I discovered Anki 12 years ago while living in Japan. I was trying my hardest and absolutely failing to remember any of the Japanese I was studying. Maybe I was due for a learning-style renaissance for myself and Anki was just the catalyst, but it really made a positive impact on my life. More than just memorizing kanji on AnkiDroid during my commute, I just started to believe I could learn anything. I was starting to take my coding hobby more seriously at the time and hacking on Anki was a big part of that too. Thanks for all the hard work Damien and David Allison. I'm so grateful for the software you've worked on.
True. It should however be noted that the most active maintainer of AnkiDroid will be joining the new entity:
> We’re currently talking to David Allison, a long-time core contributor to AnkiDroid, about working together on exactly these questions. His experience with AnkiDroid’s collaborative development is invaluable, and we’re grateful he’s willing to help us get this right. We’re incredibly excited to have him join us full-time to help propel Anki into the future.
anki has so much potential and has such a big and unique audience, it is incomprehensible to me how it has managed to be so neglected.
and then now why, of all times, when a solo developer is never more productive, would the lead maintainer cede ownership? the antidote for programming burnout has just been invented, just take it haha
My experience using AI is that it wildly increases burnout, not decreases it.
Writing code is fun. Solving interesting problems is fun.
Debugging deep problems is fun.
Debugging slop code is a painful suffering experience, having to constantly double check that the AI agent didn't just change the unit tests to "return true" and lie to you is tiring, and the feeling that you can't significantly improve the tool burns me out hard.
That last one can't be overstated. When I find a weird behavior that looks like a bug in the linux kernel or rustc or such, I find it exhilarating to read code and understand what the bug is, how it got there, and to feel like I can fix it and never see it again.
When claude code gives me a "wrong" output for my prompt, I don't feel like there's any possible way I can go and find what part of the Opus 4.5 model resulted in it not being able to give better output.
I feel helpless to debug what went wrong when claude code spirals into the deep end.
I can add more initial context, add skills, but those are tiny heuristic tweaks around the giant mass of incomprehensible weights and biases that no human understands.
The antidote for programming burnout is not to replace all the fun parts of programming with painful probabilistic suffering.
iOS one is fine, pretty good. I use it daily too. Ankidroid is much better, which I would attribute to being open source with lots of eyeballs on it and making improvements for the love of it.
What’s so bad about the paid iOS client? I remember it being expensive when I got it but it works fine for my use case (mix of getting me through part of med school, all of law school, and the just general shit I’d like to remember and learn). There’s definitely never been anything jarring about using it vs the Mac or windows clients but I’m happy for somebody to point out the problems I’ve been missing!
Just have them use it on their computer or the web?
It improved my grades so much in college that I spent the 25 bucks as a broke student so I could have it on my second hand iPad. This was before AnkiDroid even existed so it's amazing the price is still the same.
It was a fascinating symbiotic between nerdy med students from all over the world and an obscure open source flashcard app that originally targeted language learners. I've been part of that community for many years and would have never foreseen this outcome but in hindsight it seems the best path forward for anki.
If one is of a certain age, of course they studied for step I without it
and the classic method was the inspiration for Anki to begin with: making your own flashcards on index cards! You could do a version of spaced repetition by shuffling the deck.
Not sure the digital version is actually easier or more effective
Many community-oriented programs have failed after acquisition because they came out too firm, too decided, and too purposeful, only to realize the community is still skeptical and turning against them six months in.
Honestly, for a program like Anki, starting out by saying "we need to figure out what good governance looks like, as well as what might be agreeable and possible for everyone involved" is a much stronger positioning than coming up with something that may or may not fly to try make a strong first impression. Communities do not follow the conventional rules of American business.
From the posts, it sounds like the original maintainer was approaching the point where they'd just abandon it, so this overall seems like a better outcome than either abandonment or sale to a PE firm.
Community focused organizations like this are hard to run without governance transitions. I think Anki brings value to the world and anyone willing to take on a leadership role in keeping it going should be given trust and grace to make the best decisions they can with the knowledge they have. I wish them luck.
It seems like the core things that Anki needs are new user experience improvements, and algorithm updates. SM2 really shows its age as compared to other algorithms.
