Wavelength
daringfireball.net
daringfireball.net
The disclosure was made almost exactly in the middle of a long article. Not at the beginning.
Excuse me but gentlemen don't do that.
It's universal rule if you want people to actually read disclaimer. Similar method of burying the lead is used to promote quackery, info that's it's based on a online poll that five people took is "hidden" in later parts of article.
Put another way: do you really think there are people who read my article who downloaded and installed Wavelength based on the fact that I wrote about it, but who didn’t read that far into the article? And if they had seen the disclaimer, wouldn’t have downloaded it on the basis that they no longer trust my word about it? I don’t.
Jeff Atwood instead is remarkably compliant. I personally would have been instisting on „Common Markdown“. Especially with the reasoning, with John Gruber just not-replying until publishing and then being a child and applying double standards, when he himself caused the need for CommonMark due to the inconsistent description of Markdown’s grammar.
With regard to design, Wavelength's particular, native aesthetic might have a certain kind of emotional appeal to old-school Mac and iOS users (including me), but fundamentally its conversation view works a lot like Discord's, which also supports threads and basic replies. I don't think they're going to win on "quiet" design alone.
In fact, I think their focus on "nativeness" could inhibit feature competitiveness down the line. They'll need to make compromises one way or the other, either tolerating a greater sense of "cacophony" or forgoing features that other platforms support.
Their positioning and followthrough on privacy are strong.
Although Wavelength seems to be well-crafted, I think its success, like the success of many other network-oriented products, is going to largely come down to go-to-market. As an adviser and cheerleader, John gives them meaningful reach, but whether they can capitalize on that reach will depend on whether they can build a compelling enough hook to get people to move from wherever they're having conversations today.
It's not just old school. It's a "Mac-assed Mac app" [1]. Unfortunately even Apple themselves have stopped caring about making highly polished native Mac/iOS-assed apps. And that could be a huge differentiator if it wasn't just indie developers embracing this, but the platform itself.
> In fact, I think their focus on "nativeness" could inhibit feature competitiveness down the line.
Wat? Half-assed designs isn't "feature competitiveness".
[1] https://daringfireball.net/linked/2020/03/20/mac-assed-mac-a...
This new chat app is currently bait, but I'm not really lured by "threads", and the switch is a ticking timebomb.
Other red flags:
* Signup with phone number, other apps have proven that it's not universally necessary.
* Not mentioning that the author got equity of the app's company until halfway through the article.
* e2ee handwaving but there will never be proof unless they open-source it.
* Some integrated AI bot thing that nobody wants, could have been a bot API, stands at odds with e2ee, and just sounds like a lame attempt to ride the current wave.
* Mentions signal, whatsapp, telegram, but suspiciously not matrix.org, which AFAIK does have the touted feature to re-encrypt older messages to new members.This is a good point, and the Wavelength team is well aware of this.
> but suspiciously not matrix.org
I will admit I’d not heard of Matrix until your mention of it here. So I spent a few minutes downloading it, signing up, etc. It’s obviously competitive with Wavelength to some extent, but the first-class threading concept isn’t there. It seems more like a modern IRC or Slack alternative. Glad to be made aware of it though — thanks.
That being said I like what I see of Wavelength so far. I don't have anybody in my network that popped up right away, but I have a few people who are always interested in trying new chat apps so we'll give it a whirl.
What wavelength offers:
- Privacy focused (majority of users do not care about this, all modern chat apps have e2ee already)
- Natively runs on Apple devices (without Android/Windows version, good luck getting a large user base, especially with 2 developers)
- An AI chatbot (discord/slack already have this?)
- A groups -> threads -> messages layout. Like they mentioned, the same as slack but less corporate. Seems like a good idea since Slack is pretty awful for personal use.
The article could've done less without the author patting themselves on the back about how amazing their opinions offered were.
Instagram comes to mind: https://en.wikipedia.org/wiki/Timeline_of_Instagram
Surely you realize that picking one of the most successful apps created this past decade is not the resounding endorsement you think it is.
You must be new here
Of course, a paid-only model would put a hard ceiling on the total addressable market, but they have an interesting challenge ahead in finding something that a sustainable number of users will want to pay for.
But, iOS only/first is a nonstarter for a huge portion of the population. If they're just trying to experiment is a small pond first to get it right (sort of as a throwaway version), then that is excellent. If they're planning to stay in the Apple walled garden exclusively or as a first-class/peasants model, then, well, good luck with that.
Wavelength is currently available only for iPhone, iPad, and Mac. An Android app is planned (see next item) … As mentioned above, Wavelength’s development team is very small. Two people, Henry and Patterson. That puts a limit on how much they can accomplish — hence the lack of an Android app at the moment.
Regardless, I don’t personally have any interest in this App at this time.
And it's especially true early in the development of any product, when major decisions are still being made, and even more true when there are sudden bursts of uptake, which can impact infrastructure in unexpected ways, which can then affect architectural decisions about the app.
Trying to do all of that on two platforms simultaneously is almost certainly an unnecessary dilution of effort this early in the development of the product.
That doesn't make Android an afterthought. It just makes it next.
