMTP stalled): a big part of why people leave for app-based messengers is for features like voice memos, video-chat, multi-device (e.g. PC) support, better multimedia support, etc. to say "SMS forever" i think is to say "i'm okay never having these features" -- which is a fine decision but important to note.

> A little confused by this, because Beeper and other unifying clients cannot in fact make groups which have participants on multiple platforms at all.

i'm pointing to where i understand the landscape to be headed. for channel-based chat systems like Discord, irc, Matrix, XMPP/jabber, Slack, it's common enough to find channels which are bridged across 2 or more of those protocols. my experience with ephemeral group chats is that if i want to plan a large enough event i just end up starting multiple group chats, and the unimportant details are chaotic but the important ones like where/when we're meeting i make sure find their way into both chats. there's a possible future where i start two group chats and my client bridges messages between them in the same way those channel-based systems bridge.


> i admire the resolve. on the other hand i think that rules out iMessage playing much role in that long-term, right? like, they're just never going to play nicely with others, it's not easy for the broader developer base to integrate with much less improve, and so on.

Well Apple is implementing RCS, so that's good. But look, I don't really think the blue bubble stuff stems from not being able to put stickers on the conversation. It definitely doesn't come from not being able to emoji-react ("tapback" as Apple calls it) because that still works on SMS, but the SMS participant receives a text message describing the tapback. In Google Messages and other modern clients, that gets interpreted by the phone and turned back into an emoji reaction [1].

I don't think the blue bubble hate comes from people not being able to do inline replies. I don't think it comes from the inability to edit your messages when in an SMS conversation.

The source of the blue bubble hate comes from group chat splitting. When you have an iMessage group chat and you hit Add to add a new user, but that user is not an iMessage user, you are shown a prompt that says "Create a New Group? Contacts not using iMesage can only be added to a new MMS group with the same members. Contacts using email address handles will use a phone number instead."

You are given two options: "Cancel" and "New Group".

If you choose New Group, you'll now have two groups. If you do nothing else, no one knows a new group was created, since no messages were received. If you send a message, its still entirely possible for the other group members to message either or both group chats. Chaos ensues.

It's not clear that Apple is actually going to fix this with RCS. Seems most likely they will not, that group chat splitting will still occur, just replacing SMS with RCS.

> i'm pointing to where i understand the landscape to be headed. for channel-based chat systems like Discord, irc, Matrix, XMPP/jabber, Slack, it's common enough to find channels which are bridged across 2 or more of those protocols.

Bridging is hacky, and involves not showing contact information for each user. You (of course), can't start a DM with such a user, and I'd assume things like @ mentions are ambiguous or nonfunctional.

Sure it _can_ be done, but it is kind of a terrible experience. Even Matrix and IRC have the same problem, and that's one I've actively experienced from both sides (IRC and Matrix).

> my experience with ephemeral group chats is that if i want to plan a large enough event i just end up starting multiple group chats, and the unimportant details are chaotic but the important ones like where/when we're meeting i make sure find their way into both chats.

I commend you, because you take a lot more effort than most humans to make sure things end up on both ends. In my experience, with the humans I have to deal with, its about a 5-10% of the time this happens, and usually its by sending a screenshot of the other group chat with half of the first line of the next message showing more important details that they decided "weren't relevant" or just didnt fit on the phone screen.

Also it should be obvious but some kinds of planning are simply not possible or require people to perform special courier roles to complete. Things like planning for what weekend everyone's free or what elements of a potluck everyone's going to bring are pretty tedious to manage between 2 group chats.

Furthermore, in my experience events that need planning aren't given dedicated ephemeral group chats, instead they are simply planned on whatever group chats they already have. People don't tend to put a lot of thought into making sure people are included, especially if the group chat is large. Some of the family group chats I'm in are 12-14 people. Not all of those people are coming to the potluck. They still use it, and honestly I think that's better than having to juggle every combination of every participant and keep track of whos in each one.

[1] Side note here, after Google started interpreting the (fairly annoying) iPhone tapback SMS messages as tapbacks, Apple introduced a similar feature to interpret tapback SMS messages --- but only for iPhone sent tapbacks. So the scenario is a group chat with 2 iPhone users in it-- the tapbacks show as SMS to the receiving iPhone, but it gets turned back into a tapback emoji reaction. This only works for iPhone style tapback SMS messages. The slightly different format that Google Messages sends is... ignored...

Pretty much the most smug Apple way they could possibly implement that feature... but now the Pixel in the chat works in all cases and the iPhone only works in half the cases, so it actually only hurts Apple users' experiences




Consider applying for YC's Fall 2025 batch! Applications are open till Aug 4

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

Search: