Whether you, like me, beleive that QAZWSX keyboards make far more sense, especially in a machine learning world, I think we all agree a layout designed to circumvent jamming typewriter keys doesn’t make sense in modern society on modern devices.
Whether you, like me, beleive that QAZWSX keyboards make far more sense, especially in a machine learning world, I think we all agree a layout designed to circumvent jamming typewriter keys doesn’t make sense in modern society on modern devices.
I put AI in quotations because it really isn’t an artificial intelligence, it’s just machine learning.
Okay, so put it like this, let’s say you split a keyboard in two vertically and then rather than have 13 keys on each side, you have one. But your brain still sees the thirteen keys, when you type, based on a local language model, the keyboard would say, this list of words is most often used after this word. However based on an input on the left side X times and Y times on the right, the word is around a certain length and has a likelihood of these letters being used and then it replaces your gibberish with the word you were trying to type. Obviously with machine learning it gets better over time.
We walk with these mini computers, why wouldn’t we use them to process communication?
But what happens when it’s inevitably wrong? How do you type the word you actually meant to type?
You would get a selection of words, like autocorrect now. But if none of those were correct, you could fall back to legacy mode to put in the correct word and next time that word would be weighted so you’re more likely to see it.
I mean, I see how this could kind of work… but I still think most people would want to use a regular keyboard. Correct me if I’m wrong, but this seems needlessly complicated, especially since most people use and are fine with the system we have now.
I think that if you had a new keyboard enabled on every new phone, people would adapt. Unfortunately though, that would never happen.