If you use the privacy respecting Gboard alternative called FUTO keyboard, you’ve probably noticed that the built in swipe typing is HOT GARBAGE. (Typing this currently with two thumbs for this exact reason.)

Most keyboards improve their swipe algorithms by simply spying on you and logging your typing data. FUTO isn’t about that, so they have built a simple webpage based typing game that you can use to improve their system in an ethical and voluntary manner! Just swipe the website’s keyboard to type a provided sentence.

I love this, they can crowdsource the improvement without invading privacy!

Share with any relevant communities you’re a part of. The more data, the better this gets.

  • arthurpizza@lemmy.world
    link
    fedilink
    English
    arrow-up
    2
    ·
    11 months ago

    Are they going to share this data with the open source community or is this just for their proprietary keyboard?

  • LiveLM@lemmy.zip
    link
    fedilink
    English
    arrow-up
    1
    ·
    edit-2
    11 months ago

    What a cool idea!
    As an avid swipe user, I’ll try to send as many samples as I can!

  • Chais@sh.itjust.works
    link
    fedilink
    arrow-up
    1
    ·
    11 months ago

    My only gripe with this page is that it’s in the browser and it keeps recognising a back gesture (swipe from the edge of the screen) whenever a word starts with q or p.

  • kitnaht@lemmy.world
    link
    fedilink
    arrow-up
    0
    ·
    11 months ago

    They should simply allow collecting this data via opt-in. I’d gladly provide it, but I don’t want it to be opt-out, because then this feels sneaky.

    • dracs@programming.dev
      link
      fedilink
      English
      arrow-up
      1
      ·
      11 months ago

      That would require a lot of data privacy concerns to be addressed. Even if it’s an explicit opt-in. The current method uses sample text which can’t include PII. Using user supplied text would almost guarantee they’d get names and other PII in their data set.

      I also imagine it’s harder to train the model when you don’t know exactly what the user was trying to type. I.e. Was the swipe detection wrong, or did the user delete the word because they changed their mind on what to write?

    • trevor (he/they)@lemmy.blahaj.zone
      link
      fedilink
      English
      arrow-up
      1
      ·
      11 months ago

      Is it? The FTL is restrictive about who is allowed to redistribute and modify anything covered by it. Is this data covered under a different license?