The great leaps of technology that have taken place in the last decades have been astounding. The internet stands out as the greatest technological achievement and may prove to be the greatest invention of all time. While the great leaps are fascinating they would not exist without the little innovations that make them possible. One of the most fantastic innovations that we have seen in the digital age is autocorrect. While autocorrect sometimes gets it wrong and can often result in an embarrassing moment or awkward apology it very often gets it right.
Autocorrect was invented by Dean Hachamovitch who is now a Vice President at the company. It all started with the word the (or should I say teh). Hachamovitch was on the word team at Microsoft and was tasked with making everything run a little smoother. At the time Word had a glossary that you could essentially place items (pictures, logos, titles, etc.) in your documents using short cuts. For example, if you said DeanH and then pressed F3 it would replace it with Dean Hachamovitch. Hachamovitch realized that this glossary had greater potential and he started to build the Microsoft glossary to take on more tasks. The first word he targeted was the. He made a glossary rule to replace teh with the, simple and effective. Autocorrect was born.
The next step was to record every mistake that they could think of into the glossary. A long and lengthy task that was likely completed by hundreds of interns. As technology has progressed the task became increasingly automated with the computer now able to spot inaccuracies with ease. As artificial intelligence and machine learning continue to improve autocorrect now continues to learn more and more about the writer and can correct with their writing style still in place.
At this stage of machine learning, nothing is perfect but we are advancing quickly. Autocorrect is likely the first example of intelligent automated computing and it is the start of digital technology being able to correct a lot more than words. Thanks, Dean teh results speak for themselves.