I'm an Anki user, on and off since 10 years or so, but was still confused. If I understood correctly, the entities here are:
- Anki, as set up by dae aka Damien, is like the brand name and desktop implementation with the spaced repetition algorithm
- AnkiWeb is what I thought this hub thing was. It's where you download decks
- AnkiHub is a third party (started by "AnKing", now 35 employees) who sells decks as a monthly subscription and has their content on the deep web (you need to create an account and agree to terms to even see a listing of what's there besides a few featured parts). This is who is getting ownership of the former two. Because they write that Anki will remain open source at its "core", I presume that means that things will, at best, stay stable rather than anything (like AnkiWeb the deck sharing platform) becoming open
- AnkiDroid is a separate open source project (an Android app). The corporation is hiring the main developer, but it's not yet clear to me whether they're just going to get paid to work more on AnkiDroid or if they're also getting other tasks
> - AnkiDroid is a separate open source project (an Android app). The corporation is hiring the main developer, but it's not yet clear to me whether they're just going to get paid to work more on AnkiDroid or if they're also getting other tasks
----
To copy from my message on Discord:
> I’m moving to a full-time position working on Anki [incl. AnkiWeb & AnkiMobile]. I’m really excited about this, but there’s a mountain of pending, somewhat undefined work which will need to be done, and it’ll need my full-time attention for a while.
> I’ll still be contributing to AnkiDroid, but I won’t be able to commit as much time as I am doing currently (at least for the first few months while things stabilize). I’ll be here on evenings/weekends, and will be contributing in other ways (hopefully: unified Note Editor, JS addons etc… ), but I expect to slow down with code contributions to ensure I’m staying on on top of PR reviews & general force multiplier work. I’m definitely Org Admin’ing for GSoC over the summer [assuming Google gives us the greenlight], it’s historically been a VERY light role.
> In all honesty: I’m expecting things to be business as usual, I have more than enough capacity to keep up with the notification queue. Even if I completely dropped off the planet, we’re a great team and the improvements would keep on flowing. AnkiDroid’s bus factor has been >>> 1 for a LONG time now.
Worth noting you don't need to use it. Anki comes with a syncserver implementation for a while now, and there are docker images too. It's worth it for the transfer speeds alone IMO.
Anki is under AGPL too, which has an anti-DRM clause, so many type of enshittification of anki or their addons (e.g. to prevent sharing of their decks) would be unenforceable too.
As such I see no obvious things that would be susceptible to enshittification here.
(Same person as above but felt that this part had a separate purpose so I've moved it into its own comment)
The ecosystem is currently such that it seems hard to enshittify it. They say they have no intention of doing that and I believe it, but their vision of a healthy and good product might involve a fair price (for rich countries at least) whereas it was always free so far
Time will tell; it sounds like there's currently no plans either way, but it's also simply open enough that users can always just install the open source software and share decks with each other by whatever file transfer/sharing means. Everything that's already there won't simply go away. I'm going to keep using AnkiDroid and building the language deck I am working on
This may be true, but as someone who picked up Anki as a desktop app back around 2009 it feels a little crazy.
I also can’t imagine making cards on a phone, given how much switching between apps/windows is involved and how poor mobile platforms are at multitasking. It’s difficult to envision it being anything but maddening.
In America perhaps. Android is more popular in other countries, most people I know use Anki for free. The desktop app and sync are useful for editing cards and managing a large collection. Both of those are free too, but for how long?
Mochi has really come a long way and has some wonderful built-in functionality. The automatic text-to-speech template options are high quality and randomize a variety of voices in each language. It's also doing some cool things with dynamic cards and LLMs.
Impressively, Mochi now offers FSRS (beta but still available in the app's main settings) and both the type of scheduler (Mochi default or FSRS) and the schedulers' settings are configurable on a deck-by-deck basis.
The developer is very responsive to folks on the forum and often quickly adds requested features.
Overall the app is well-designed and fun to use. I appreciate the swipe left/right to fail/pass cards on iOS. My one complaint is that the web clipper only works with Chrome and Firefox, but not with Safari (surprisingly). It would also be useful to have a global hotkey/palette to quick-add cards to various decks, similar to how OmniFocus lets you quickly add items to the OF Inbox.
I've been paying for the pro version for a while. It's templating is really powerful and easy to use. For my vocab deck, I set up a input field (e.g., word) and a bunch of derived fields (dict definition, AI-generated example, TTS audio). To add a new card, I just input the word and other fields will be automatically populated.
Technically this can be implemented in Anki as an addon. But only the desktop version supports addons and the default UI is a bit too complicated.
> Absolutely. Anki’s core code will remain open source, guided by the same principles that have guided the project from the beginning.
Already caveating with the "core" code. Even without PE and VC, it's clear that a company with 35 employees is bound to take this in a different direction than 1 guy, and not a good one. If there comes a day where those 35 employees can't be sustained anymore by revenue, and the choice is between enshittification and shutting down/firing everyone, we'll see what happens. That's the big difference - such a decision was never on the cards, or at least much less likely, when run by a single person. Now it will be.