In terms of addressable market, Android isn’t a stumbling block to begin with.
By making Groups hierarchical they can even reinvent Forums.
MS Teams is groups > channels > threads > messages
But yea this does feel like forum arrangement. Which is really the most useful format for longer term conversations. Discord is great for “today and tomorrow matters yesterday doesn’t (yolo).
Team's threads vs messages are so similar users often don't understand them
No one has usenet style threads yet
I don't understand what could possibly drive anyone to making a new messaging app at this point. Even if your app has impenetrable privacy, impeccable ui, and amazingly useful features - it's not going to improve the horrible experience of wrangling a zillion of other messaging apps - and if your users have friends, they'll have to do that anyway. At this point, making another one of these feels like some kind of sadistic joke.
I wish to incinerate all messaging apps.
P.S.: to clarify - it's not like I even _want_ to use more than half of the messengers I listed. I am _forced to_, because some people I want to talk to are only available on certain platforms.
Similarly, if someone new can join and they can read what you wrote, it doesn't seem very private. Your audience is not only the people in the group, but a person to be named later. (Screenshots are always a possibility, but this is normalizing it.)
This goes double for having AI in the chat. The AI needs to read all the messages, so doesn't that mean the Wavelength needs a private key?
At that point, I'm not sure that end-to-end encryption matters. The chat might as well be stored on Wavelength's servers.
Only messages that actually mention @AI are sent in plaintext to us, and then to OpenAI's API. All content shared between people without invoking the AI is fully end-to-end encrypted. Our servers don't have access to any users' private keys -- those are generated on each client and never transmitted, except in the case when a user logs in on secondary device and uses our QR-code transfer flow (in which case the sensitive information is communicated peer-to-per in that QR code). We'll share much more on this when we publish our tech blog in the coming weeks.
Regarding privacy around strangers joining a group -- Wavelength supports locking down groups to the degree the group creator chooses. You can disable history syncing, so that new members don't get a copy of prior history (which, when enabled, is reencrypted and sent by prior members of the group) when they join. And groups only grow to a size where there would be people you don't know in them, if someone chose to share an invite link semi-publicly, or allow all members to invite new members (which is also a setting the creator can toggle). Personally, I'm in a number of quite large groups including people I don't know in real life, as well as plenty of groups with 2-10 friends I know in real life, where I obviously trust everybody not to screenshot/share what I'm writing, and I appreciate having strong e2ee.
If you keep controlling equity stakes in the business and don't need to raise significant outside investment it will work out as per your business model.
If you get into enterprise selling, you'll need a sales workforce and significant cost of Sales/Goods/Administration. If you want only organic growth, you'll be ok. But what will happen is that investment from, or selling to a large corporate, will turbo boost such sales; the temptation proves irresistible to many. If you are in a land-grab dynamics with other companies, growth may actually mean survival.
There is a scalability tier above the best subscription businesses, which is mass-market consumer adoption. The twitter experience demonstrates this. The ad revenue across billions outweighs a tidy subscription revenue from a small highly engaged subscription crowd. Any significant co-investor will find the lure of this irresistible. If you gain private equity investors, you will become ad based with high probability.
I welcome your contribution to the field of messaging apps and wish you well. I hope that this can inform your judgements about business models going forwards. If you think this is incorrect, please follow up with your alternative thoughts.
There are so many corporations that made similar pledges only to break them when it became advantageous for them to do so. Is there any mechanism that actually would enforce it? Mechanism that wouldn't disappear with a single stroke of pen?
Going for network effect with a free offering means there has to be investment money keeping the lights on. Those investors will want their returns.
I tested it, and indeed it was entirely wrong about how to force a provisioned concurrency AWS Lambda function to reload.
On a mobile device.
And called a messaging app.
But it's a forum.
This is an encrypted chat app where group chats are the main focus.
Forums have had private forums, public forums, paid forums, encrypted, shared, only trusted device networks, for everyone, federated.
My point is that this isn't new, and that's a good thing.
It resonates, "it clicked", because it's something we recognise.
When we recognise something then we can also tap into the deep history and learn from the possibilities of that and also reflect on how they failed at times (against social media).
I love forums, I would love to see forums become the predominant form of messaging again. Forums yield slower and richer conversations between groups as small as a few people or as large as hundreds of thousands.
Calling it what it is helps access the past and peer into the future.
If they could somehow fix that issue, that would be awesome.
The same number of messages are made, why arbitrarily force the discussion to be broken up? Embracing what the users want to do, how they want to us it, is a great way to discover non-obvious uses and to be low friction for them.
Just looked, and on one of the forums I run there's a thread with more than 130K replies https://www.lfgss.com/conversations/133015/ and so long as it loads fast and people are enjoying it, it's fine.
If there's a thread with 100 pages the title probably says little about the topic. There's probably some interesting comments in there, but I don't want to go through 100 pages...
It's good. I hope it becomes more a popular thing.
But yeah, article gives some glimpse of hope for communities.
This may be important for "threaded chat" but it's also a privacy killer. Everything you write might than be read by someone who later joins and wasn't the intended audience.
Only messages that actually mention @AI are sent in plaintext to us, and then to OpenAI's API. All content shared between people without invoking the AI is fully end-to-end encrypted.
