Comment on The most popular Chinese keyboard app which is used by more than 450 million monthly users sends every key typed to Tencent in China.

Goodie@lemmy.world ⁨1⁩ ⁨year⁩ ago

• toofpic@lemmy.world ⁨1⁩ ⁨year⁩ ago

  You just save the first 50 digits typed after some email is typed, and you have all the passwords you need!

  source

  • Goodie@lemmy.world ⁨1⁩ ⁨year⁩ ago

    This only applies if a username is a email

    And if it is then what happens when people actually email someone? Autocorrect during login?

    source

    • ultimate_question@lemmy.world ⁨1⁩ ⁨year⁩ ago

      I don’t think they’re saying that method would yield 100% clean data but it would give you all the “necessary” data with the absolute bare minimum storage requirement

      source

      • toofpic@lemmy.world ⁨1⁩ ⁨year⁩ ago

        Yep, I only reacted to a “new requirement”: save space :)

        source
    • WarmSoda@lemm.ee ⁨1⁩ ⁨year⁩ ago

      They weren’t describing a use case for every single type of situation.

      source
• WarmSoda@lemm.ee ⁨1⁩ ⁨year⁩ ago

  I could be wrong, and this is a generalization of any country you can name, but my impression is data is stored on everyone so when they decide someday to look you up they already have all the data collected. It’s not really processed until needed.

  source

  • TheYear2525@lemmy.world ⁨1⁩ ⁨year⁩ ago

    And in hopes of it being useful later, when processing power is better.

    Hey GovGPT8, please rank the 10 citizens most likely to organize protests if we institute curfews.

    source

    • WarmSoda@lemm.ee ⁨1⁩ ⁨year⁩ ago

      Exaaaactly

      source
• perviouslyiner@lemm.ee ⁨1⁩ ⁨year⁩ ago

  And how can autosuggest / autocorrect be so bad with so much training data

  source

  • TheEntity@kbin.social ⁨1⁩ ⁨year⁩ ago

    Did you ever see how an average person types? It's not the amount of data that is the problem. We have too much dumb data!

    source
  • Steeve@lemmy.ca ⁨1⁩ ⁨year⁩ ago

    The real answer is compute power. At the moment it’s very expensive to run the computations necessary for big LLMs, I’ve heard some companies are even developing specialized chips to run them more efficiently. On the other hand, you probably don’t want your phone’s keyboard app burning out the tiny CPU in it and draining your battery. It’s not worth throwing anything other than a simple model at the problem.

    source
• BobKerman3999@feddit.it ⁨1⁩ ⁨year⁩ ago

  They can a “rollup” of the data to coalesce a lot of stuff and still maintaining precision

  source