Big conflict of interests too. AnkiHub makes money from selling paid addons. No chance any of those will ever end up in Anki now.
Also not a good look that they immediately locked the thread in their most popular community.
At a fundamental level the algorithms predict the probability of a learner to correctly recollect a factoid at a given point in time given a history of sampling that recollection / presentation.
It would be interesting to have machine learning predict these probability evolutions instead. Simply recollecting tangential knowledge improves the recollection of a non-sampled factoid, which is hard to model in a strict sense, or perhaps easy for (undiscovered) dedicated analytic models. Having good performing but relatively opaque (high parameter counts) ML models could be helpful because we can treat the high parameter count ML model as surrogate humans for memory recollection experiments and try to find low parameter count models (analytic or ML) that adequately distill the learning patterns, without having to do costly human-hour experiments on actual human brains.
FSRS just works, even without a GPU so it's not the cool kind of AI / machine learning these days.
No joke though: the FSRS model is marvelous, and Anki remains one of the best free + open source implementations around.
I've been learning German recently and Anki (in FSRS mode) is one of the most important learning tools I have. No joke.
------
Every card remembers every rating you give it, as well as the time / date. This allows for Anki to solve for a 'forgetting curve', and predict when different cards have a chance to be forgotten.
There is furthermore the machine learning / stochastic descent algorithm to better fit the assumed forgetting curves to your historical performance. This is the FSRS Optimize parameters button in the settings panel.
In a nice and controlled manner, so seemingly no reason to panic just yet:
> I ended up suggesting to them that we look into gradually transitioning business operations and open source stewardship over, with provisions in place to ensure that Anki remains open source and true to the principles I’ve run it by all these years.
> This is a step back for me rather than a goodbye - I will still be involved with the project, albeit at a more sustainable level.
From AnkiHub:
> No enshittification. We’ve seen what happens when VC-backed companies acquire beloved tools. That’s not what this is. There are no investors involved, and we’re not here to extract value from something the community built together. Building in the right safeguards and processes to handle pressure without stifling necessary improvements is something we’re actively considering.
Relieved at that part where they say there are no investors involved, makes the whole thing a whole lot less risky. Good for everyone involved, and here's to many more years with Anki :)
Might as well give a recommendation then: I've been using hashcards [0] for a few weeks now and have enjoyed its simplicity and the fact that it all stays forever in raw markdown files and versioned git. A simple justfile has also been helpful.
Well every company claims no enshittification when they get acquired, in the end that's rarely the case. It's like Private Equity buying a company out and say "Nothing will change"
AnkiHub people seem kind of slimy in my experience (at least their leader, "The AnKing"). I hope they figure out a good leadership situation, and make stronger commitments to openness.
Anki is in a very solid position to be forked if anything happens, so even if this is bad news I have faith in the larger community.
I feel like there's a lot of confusion, so not sure if anyone can clear up mine, but I'm really struggling to see a significance of this.
Obviously, like all ignorant people do, I am going to oversimplify things here. But still, to me, the "platonic idea" of Anki seems a dead simple thing. All what I care about when using Anki is what's on the 2 sides of a Card, a question + answer, which can only be some visual image (possibly encoded as text, possibly just JPEG, I really don't care as long as it fits in my mobile device memory) + optional sound. That's really it. If it should be bi-directional or uni-directional card is a detail of how the deck is generated/encoded, and the spaced repetition algorithm is a detail of the app that I use to study (so, usually AnkiDroid, I imagine — an unaffiliated 3rd party; who even uses desktop apps nowadays?).
So, I imagine there can exist (and do exist) some minor additional features, like an ability to require a typed answer for a card, but it seems pretty minor, and I really don't see a lot of room for the app to evolve.
So, ultimately people need only a common .apkg format, which exists and is relatively simple (although I suppose it could've been even simplier), and a place like AnkiWeb, where people can share their decks, so Spanish top-2000 or basic integrals deck isn't re-invented over and over again. It's a pity that AnkiWeb isn't more open and will be even less open from now on, but as long as someone is willing to just host it (which is ultimately just paying for downloads traffic) it's easy to replicate, so no super-valuable IP here.
Of course, a primary use-case for Anki is a tool to make decks, but you could really do with pretty simple python script + YAML/JSON/CSV/whatever metadata file to convert it to AnkiDroid-compatible .apkg file.
IMO, the add-on value is the repository of decks that exist (and may or may not be free).
So an app-store of sorts.
As others have said, there are some provisions in place that make it allegedly harder to do a hard landgrab and keep people from freely sharing decks, to to me, even if it were so, I would not be too concerned.
