Will Anytwo still be for "Any" type of user? Thoughts & Questions

Talk Lounge
1

Well… I’m not really the type to comment much, especially on forums, but this time I need to share my thoughts and concerns, and I hope to get some feedback from the team on what’s next. With the recent news about developing completely from scratch, which will become Anytwo, I do believe it’s a good thing that will generally improve the development and the final product. What makes me uneasy is how things will unfold from here. I know the original software had its issues and everything, but how is it going to be now? As you mentioned in the Town Hall, it took 6-7 years of development to get to where it is today. I know a lot was learned during that time, but is this new development going to take a long time and evolve slowly?

Honestly, during the time I’ve been following the project, there has been a lot of back and forth, instability, issues, and radical changes to the core vision of the app. From what I understood initially, it was supposed to be a productivity system to help individuals in their day-to-day lives, connecting knowledge and acting as a second brain while adapting to other areas of life—an all-in-one tool (which always appeals to me). But as development went on, features like collaboration, chat, and business focus started popping up, and now AI is being put at the center of the app (I’m afraid of this just being another AI flop). I’m not saying I’m against that—on the contrary, I really like the idea of collective work and collaboration. The problem is that Anytype itself was lacking and immature in the basic foundations of a “superpowered” note-taking app (like it or not, its core is still note-taking, even with AI and collaborative features).

What I want to know is how Anytwo’s development will play out—will it still be geared towards the individual? Because in the Town Hall, you mentioned it would be business and enterprise-focused. What about the person who just wants to use it for logging things, connecting information, and using it for different situations like studies, work, leisure, finance, and various lifestyles, catering to each person’s unique needs? After all, if it’s going to be “any”, it has to be “any” “type”. I hope Anytwo keeps that same essence.

To wrap up, I have a few questions that I’d like the team (or someone from the team) to answer. For me, these are essential for adopting the app and supporting the product:

  • What are your plans for cross-platform support? Is this already being thought about? The current version of the Anytype app is very limited, for instance. I switch between different systems a lot, and the way it is today is a problem. Will large screens on Android (I mean tablets) have full desktop-like support, using a mouse and keyboard? Portability is crucial to me (and I believe to many others as well).
  • Will the application’s structure be based around communities, companies, and businesses? Will a single person who just wants to log their own information struggle with a system designed for a collective?
  • The beginning of the project will probably be a bit turbulent, but will things get more stable as it progresses? Like, without radical changes to the app’s core essence? Will the update cycle be faster and more evolutionary than what we have today?
  • Will it be possible to structure knowledge in a fluid, organic way? (Using links, connections between subjects and topics, macroscopic analysis of information, tags, and the like?)
  • What about the AI aspect? Ever since the AI boom with Big Tech, everything attached to this terminology has become exhausting, frustrating, and a total flop.

So, I think that’s it for starters. Since this radical change happened, I couldn’t help it—I had to comment and express my thoughts and questions about this new path.

A big hug to the team, and I’m rooting for the success of this new phase of Anytype.

2

Welcome in our community, @wandersonrms !
– Great first post, btw.!

Although I’m only a normal user, I’ll try to give answers to your questions, based on what I believe to “know”.

Oh, I’m VERY convinced, that the development will become dramatically faster!
Why?
– Because of mainly three changes:

  1. The new codes doesn’t have all the old relics from that the old code suffers.
    It’s a fresh start without the ballast.
    .
  2. As the team has already explained, the world has changed. Today we have AI coding.
    The old head of the devs team already used it, but AI coding has its limits, especially if you confront it with existing and highly tangled old code with lots of complicated dependencies, where changing one little part of the code may have unexpected influences on other parts.
    But AI can shine if you use it to write new code that’s free from such effects.
    .
  3. Although I don’t dare to say a single bad word against the old chief developer, I dare to say that it’s a good thing that he’s not involved in the new project.
    Why?
    – Because, “new brooms swipe clean”.
    Honestly, one becomes blind to the obvious if one works for too long on a certain thing. Also, one becomes reluctant against general changes if one has invested a lot of the own heart blood over years into self written code that somehow feels like an own child.
    Can you image what a programmer must feel who has worked day and night for years, if his boss tells him that he needs to give up his whole code and to start freshly?
    – Only a new programmer can do that!
    The old programmer would always have the tendency to reuse big parts of his old code.

Nope, that’s not true.

  1. You can use AnyTwo completely without any AI! That was communicated very clear and very often.
    .
  2. As @Kaye has already explained, AI is not only optional; it does not even have access to your data!
    Instead, the AI writes on your demand a small program for you that does the wanted things with your data.
    Therefore, the whole AI is completely encapsuatled and no data can leak into the internet or something like that.
    The principle is similar to the situation where you say to a friend (who can write code) that you need a certain routine from him that you describe. He does it for you without ever seeing with what data you feed that routine.
    For example:
    “Write me a code, that can sort the lines in a text file alphabetically”.
    The programmer will never know what you have in that text file(s). He simply writes the needed routine for you.
    The Ai is this programmer. It can’t see your data, it has no access to your data.
    Therefore, you can’t say that AnyTwo has AI at the center. It has only the software interface that an AI enables to react on the commands that the routines contain, that the AI writes for you (if you allow that).
    – The fact that your computer has an USB port doesn’t mean that you’re forced to use it. The port itself doesn’t do anything.
    Same with the software interface. It is simple there, not more. An AI can use that interface, if you allow it, but if you ignore the AI it can’t do anything.

