Even thought "AI" has been a study of scientific research since the 1970s, it's only in the past few months that it's appeared seemingly everywhere you look.
AI suddenly crossed a threshold from being theoretical and impractical to useful and feasible. Across domains like images, audio, programming code and the written word, was suddenly able to provide near-human - and sometimes better than human - performance at different tasks.
So what changed? In a word - graphics. Computer graphics cards, the kind used in your PC or video game console, have become incredibly performant in order to pump all of those pretty pixels to your screen.
It also turns out that those same cards, known as GPUs, are also very, very good at training artificial intelligence "models" - the brain-inspired blobs of data and code which calculate results based on the data they've been previously exposed to.
That, combined with the fundamental insight that more data and more computation is enough to keep improving the performance of these models without the need for any new major scientific discoveries. Using essentially the same techniques which were developed in the 1970s and 80s, scientists and engineers have spent the past decade using more and more GPUs and more and more data to create bigger and bigger models. Now, we've reached the point where those models are mimicking human performance. Very soon, we'll be a point where it exceeds human performance.
Goldman Sachs estimates that more than 300 million jobs could be impacted by tidal wave of AI. It sounds absolutely terrifying, but society has undergone major transformational changes before. Horses used to be a major part of the economy as well - until we've got fully automated luxury communism, there will always be work to be done.
Still, it's good for you to familiarize yourself with this stuff. If you're interested in building some of your own AI Systems to help you solve some of the problems in your life, the best way to start is with AI Studio.