In my opinion, the very act of creating one's deck is a key part of the learning. Maybe it's different for larning vocabulary, but as you said, it will be very hard to make those hard to share.
Learning a deck generated by someone else has never been as effective with me, so I think it's a false sense of time saving to use those.
As someone who has used Anki for a decade, a thank you to dae for everything. Best of luck in your future endeavours.
With that out of the way, some thoughts:
- Anki is in a really good position to work around enshitification. The app, at least to me, is "complete" - the only additional features that might pique my curiosity is a different scheduler (at the moment, they're integrating a newer one, although I don't follow enough to know the state of it). Additionally, modern Anki is really well architected: the core of it is a Rust library, that is used by all of the platforms [0]. You can write new front ends using that, or just fork the existing FOSS ones. Maybe dae does a gorhill and gives us Anki Origin.
- Really the only service-y part of Anki I use is AnkiWeb, which is basically a backup and sync system. Wonder how that'll evolve (if they do end up charging for it, I hope it is "Obsidian" reasonable). EDIT: Ooo, Anki has public server software for running your own version. Awesome! [1]
- The idea outcome in my opinion would have been some form of charitable organisation (Linux Foundation?), with people donating to support Anki.
- So, AnkiHub is a company that produces Anki flashcards, and they've scaled that quickly? Jeez. Obviously Quizlet proved there was a market for flashcards, but I didn't realise this was possible for Anki.
- No outside investment is... hopeful. Not quite sure what indicates that this company has the technical know-how to maintain it.
- I've heard too many stories of a maintainer or creative being "hopeful" about their new acquirers, only to regret it years down the line.
Can't help but wonder if it might not be time to start a FOSS alternative just in case Anki begins to decline.
It might not be the worst idea to do that anyway. Anki is great, but there's plenty of room for improvement. Off the top of my head, an architecture that doesn't involve fragile and finicky python bits and is designed to support multiple independent clients would be a nice step up (Telegram is a good model here — make a core lib with all the nuts and bolts which devs build clients around).
I discovered Anki 12 years ago while living in Japan. I was trying my hardest and absolutely failing to remember any of the Japanese I was studying. Maybe I was due for a learning-style renaissance for myself and Anki was just the catalyst, but it really made a positive impact on my life. More than just memorizing kanji on AnkiDroid during my commute, I just started to believe I could learn anything. I was starting to take my coding hobby more seriously at the time and hacking on Anki was a big part of that too. Thanks for all the hard work Damien and David Allison. I'm so grateful for the software you've worked on.
On the plus side, the actually good mobile Anki client, AnkiDroid, remains out of the hands of this potentially questionable new entity.
(AnkiDroid has always been run independently, which is good, considering the state of the iOS client, which has always been neglected.)
True. It should however be noted that the most active maintainer of AnkiDroid will be joining the new entity:
> We’re currently talking to David Allison, a long-time core contributor to AnkiDroid, about working together on exactly these questions. His experience with AnkiDroid’s collaborative development is invaluable, and we’re grateful he’s willing to help us get this right. We’re incredibly excited to have him join us full-time to help propel Anki into the future.
anki has so much potential and has such a big and unique audience, it is incomprehensible to me how it has managed to be so neglected.
and then now why, of all times, when a solo developer is never more productive, would the lead maintainer cede ownership? the antidote for programming burnout has just been invented, just take it haha
My experience using AI is that it wildly increases burnout, not decreases it.
Writing code is fun. Solving interesting problems is fun.
Debugging deep problems is fun.
Debugging slop code is a painful suffering experience, having to constantly double check that the AI agent didn't just change the unit tests to "return true" and lie to you is tiring, and the feeling that you can't significantly improve the tool burns me out hard.
That last one can't be overstated. When I find a weird behavior that looks like a bug in the linux kernel or rustc or such, I find it exhilarating to read code and understand what the bug is, how it got there, and to feel like I can fix it and never see it again.
When claude code gives me a "wrong" output for my prompt, I don't feel like there's any possible way I can go and find what part of the Opus 4.5 model resulted in it not being able to give better output.
I feel helpless to debug what went wrong when claude code spirals into the deep end.
I can add more initial context, add skills, but those are tiny heuristic tweaks around the giant mass of incomprehensible weights and biases that no human understands.
The antidote for programming burnout is not to replace all the fun parts of programming with painful probabilistic suffering.
I use the official iOS client everyday. What’s wrong with it?
iOS one is fine, pretty good. I use it daily too. Ankidroid is much better, which I would attribute to being open source with lots of eyeballs on it and making improvements for the love of it.