Yes, it will still (also) be for individuals.
Understand, that business customers have some special needs that private persons not only don’t have, but even explicitly don’t want!
The given example was this:
A company has multiple workers that use Anytype. Normally, each user has his own password and no other person on planet earth can access his data without that password.
But if a worker leaves the company, things are different. Because the data belong to the company, not to the worker.
That means, the company must have a way to get access and even to expel the worker from the access.
There are some complicated details in such a situation that no normal user needs to know, it would only overwhelm him with unnecessary ballast. But a company needs such special functions.
This example is only one single example.
And the devs have decided that it’s better to make two “separate” products out of AnyTwo: one for normal users and one for business customers. But both share at least 99% of the same code.
As an alternative, they could also have decided to configure AnyTwo during the installation in this or that way.
But this could also cause headache for users, if they are suddenly confronted with the question, if they are a “normal user” or a “business customer”. That could trigger them to click deliberately the wrong button in the expectation that they then will get “something more”, although that wouldn’t be true.
So the team decided, to make two “separated” products, although they are mostly identical.

Here I can’t give a plausible answer.
I can only guess that the situation becomes better with the new and cleaner code.
As far as I know, the support for landscape mode is in work, or already finish.
I expect a lot from the cleaner and ballast-free code. It should improve the general portabllity.
– But that’s only my personal guesswork.

I can’t see a reason why a single person could struggle?
If you don’t need the chat, simply ignore it.
If you don’t want AI, simply ignore the fact that you could use it.
And if you are a single person, don’t install the business version, install the normal version.

As already explained, I foresee for good reasons heavy improvements.

From what I’ve seen in the Town Hall and from what I’ve understood, I’m absolutely convinced that you have only lots of benefits, but no losses in functionality!
The main part of the Town Hall was for me a single long "that is it!!! -moment!
All the things with that I and others struggle since years, became addressed.
AnyTwo will finally play out the benefits of a mesh structure and a database. Anytype on the other hand, never really used that potential, it did it only partially and even that only half-baked.
For example, we have never had bidirectional linking. We have never had real relations. We have never had Tags with more that one dimension. And so on, and so on.
AnyTwo solves all that, and so much more.

As written at start, I’m not a team member; my explanations are therefore not an official statement.
But I believe that I relatively good “got the message” during the Town Hall. I understood the idea behind that all.
What was shown there, reflected our accumulated feature requests from years!

What was shown was not only “some more features”. What I saw was finally a clear concept how to make everything better in the real practice.
– I’m absolutely convinced that this is the way to go!

3

Taking 6-7 years of development of Anytype does not mean it will take 6-7 years of development to get Anytwo to the same stage. Why?

  1. Anytwo still uses many things from Anytype such as any-sync, which is a crucial piece of infrastructure that’s already built which we’ve improved on over the years. This can be said for many other components too, such as any-store, files nodes, etc.
  2. Development is not just about code, it’s about learning and decision making. We can take many things from Anytype and just execute it on Anytwo.

We’ve already had many posts about feature parity, so there’s no need to go at more length.

There’s what’s being said, and there’s what’s being done. In both cases, we have not positioned Anytwo as being enterprise-focused. I think there’s quite a big a misinterpretation of what was expressed.

Anytype for Business was created explicitly to separate features that organisations require that we believe would compromise the experience for individuals. The Personal Anytype and Business Anytype share almost identical features today. The example we used was on authorisation and member management (holding keys), which is something organisations need a lot of parameters for that conflicts the sovereignty for individual usage. We had to build a separate node on our infrastructure entirely to separate how we handle this for orgs vs. personal.

If we wanted to abandon personal users and focus entirely on businesses, I don’t see why we would make such a decision to separate these components—we would just shoehorn all organisation features into the main product. Make no mistake, we certainly see value in businesses—but it doesn’t mean they’re the only customer we care about.

If you watch the demo, you can see what is possible in the app and make a decision for yourself. There was a journal mini app in there for example. Like with most products, if there are features that you don’t need, don’t use them. I personally don’t journal at all, but that feature existing doesn’t prohibit me from using the product. I also think that’s the case for the vast majority of collaboration features.

As mentioned during the town hall, at ~25% of the respondents to the community survey use collaborative features. That’s more than some device platforms. Saying we shouldn’t support collaborative features is like an iPhone user saying we shouldn’t support Android. If you don’t use it, no problem, but others do and it’s desirable for them.

