Oh. So apparently we've been using autocorrect wrong this whole time.

Autocorrect doesn’t exactly have a reputation as the most reliable toot…tune…TOOL around. We’ve all had the phone-squeezing, teeth gritting moment that comes with typing a mildly obscure name or suburb.

(TIP for Sydneysiders: Avoid Woolloomooloo. Your phone will combust.)

Look, it’s trying to be helpful. It knows barely anyone can spell anymore nor has the time/inclination to look words up. And so it intervenes, using three factors to guide it:

  1. What dictionary words have similar tap patterns, ie. letters in a similar place on the keyboard;
  2. Which of those words are more common in your chosen language;
  3. Which of those words are most commonly typed by you.

Most of the time it gets it right. But others, it gives the internet meme-worthy gems like this:

And this:

But according to the bloke who helped invent it, all the embarrassment, all the rage could be avoided if we just used it properly.

Speaking to LifeHacker, Ken Kocienda, former Principal Engineer of iPhone Software for Apple, explained that key to ensuring autocorrect performs at its best is to STOP FIXING ITS STUFF UPS STRAIGHT AWAY.

So rather than immediately retyping the word that's been incorrectly, erm, corrected, complete your message and fix any errors at the end.

By allowing your keyboard to examine longer passages of text, you're helping it learn more about your language patterns. This, Kocienda insists, will ultimately make your keyboard smarter and more intuitive.

In other words, stop being a helicopter parent and let it learn from its own mistakes.

00:00 / ???