WaniKani mnemonics in image-form done by AI

It’s just a screenshot from the Mad Max game that came out a few years ago. :laughing: It’s what the mnemonic conjured in my head.

At first I used the exact sentence and this came out

But I wanted to see someone jump into that car so I slightly rephrased it :sweat_smile:

Jumping into a driving car from that distance is a 100% manly and only 80% deadly lol

A fang on a building shows that it’s wicked. Something wicked probably lives in the building too.

Is the subscription with 10$ a month sufficient?

I think so, yes. I started with the biggest first but I’ll most likely downgrade at some point

This is pretty cool!

Btw is this you on your profile picture holding artwork?

Yes ^^
It’s a book I’m holding

I thought it might be your drawing :slightly_smiling_face:

I just found out that one of the main advisors for Midjourney is the founder of Second Life. Interesting.

Yeah, I also looked at their lineup the other day, and now I’m not surprised any more that they got those stunning results!

If that’s a farmer, I’m Brad Pitt.

@saraqael MVP :heart_eyes:

Haha thanks, I’m glad you seem to like it! :smiling_face:

I have a request if it’s okay. I would really like to see the kanji 済 (come to end and) as an image! The rephrased mnemonic is:

A cyborg stops two simultaneous tsunamis from ending the world

If that is too much in one sentence you’re free to shorten it. Thanks! :smile:

Here you go :blush:

Oh, wow! :open_mouth: That looks way more amazing than I would have thought. Thank you!

@Mo-Kanji if you want to collaborate, I have an access to DALL-E 2 which is the ‘best’ text-to-image model yet with its own limitation.

Edited: I forgot that I asked this several months ago :sweat_smile: Anyone tried creating WK visual mnemonics using AI (DALL-E etc.)?

What does one need to be able to collaborate?

I don’t know, maybe to create some script(?)

I cannot do it :sweat_smile:
But I could test a lot :innocent:

This is what DALL-E first generate with your prompt.

One thing to consider is that DALL-E 2 research team really limits its model so the generated image will not look wild enough compared to Craiyon or Midjourney.