We have mac, windows, linux, android, and iOS today—that is a lot of cross platform support already? I think you must be talking about something else. If you’re talking about landscape mode, tablet, etc. then of course this is something we want to support but it’s frankly not a priority given the state of affairs.

It should be clear: the more platforms you support, the more resources it costs, which means it slows development across the board because you are spread more thin. Users want faster development, but they want more platform support. You can see the predicament.

Words are no useful here. You can see from the demo what is possible already and the future of the product to make a decision for yourself if it suits your individual needs.

What individuals need greatly varies, it’s not one collective unified body of features. For example, not every individual wants to store their tasks in Anytype/Anytwo because they want a separate dedicated app. Not every individual journals. Not every individual builds photo galleries. Not every individual stores audio notes. If we want to build a tool for only individuals where you’re not inundated with features you don’t need, which to keep and remove? There is no such thing as one type of Anytype user, even if you try to confine it to ‘individual users’ only.

You can see from this framing that the whole individual use vs. collaborative use is not very helpful. For my task manager/to-do app, I personally do need collaboration features so my partner and I can stay coordinated on groceries, shopping, etc. For others, not.

This framing also doesn’t make sense for any software company. Think about Apple, how many radical changes have they made over the decades to their products regardless of their maturity level. For Anytwo, we of course believe that it’ll become less turbulent over time on average. But to promise from 2026+ we’d never have any radical changes to the core essence? Not possible to answer.

Yes our update cycle will be faster, and it already is. I don’t know what evolutionary is meant to mean?

I can’t answer the question because it’s quite vague. We have a lot of posts on Collections 2.0 and in the demo we’ve shown you how organisation can be organised. On average, you should not have less ways than your experience on Anytype today—it should be more.

There’s nothing more for me to add that would be different from what has already been said. If you find AI to be a total flop, you can not use the features. For us, it’s clear that many people who are well versed in AI find it incredibly helpful. Our role isn’t here to convince people to use AI, but we believe that AI is here to stay in the world—you can ignore it, but it’ll become increasingly hard for anybody to avoid it.

4

So this needs clarification, I will give it a very high level overview.

First, you choose if you want to use AI at all or not. If no, then the rest doesn’t matter.

Second, you choose what kind of model you want to use. If it’s a local model, then data doesn’t leave your device and nobody has access to it. If it’s self-hosted, then data leaves your device but to your server that you secure and own. If it’s a cloud model, then data leaves your device and you need to trust your provider.

Third, you choose the space and usage. In the case of it being a local/self-hosted model, it really doesn’t matter so much for most people.

Fourth, what you’re describing is our we’ve built our agentic system. In this case, the LLM does have access to your data, it’s just that it doesn’t run inference (and burn through tokens) every time it performs certain tasks. Instead, the LLM performs the task the first time through inference, builds a reuseable program that it can improve over time, and each time you want that task to be done, it uses the program. The solves not for privacy but primarily for:

  • Consistency of output. Using code to perform tasks, not LLM reasoning.
  • Better output. It learns from failures and can improve the code over time.
  • Reduction of LLM usage. If you use a cloud model, this means less burning of tokens.

In short, you have self-improving programs managed by your LLM model of choice in a privacy model that you choose. It’s a way to customise your exact usage needs (at least with automations) without relying on UI features.

5

Thanks for the reply! Just to clarify—in case I didn’t express myself well—I don’t see these points as mutually exclusive choices, like strictly collective vs. individual, or strictly with AI vs. without AI (I’m by no means against AI; what bothers me is how it’s implemented in certain situations. In Anytype’s case, it makes sense, and if done right, it really helps with development and productivity). My main question was whether you would take other factors heavily into account when designing the project, and overall, I understood that yes, you will. It won’t be a system niched down for a single audience, but will bring true versatility (and I really like that).

Regarding cross-platform support, the current Anytype does indeed support various systems (if not all of them). The real issue is how it’s structured for each operating system. I asked about Android or even iOS because many programs and apps originally built for desktop suffer from limitations and lose a lot of features when ported to other systems (like mobile devices), acting as a “lite” version for phones. In my view, the current Anytype mobile app feels like a “lite” version just to get by for anyone needing phone access (which is understandable since it’s still in beta, and as you mentioned, legacy code must be an issue for that kind of thing). Cross-platform should be full-featured and adaptable to the user’s workflow; often, you leave the desktop and need the same experience and productivity on a large-screen device (like a tablet or foldable) that offers high portability. Google, for example, wants to change the impression that Android is just for phones, especially with the release of the Googlebook.

I expanded on my comment to make my point clear. I’m not talking about restrictions or anything rigid, but rather trying to understand what’s being built from here on out. I understand that cross-platform optimization is something for further down the line, as the system needs a solid foundation first before exporting to other form factors.

I feel like this new version being developed will be very promising. I’ll most likely be right here following the updates closely and supporting the project.