Very useful Addon

The Link :

:gear:FSRS Smart-Pass​:brain: - AnkiWeb :backhand_index_pointing_left:

1 Like

If you are the developer of that add-on, you can use this AnkiForums thread (or Github) as the support thread for that add-on and link to it from the add-on page, like this:

2 Likes

But in that case, the title should be the name of the add-on, and we should move this to the appropriate category.

3 Likes

I’m concerned about both the functionality you’re pushing with this add-on, and the way you’re marketing it. Just because you call your add-on “Smart” and say it is “intelligent” doesn’t mean that it is.

Having the app auto-grade your cards based on FSRS’s predicted likelihood that you’ll remember them is completely the opposite of what a user should be doing. First, it assumes that the prediction is correct for all cards – which it won’t be. And then it assumes if R is above a certain line (88%), the card should always be graded as correct – but even assuming 88% was a perfect prediction, that means you would still get the card wrong 12% of the time. (Similarly, even if R is 75%, you would only get it wrong 25% of the time, not 100% of the time.)

FSRS can’t function correctly without real grading information from the human user. This add-on just feeds FSRS’s own predictions back to it, which basically creates garbage review history that will be useless for re-optimizing in the future. [You also say some conflicting things about whether the card will be marked Good or Easy.]

Consider that the only research that has been done about whether time-to-answer correlates to correctness of answer found that it could not be used as a proxy for grading. Your add-on takes that a step further and draws a single line across which the app simply assumes whether you got the answer right or wrong.

I’m also concerned that you’re mimicking the very popular (and useful) FSRS Helper – by putting a :gear: at the front of your name, and having a similar page layout. Imitation might be the sincerest form of flattery, but here I think it is misleading, and will cause confusion among users.

I encourage you to rethink what you’re doing here. If this is truly well-intentioned, you should take this add-on down and learn more about FSRS before trying to contribute.

4 Likes

Since Anki and add-ons are open source, most developers actively permit copying and code reuse under their licenses (sometimes there are exceptions). So it’s very common for users to get confused by the many duplicate Anki add-ons.

Non-developer Anki users are unfamiliar with these licenses, so when developers reference code or ideas from other add-ons it’s common practice to give credit on the add-on page. e.g. “this add-on references code from A”, “this idea was inspired by B”. It can make it clear to general users that you are not the author and express gratitude for their contributions.

The add-on doesn’t copy any code from FSRS Helper (apparently), so this is not relevant. It appears to be written by an LLM.

3 Likes


mr fsrs’s take on the matter of this addon.

this is just pure ai slop.

1 Like

How do you know btw? Are there systematic errors or issues that LLMs tend to make present in this code? Or is it specific things that are good, but typical for LLMs?

I know this is off-topic, but maybe I have a chance to improve my knowledge and thus programming abilities.

Edit: Just to make it clear: I am not the author of that addon (or any addon). I just want to learn more about programming so that my future code contributions can be improved.

2 Likes

The description in the AnkiWeb page is a dead giveaway. The code itself appears to be well-written. The only hint I can find is trivial comments like this (LLMs love comments):

2 Likes


this very obviously ai generated response on the addon page to someone pointing out that this addon is broken by design is a pretty good tell that the rest of it is also junk.

4 Likes