What’s so bad about the paid iOS client? I remember it being expensive when I got it but it works fine for my use case (mix of getting me through part of med school, all of law school, and the just general shit I’d like to remember and learn). There’s definitely never been anything jarring about using it vs the Mac or windows clients but I’m happy for somebody to point out the problems I’ve been missing!
The (paid!) iOS client has always been a disappointment to me, and I've long been jealous of the open source Android one.
I don't mind so much that it's paid, given how much use I get for the price, but it sucks knowing it sucks and not being able to help make it better.
I've just bought it to support the developer, as it was according to his website his preferred way to support him.
Is it still 25$ price? Makes it impossible to recommend Anki to friends/students to "try spaced repetition".
Just have them use it on their computer or the web?
It improved my grades so much in college that I spent the 25 bucks as a broke student so I could have it on my second hand iPad. This was before AnkiDroid even existed so it's amazing the price is still the same.
The web client works just fine!
Agreed. I’m particularly excited that they’ll be investing in the UI/UX.
It was a fascinating symbiotic between nerdy med students from all over the world and an obscure open source flashcard app that originally targeted language learners. I've been part of that community for many years and would have never foreseen this outcome but in hindsight it seems the best path forward for anki.
I don't think it's just for nerdy med students nowadays. Who studied for Step I without it? And How? (And Why? :)
If one is of a certain age, of course they studied for step I without it
and the classic method was the inspiration for Anki to begin with: making your own flashcards on index cards! You could do a version of spaced repetition by shuffling the deck.
Not sure the digital version is actually easier or more effective
Step 1 is pass/fail now. If I had to redo it and just pass, I don't think it would be necessary to use anki now (except maybe for something Sketchy).
Amboss
Name checks out
> What We Don’t Know Yet
> Governance and decision-making: How decisions are made, who has final say, and how the community is heard
> Roadmap and priorities: What gets built when and how to balance competing needs
> The transition itself: How to bring in more support without disrupting what already works
In other words: they have no clue what to do next (https://forums.ankiweb.net/t/ankis-growing-up/68610/2#p-1905...)
Many community-oriented programs have failed after acquisition because they came out too firm, too decided, and too purposeful, only to realize the community is still skeptical and turning against them six months in.
Honestly, for a program like Anki, starting out by saying "we need to figure out what good governance looks like, as well as what might be agreeable and possible for everyone involved" is a much stronger positioning than coming up with something that may or may not fly to try make a strong first impression. Communities do not follow the conventional rules of American business.
From the posts, it sounds like the original maintainer was approaching the point where they'd just abandon it, so this overall seems like a better outcome than either abandonment or sale to a PE firm.
Knowing what they don't know puts them ahead of many organizations.
Community focused organizations like this are hard to run without governance transitions. I think Anki brings value to the world and anyone willing to take on a leadership role in keeping it going should be given trust and grace to make the best decisions they can with the knowledge they have. I wish them luck.
The community has been in a deadlock over making FSRS the default (https://github.com/ankitects/anki/issues/3616), and I wonder if this will lead to some resolution.
It seems like the core things that Anki needs are new user experience improvements, and algorithm updates. SM2 really shows its age as compared to other algorithms.
Is there that much of a difference?
As a user, it’s a HUGE difference. FSRS leads to an incredibly reduced workload.
I'm an Anki user, on and off since 10 years or so, but was still confused. If I understood correctly, the entities here are:
- Anki, as set up by dae aka Damien, is like the brand name and desktop implementation with the spaced repetition algorithm
- AnkiWeb is what I thought this hub thing was. It's where you download decks
- AnkiHub is a third party (started by "AnKing", now 35 employees) who sells decks as a monthly subscription and has their content on the deep web (you need to create an account and agree to terms to even see a listing of what's there besides a few featured parts). This is who is getting ownership of the former two. Because they write that Anki will remain open source at its "core", I presume that means that things will, at best, stay stable rather than anything (like AnkiWeb the deck sharing platform) becoming open
- AnkiDroid is a separate open source project (an Android app). The corporation is hiring the main developer, but it's not yet clear to me whether they're just going to get paid to work more on AnkiDroid or if they're also getting other tasks
> - AnkiDroid is a separate open source project (an Android app). The corporation is hiring the main developer, but it's not yet clear to me whether they're just going to get paid to work more on AnkiDroid or if they're also getting other tasks
----
To copy from my message on Discord:
> I’m moving to a full-time position working on Anki [incl. AnkiWeb & AnkiMobile]. I’m really excited about this, but there’s a mountain of pending, somewhat undefined work which will need to be done, and it’ll need my full-time attention for a while.