Imagine a group chat on Signal, but in the middle of the chat dangles a big carrot enticing other group members to forward the discussion to OpenAI conveniently.
Sure, there's various other ways someone you talk to could do this silently. But ignorance is bliss vs. an inline feature to flaunt it.
All of the state of the art end-to-end encryption in products I know of is free software.
Much more difficult to verify these claims otherwise.
Furthermore, the app privacy label on the App Store claims that this service collects your contacts. So much for privacy, then.
Any conversation that includes OpenAI's GPT also necessarily can't be "completely private" or end to end encrypted as the service must have a plaintext of the messages to proxy them (in cleartext from their view) to their OpenAI APIs. It also then, you know, uploads your messages to OpenAI.
What's the point of e2ee messaging if the service provider and OpenAI can read it all the moment you loop in the headline feature (GPT in your groupchat)?
Ultimately, threaded e2ee groupchat is a feature (ideally a feature added to Signal), not an app/company.
FWIW, it’s up to the user whether to allow that. I don’t recall the language the app uses to describe the contact request, it seemed reasonable but I didn’t allow it.
And it’s a tiny amount of work if you just want to know which users have a contact with one specific phone number.
Much better would be to store no contact information at all server-side.
But anyway I'm not looking for the next chat app anymore.. there's already too many. I bridge everything to matrix now to have everything in one place. WhatsApp, signal, telegram, discord, slack etc. I don't want to use this new thing until a bridge becomes available to it and some of my contacts prefer it as their main contact method. It sounds like this is a long way away.
> “something like Twitter but with real names and enforced civility, and with hashtags as Slack-like channels of interest.”
but why is having real names on a chat app a good thing? Considering what's happening worldwide, anonymity online should always come first.
Most social networks currently don't allow reporting low-level offenses so it's either a major violation or just ignored entirely. I would greatly prefer Reddit and Twitter if I could cut out large chunks of low quality posters.
- this obnoxious user, anywhere
- this obnoxious user, in this thread/group
- this obnoxious user, using any of these keywords
- any user using these keywords in this thread/group/anywhere
Any social media system that doesn't offer these needs a much more expensive moderation system.
Arguably everything is E2E encrypted but that won't prevent someone in the discussion from taking a screenshot (thus removing context) and tweeting it, will it?
Considering the amount of politically-incorrect, politically-sensitive etc. stuff (ukraine-russia war, anti government protests, recently anti-covid protests, basic software/... piracy, ...), having anonimity is a huge feature.
EDIT: I thought his "it's AI in a group chat" pitch sounded familiar, then I realized it's because it's a Discord blog post from a couple weeks ago: https://discord.com/blog/ai-on-discord-your-place-for-ai-wit...
a) better threading UX
b) E2E encryption
admittedly I'm a very casual Discord user, and haven't used this app at all yet, so I may be missing something.
Also, sounds like the Android app is coming, so I don't think "don't want to talk to Android users" is an intended sales point?
Something that has a UX more like traditional forums as a chat app sounds neat to me.
I joined a couple of servers related to games I play and some sports. But I feel like there's conversations going on that I can't see. Do I need to message some special bot? Do I need an invite or be admitted before I can see some hidden channels? There are voice chats, there are so many settings.
I definitely feel like I need someone to handhold and clue me in to a few more things about it before I feel "comfortable" using it. Discord is a lot to deal with.
https://wavelength.app/invite/BHRaKD7Lffh#rjZY5RlPZmg_l7UhDP...
Edit: Oops, there is an earlier invite to a HN group. Should have used the find command before posting...
What happened to "if you're not paying, you're the product"?
To register, you need to enter your phone number.
Even telegram now allows creating an account without a phone number, why is this still a requirement?
Thanks, but no thanks.
[downvote all you want, but we're complaining about facebook and others, while replacing them with a new, shiny social network, that has the same pitfalls, just because Gruber recommends it - we never learn]
WhatsApp and Telegram both got “rooms” inside groups (“communities” and “topics”, respectively) to organize groups around topics/subjects, pretty much like Slack and Discord.
Hard to justify a whole new app just to mimic features all other apps, with people already using them, have or will eventually have.
> Me: @AI What weighs more: a pound of flesh or a Great British Pound?
> AI: They both weigh the same amount, one pound.
> Powered by OpenAI GPT-3.5. Be careful, the bot may be completely wrong. Only messages that mention @AI are visible to the bot in this group.
This kind of reasoning is easy for GPT-4.
It's great to see messengers incorporate AI.
What's your moat here against a GPT Telegram bot?
The one point in the article I disagree with. IMO the best chat interfaces are in desktop web apps. So many things are easy to do in a web chat app that are a pain to do in native apps. And I'll take typing on keyboard over a touchscreen any day of the week.
Is what Zulip does. And it works great. Even for large orgs.
threads become like subgroups I guess, with names, changing member lists - and ... ? retention policy? delete permissions? ai summaries?
App Store search is broken, they are doing the Google thing where they want you to buy ads so people find your app when you type its name in the search box.
It requires iOS 15 which cannot (currently) be jailbroken.