1. Crowdsourced word weighting: your keyboard's stochastic predictions are no longer mostly based on your typing, but rather on what 'everyone' is typing as their next word. This makes the word replacements it does often suboptimal to downright nonsensical.
2. Aggressive lookbehind correction: these days you have to be seriously on your guard for your keyboard to not sneak-edit something you typed 5 words back, because autocorrect suddenly decided that the probability is high you meant to say something else there (which it clearly isn't, as your eyes and brain exist)
The problem your encountering is downstream from point 1. Basically your keyboard thinks due to the way most people construct a particular sentence, you're gonna want to type "bold" next, despite "hold" clearly clearly making more sense. So it'll force "b" on you 4 times in a row until it realizes you really want to type "h".
Going back to the old style of doing keyboards (mostly user-learned dictionaries and probability weighting, and little lookbehind autocorrrect) could be done, but within Google and Apple there are probably people who got promoted by switching to the current shitty system. They'll block off any attempt at someone messing with their pride.
(There is a third 'problem' where your visual keys do not correspond to the touchmap at all. Swiftkey has a feature where it can show you what your touchmap and heatmap look like versus the actual layout and it its often staggeringly different, with many keys vastly tilted. When you try to desperately type "h" after 4 misses, you're doing that with your index finger in "hunt and peck" mode, which does correspond to the visual layout but not with your usual typing on the touchmap layout. There is no way for your keyboard to know you're in "hunt and peck" accuracy mode.)
In the video, the user is typing 'Thumbs up', and when they get to the first 'u' the keyboard shows a 'u' being pressed but a 'j' is inserted instead. Are you suggesting that, due to the way most people construct sentences, the OS thinks that 'thjmbs' is the most likely word? And then the next time the OS thinks that 'thhmbs' is the most likely word?
Both of the issues you've mentioned are common, and irritating, but if you watch the video you can see that that's not what's happening here. Before any autocorrection or adjustment is being done, the keyboard is registering a 'U' and the OS is inputting a J or H or I or some other nearby letter.
The video also debunks the touchmap discontinuity issues as well, because you can clearly see which key the keyboard is registering; it's not assuming that you meant to press J or it would highlight the J; it's registering a U, highlighting U, and inputting J.
It sounds to me as though you didn't watch the video and just assumed what issue was being discussed; please do watch it, because this is another, relatively new, issue that lots of people have seen and which is far worse and more frustrating than the other legitimate issues you mentioned.
Apple additionally may have just bugged up their implementation as well, but the above mentioned issues exist even on Android, and didn't a decade ago.
I still contend that the single best touchscreen keyboard and autocorrect implementation was the onscreen keyboard on the Microsoft Zune HD. A tiny tiny screen, and you could still type without looking and nearly always end up with the right text. It was magical, and creepy in retrospect.
But nobody bought it so we had less good keyboards for a decade. Then companies insisted that they could throw "Algorithms" at the problem (which is what we had been doing for a decade but whatever) and make it magically better and now everyone gets worthless autocorrect because of the everpresent "Nobody is actually average so tuning your system to the average makes it bad for everybody" problem that has infected literally all "Data driven" product decisions.
We literally had better text prediction using boring methods. We literally had working voice control on flip phones from the 90s. All on device too.
In addition to the other problems (the keyboard being too prone catching extremely subtle slides below UI response time), there certainly is the problem of when you crowd source enough data you crowd source all of their collective mistakes, too. In a lot of that raw data mistakes are going to be as common or more common than corrections and/or originally correct spellings.
We do have a great filter for this called a "dictionary", but as the above commenter laments companies have given up on "just autocorrect to dictionary words" for much more complex "learning" models and filtering them back to just dictionary words is antithetical to the (sunken cost) expense that went into training these models, and/or the KPIs and promotion incentives that keep prioritizing "AI" and giant crowd sourced data vats over simpler mechanics and local user specifics.
The voiceover is deceptive (unintentionally?)...
They touch the [u] which shows the popover U but you can see them slide their thumb down off the [u] key onto the [j] key.
I guessed that was the issue, repeated it on my phone (SE) and only then looked at video and it's obvious when you see him do it in slo-mo. Edit: I have most prediction turned off (I mostly find slyde typing to be fastest, and I hate automiscorrect on uncommon words).
iPhones are very very sensitive to tap-slides which causes many UI gremlins (a variety of terrible side effects that you can't avoid if you're designing a UI).
Over time, most people seem to intuitively learn not to slide when tapping.
I'm unsure how many designers/developers even notice the effects of slide since they have learnt to avoid sliding? When I watched beginners on iPhones you see them get frustrated by things not tapping and other subtle effects (HTML event interactions, scrollable areas, buttons, inputs).
Same thing can happen on Android. One menu button repeatably failed if I used my left hand - took me a while to work out the issue (and a bit of work to increase the tappable area so a bit of slide was accepted and worked better for neophyte users).
If I ever meet the person that invented lookbehind correction, I’m not sure I’ll be able to restrain myself. This person has robbed me of my peace of mind as I now have to be on guard every time I type anything on a mobile keyboard
your comment reminds me of this comic from the 2000s that became a bit of a meme back in the day
swap out "comic sans" with "aggressive lookbehind correction" and it'd fit perfectly ;-)
It's real frustrating that Apple has decided to put just about everything under only a single Settings switch and won't break it out into individual things.