> I’ll still be contributing to AnkiDroid, but I won’t be able to commit as much time as I am doing currently (at least for the first few months while things stabilize). I’ll be here on evenings/weekends, and will be contributing in other ways (hopefully: unified Note Editor, JS addons etc… ), but I expect to slow down with code contributions to ensure I’m staying on on top of PR reviews & general force multiplier work. I’m definitely Org Admin’ing for GSoC over the summer [assuming Google gives us the greenlight], it’s historically been a VERY light role.
> In all honesty: I’m expecting things to be business as usual, I have more than enough capacity to keep up with the notification queue. Even if I completely dropped off the planet, we’re a great team and the improvements would keep on flowing. AnkiDroid’s bus factor has been >>> 1 for a LONG time now.
https://discord.gg/qjzcRTx => https://discord.com/channels/368267295601983490/701922522836...
GSoC: https://summerofcode.withgoogle.com/
> AnkiWeb
Worth noting you don't need to use it. Anki comes with a syncserver implementation for a while now, and there are docker images too. It's worth it for the transfer speeds alone IMO.
Anki is under AGPL too, which has an anti-DRM clause, so many type of enshittification of anki or their addons (e.g. to prevent sharing of their decks) would be unenforceable too.
As such I see no obvious things that would be susceptible to enshittification here.
Aha [0], that is neat.
[0]: https://docs.ankiweb.net/sync-server.html
I've tried several times before to install syncserver using those pip instructions, on multiple platforms, without success.
Try docker: https://github.com/ankitects/anki/tree/main/docs/syncserver#...
(Same person as above but felt that this part had a separate purpose so I've moved it into its own comment)
The ecosystem is currently such that it seems hard to enshittify it. They say they have no intention of doing that and I believe it, but their vision of a healthy and good product might involve a fair price (for rich countries at least) whereas it was always free so far
Time will tell; it sounds like there's currently no plans either way, but it's also simply open enough that users can always just install the open source software and share decks with each other by whatever file transfer/sharing means. Everything that's already there won't simply go away. I'm going to keep using AnkiDroid and building the language deck I am working on
The iOS app has never been free and that's the way most people use it these days. Desktop computing is a niche.
This may be true, but as someone who picked up Anki as a desktop app back around 2009 it feels a little crazy.
I also can’t imagine making cards on a phone, given how much switching between apps/windows is involved and how poor mobile platforms are at multitasking. It’s difficult to envision it being anything but maddening.
Desktop for creating cards, mobile for reviewing them.
I prefer a laptop for reviewing because it’s still portable, but also more amenable to comfort for longer sessions and makes spot corrections easier.
Who is this
In America perhaps. Android is more popular in other countries, most people I know use Anki for free. The desktop app and sync are useful for editing cards and managing a large collection. Both of those are free too, but for how long?
Worth mentioning too is the FSRS algorithm for scheduling cards is implemented in separate libraries which are released under MIT license.
This sounds concerning. Someone ought to back up the public AnkiWeb decks while we still can.
Good on him, 19 years is a long time to carry the flame. Thanks for getting me through school!
Possible alternative to check out (not affiliated):
https://mochi.cards/
Mochi has really come a long way and has some wonderful built-in functionality. The automatic text-to-speech template options are high quality and randomize a variety of voices in each language. It's also doing some cool things with dynamic cards and LLMs.
Impressively, Mochi now offers FSRS (beta but still available in the app's main settings) and both the type of scheduler (Mochi default or FSRS) and the schedulers' settings are configurable on a deck-by-deck basis.
The developer is very responsive to folks on the forum and often quickly adds requested features.
Overall the app is well-designed and fun to use. I appreciate the swipe left/right to fail/pass cards on iOS. My one complaint is that the web clipper only works with Chrome and Firefox, but not with Safari (surprisingly). It would also be useful to have a global hotkey/palette to quick-add cards to various decks, similar to how OmniFocus lets you quickly add items to the OF Inbox.
I've been paying for the pro version for a while. It's templating is really powerful and easy to use. For my vocab deck, I set up a input field (e.g., word) and a bunch of derived fields (dict definition, AI-generated example, TTS audio). To add a new card, I just input the word and other fields will be automatically populated.
Technically this can be implemented in Anki as an addon. But only the desktop version supports addons and the default UI is a bit too complicated.
Looks nice but being an Electron-based app is a huge dealbreaker for me.
I imagine it's for Anki users and Anki already embeds a HTML engine (for the cards).
Andrew (the CTO of AnkiHub) is smart as a whip and has built an impressive and seemingly sustainable business. Seems like the ideal steward for Anki!
