Hacker News Reader

They have a few sentences about each kanji as a mnemonic which helps you remember the pronunciation and meaning.

Perhaps this could be somehow plugged into DALL•E. Create an image for each kanji for yet another data point which helps memorize it.

View on HN

There's a thread on the user forums with an example of doing that, with some images from various of these AI image generators. I think the most obvious flaw with the idea is that you can feed in a prompt like "a prawn and a horse are waiting in the railway station", and the image generator will happily produce an image missing at least one of the keywords, or weird things like a horse coloured like a prawn... So the images look cool but fail as cues to remember all the keywords that are supposed to be prompting you to remember the pieces of the character.

As a WaniKani user, I don't think I'd want AI-generated content in there. The entire value proposition is that all the content is 100% hand-crafted by people who know Japanese way better than me.

Reply on news.ycombinator.com