It's also frustrating that for about half an iOS version Apple seemed to have caught on that the update behind was catching people off guard and implemented an extra, more obvious change animation. The whole word flashed in a bright blue or yellow when it changed and had a visible undo button. That was useful. But then the button didn't survive the next point release and the animation kept getting subtler again until it disappeared.
Unfortunately this falls apart when I try to type anything that isn’t common English words: names, code, rare words, etc.
I also think that the keyboard could learn the different “rhythms” of typing - my normal typing which is fast and practically blind, and the careful hunt and peck which is much slower and intended for those out-of-distribution inputs. I bet the profile of the touch contacts (e.g. contact area and shape of the touches) for those two modes looks different too.
So I realized I had exchanged correcting the same word four times in a row to correcting the same letter four times in a row.
> pick adyacent keys
But there is a way for your keyboard to simply show the real size/position of buttons so that in hunt and peck mode you'll be correct
Yes, but the tradeoff in that case is that for most casual typing it will be less forgiving and you'll make more (uncorrected) mistakes
My understanding is it's not just about including or not including some extra fixed clickable padding around each square (in which case it indeed wouldn't harm to show the whole area), but about dynamically adapting the area, based on frequent off-target clicks, probabilities, etc.
The big reason after years of SwiftKey use I finally uninstalled it is because it became too much of an ad vector for "you should use Mobile Edge and have you tried our new Bing Mobile app yet". I also haven't used it in a couple of years, but I'd be surprised if it doesn't have some Copilot button or buttons somewhere now.
The industry really does forget all the lessons it learned...
I still feel the pinnacle was ~2011 Windows Phone. It was some kind of swipe-to-type, but maybe not Swype specifically? At any rate, it seemed to use "how humans actually talk" as a guideline, because it was do a great job of predicting what words I would actually mean to use in a row.
Modern keyboards are like, I know you just said "I want" but instead of predicting "to" I predict "rip". I mean the letters are close. And "I want rip" makes way more sense than "I want to." You're welcome!
The fact that Apple will as often as not autocorrect grammar from actually-correct to wrong -- and systematically screw up spelling -- in not just transcribed Siri but also in typing is just inexcusable at this point. It will even Randomly capitalize Certain words!
And i used to be able to backspace the wrong word and fix it and it would learn thats what I meant. Now if I try that, it'll frequently keep trying to edit to the word I didn't mean unless I press the little checkmark in the autocorrect panel. Just annoying UX.
I'd be giving the keyboard a try!
Apple keyboard is shit. Swype (the one Microsoft bought) is better but still shit. Gboard is ok. But none of them are close to that windows phone keyboard. I still miss it.
Suggests words that make no sense, preferring rare words to much more widely used and obvious matching picks. Has the vocabulary of a poorly educated five year old idiot savant — fails to complete many words you use fifty times a day, but sometimes surprises you by suggesting something you'd hear a couple times per decade. Doesn't know other forms of the same word, forcing you to correct it manually over and over again, often failing to remember the word until you type it in four or five times.
Yes, I've downloaded all the dictionaries, tried it on many phones, and my friends are of the same opinion: it really is just bad.
This is a patent case where IA made a function worse instead of better, yet companies clinged to it for some reason.
The swiping keyboard from Apple simply refuses to do "and" for me. I get "abs" (I'm not a gym rat; I don't talk about that) or "Abbas" (the only one I know is the Palestinian president, and I don't talk about him either) almost every time. I hate the autocorrect-something-five-words-back problem, but not being able to recognize one of the most common words in the language is unacceptable crap.
I'll give Swiftkey another try.
Troll answer: A-Z label maker keyboard
Even switching to the Hacker's Keyboard and tweaking some settings still has me smacking the "tab" key or whatever when hitting space.
Just out of curiosity, who here is a one-handed texter, like me? I just assumed my constant need for error correction was because I only use one hand (and thus, one thumb) to type, but this thread has me wondering.
Absolute perfect typing experience, better responsiveness and almost entirely free of mistakes. It's mind-boggling.
Frustrating if you are a 13 mini user
Guess they’ll want us to carry iPads in our pockets for these UIs to actually work :)
Perhaps they wanted to sell more Smart Keyboards.
* I type a word, it shows up correctly
* I type a second word, my phone CHANGES THE PREVIOUS WORD
* A silent tiny rage removes several seconds from my life
One can find many iPhone sourced typos in my HN history which I leave, usually, as a method to preserve sanity.
By contrast, the typing experience on a 2.5” Unihertz Atom screen is shockingly acceptable…
> The Unihertz Atom's 2.45-inch screen (240x432 resolution) is "shockingly acceptable" for its niche.
Though of course Google's Gemini-whatever does manage to subtly miss the mark even there: I said (and think) that the typing experience is acceptable, I said nothing about the screen. If I remember correctly, the last one I handled, the screen was resistive rather than capacitive, and it felt weird and squishy. Still not bad for the price, and it's a minor miracle how much Android software can still draw a coherent layout with that kind of resolution, but...
I'd never come anywhere close to trusting it with anything important, but then again maybe that's not such a bad relationship to have to a smartphone...
My first iPhone was a 4S and i was astonished how correctly i'm typing. At least in English.
I even managed to bully the spell checker into reasonably accepting both English and Romanian, back when they didn't have multiple languages at the same time on the keyboard.
I'm not sure when it started to go downhill, but I was using an XS and it was at at least one more version after whatever XS shipped with.