I love anki.
But upon reading this I think it's high time I exported all my notes in simple text format, just in case.
Maybe also try Fernando Borretti's flashcard app I saw (and dismissed) recently here
Anki uses a SQLite database for all its data, so it's already an open format.
> Absolutely. Anki’s core code will remain open source, guided by the same principles that have guided the project from the beginning.
Already caveating with the "core" code. Even without PE and VC, it's clear that a company with 35 employees is bound to take this in a different direction than 1 guy, and not a good one. If there comes a day where those 35 employees can't be sustained anymore by revenue, and the choice is between enshittification and shutting down/firing everyone, we'll see what happens. That's the big difference - such a decision was never on the cards, or at least much less likely, when run by a single person. Now it will be.
Big conflict of interests too. AnkiHub makes money from selling paid addons. No chance any of those will ever end up in Anki now.
Also not a good look that they immediately locked the thread in their most popular community.
I've used Flashcards Deluxe for years and been very happy (no affiliation):
https://orangeorapple.com/flashcards/
Easy to import and export my cards, plenty of options for tweaking the algorithms for my use.
At a fundamental level the algorithms predict the probability of a learner to correctly recollect a factoid at a given point in time given a history of sampling that recollection / presentation.
It would be interesting to have machine learning predict these probability evolutions instead. Simply recollecting tangential knowledge improves the recollection of a non-sampled factoid, which is hard to model in a strict sense, or perhaps easy for (undiscovered) dedicated analytic models. Having good performing but relatively opaque (high parameter counts) ML models could be helpful because we can treat the high parameter count ML model as surrogate humans for memory recollection experiments and try to find low parameter count models (analytic or ML) that adequately distill the learning patterns, without having to do costly human-hour experiments on actual human brains.
This is being actively researched (in the open!). https://github.com/open-spaced-repetition/srs-benchmark
Isn't FSRS (the new algorithm used in Anki since a few years ago) already based on machine learning?
Too old school and too effective.
FSRS just works, even without a GPU so it's not the cool kind of AI / machine learning these days.
No joke though: the FSRS model is marvelous, and Anki remains one of the best free + open source implementations around.
I've been learning German recently and Anki (in FSRS mode) is one of the most important learning tools I have. No joke.
------
Every card remembers every rating you give it, as well as the time / date. This allows for Anki to solve for a 'forgetting curve', and predict when different cards have a chance to be forgotten.
There is furthermore the machine learning / stochastic descent algorithm to better fit the assumed forgetting curves to your historical performance. This is the FSRS Optimize parameters button in the settings panel.
Yes. Stochastic gradient descent, to be precise.
https://github.com/open-spaced-repetition/fsrs4anki/wiki/The...
In a nice and controlled manner, so seemingly no reason to panic just yet:
> I ended up suggesting to them that we look into gradually transitioning business operations and open source stewardship over, with provisions in place to ensure that Anki remains open source and true to the principles I’ve run it by all these years.
> This is a step back for me rather than a goodbye - I will still be involved with the project, albeit at a more sustainable level.
From AnkiHub:
> No enshittification. We’ve seen what happens when VC-backed companies acquire beloved tools. That’s not what this is. There are no investors involved, and we’re not here to extract value from something the community built together. Building in the right safeguards and processes to handle pressure without stifling necessary improvements is something we’re actively considering.
Relieved at that part where they say there are no investors involved, makes the whole thing a whole lot less risky. Good for everyone involved, and here's to many more years with Anki :)
Yeah my first thought on seeing the headline was “Uh-oh. Time to replace Anki.”
But finding out there are no VCs, no investors, I’ll stay with Anki for now.
But still, these HN comments - after an announcement like this - are usually a good place to find out about replacements.
Might as well give a recommendation then: I've been using hashcards [0] for a few weeks now and have enjoyed its simplicity and the fact that it all stays forever in raw markdown files and versioned git. A simple justfile has also been helpful.
[0]: https://news.ycombinator.com/item?id=46264492
Well every company claims no enshittification when they get acquired, in the end that's rarely the case. It's like Private Equity buying a company out and say "Nothing will change"
The key difference is the outside investors, who more times than not has no interest in what's best for users ultimately.
AnkiHub people seem kind of slimy in my experience (at least their leader, "The AnKing"). I hope they figure out a good leadership situation, and make stronger commitments to openness.
Anki is in a very solid position to be forked if anything happens, so even if this is bad news I have faith in the larger community.
I feel like there's a lot of confusion, so not sure if anyone can clear up mine, but I'm really struggling to see a significance of this.
Obviously, like all ignorant people do, I am going to oversimplify things here. But still, to me, the "platonic idea" of Anki seems a dead simple thing. All what I care about when using Anki is what's on the 2 sides of a Card, a question + answer, which can only be some visual image (possibly encoded as text, possibly just JPEG, I really don't care as long as it fits in my mobile device memory) + optional sound. That's really it. If it should be bi-directional or uni-directional card is a detail of how the deck is generated/encoded, and the spaced repetition algorithm is a detail of the app that I use to study (so, usually AnkiDroid, I imagine — an unaffiliated 3rd party; who even uses desktop apps nowadays?).
So, I imagine there can exist (and do exist) some minor additional features, like an ability to require a typed answer for a card, but it seems pretty minor, and I really don't see a lot of room for the app to evolve.
So, ultimately people need only a common .apkg format, which exists and is relatively simple (although I suppose it could've been even simplier), and a place like AnkiWeb, where people can share their decks, so Spanish top-2000 or basic integrals deck isn't re-invented over and over again. It's a pity that AnkiWeb isn't more open and will be even less open from now on, but as long as someone is willing to just host it (which is ultimately just paying for downloads traffic) it's easy to replicate, so no super-valuable IP here.
Of course, a primary use-case for Anki is a tool to make decks, but you could really do with pretty simple python script + YAML/JSON/CSV/whatever metadata file to convert it to AnkiDroid-compatible .apkg file.
So, basically, who cares? What is to "own" there?
IMO, the add-on value is the repository of decks that exist (and may or may not be free).
So an app-store of sorts.
As others have said, there are some provisions in place that make it allegedly harder to do a hard landgrab and keep people from freely sharing decks, to to me, even if it were so, I would not be too concerned.
In my opinion, the very act of creating one's deck is a key part of the learning. Maybe it's different for larning vocabulary, but as you said, it will be very hard to make those hard to share.
Learning a deck generated by someone else has never been as effective with me, so I think it's a false sense of time saving to use those.
Even in the worst-case scenario, Anki is already perfect for me as is.
I always though Damien earns well with the iOS mobile version. Does he also pass it to AnkiHub too so they can earn from the app sales?
As someone who has used Anki for a decade, a thank you to dae for everything. Best of luck in your future endeavours.
With that out of the way, some thoughts:
- Anki is in a really good position to work around enshitification. The app, at least to me, is "complete" - the only additional features that might pique my curiosity is a different scheduler (at the moment, they're integrating a newer one, although I don't follow enough to know the state of it). Additionally, modern Anki is really well architected: the core of it is a Rust library, that is used by all of the platforms [0]. You can write new front ends using that, or just fork the existing FOSS ones. Maybe dae does a gorhill and gives us Anki Origin.
- Really the only service-y part of Anki I use is AnkiWeb, which is basically a backup and sync system. Wonder how that'll evolve (if they do end up charging for it, I hope it is "Obsidian" reasonable). EDIT: Ooo, Anki has public server software for running your own version. Awesome! [1]
- The idea outcome in my opinion would have been some form of charitable organisation (Linux Foundation?), with people donating to support Anki.
- So, AnkiHub is a company that produces Anki flashcards, and they've scaled that quickly? Jeez. Obviously Quizlet proved there was a market for flashcards, but I didn't realise this was possible for Anki.
- No outside investment is... hopeful. Not quite sure what indicates that this company has the technical know-how to maintain it.
- I've heard too many stories of a maintainer or creative being "hopeful" about their new acquirers, only to regret it years down the line.
[0]: https://news.ycombinator.com/item?id=46299897
[1]: https://docs.ankiweb.net/sync-server.html
Oh for fucks sakes. No please no.
AnkiHub was already annoying with shoving AI into their Add-On without anyone asking for it.
I don't think this will go well
Can't help but wonder if it might not be time to start a FOSS alternative just in case Anki begins to decline.
It might not be the worst idea to do that anyway. Anki is great, but there's plenty of room for improvement. Off the top of my head, an architecture that doesn't involve fragile and finicky python bits and is designed to support multiple independent clients would be a nice step up (Telegram is a good model here — make a core lib with all the nuts and bolts which devs build clients around).
> make a core lib with all the nuts and bolts which devs build clients around
That is modern Anki. The core is a Rust library, which all the clients (desktop, web, Android and iOS) use. [0]
[0]: https://news.ycombinator.com/item?id=46299897
I learned something today, thank you.
Even so, I believe there's room for another open competitor or two in this space.
It's not FOSS but Mochi [0] is a pretty good alternative.
[0] https://mochi.cards/