It’s undeniable that in the nearly 15 years since Google became a Real Thing and unleashed search engine greatness on the world, search has evolved. Google has done a solid job of “understanding” search queries these days, to the point where we no longer have to fuss with things like correct spelling or Boolean variables. Further, voice search has improved dramatically in recent years; it no longer seems incredible that you can ask your phone a question and hear it spit out an answer in a small robotic voice imbued with hints of human inflection. And of course, the amount of data that Google has gleaned over the years helps predict what we’re looking for.
However, search is still fundamentally about submitting a query and hoping that you get a bunch of links to help you find what you’re looking for. According to Slate’s Farhad Manjoo, however, Google has much higher aims: the company wants to build the computer from Star Trek. Manjoo says that this isn’t slang for some cool thing Google wants to do with search; Googlers he’s spoken with in recent years are actually using the Star Trek computer as a template and a goal, and they refer to it all the time as they’re developing new search capabilities.
Google in its infancy
Essentially, Google wants users to be able to simply say out loud what they want and have helpful conversation with high-level artificial intelligence. This AI would understand what you’re asking and give you the answer (instead of providing a path for you to discover the answer on your own), but it would also anticipate what information you need. There are glimmers of this potentiality with developments such as Google Now, the predictive virtual assistant.
Google Now: A shadow of things to come
To accomplish this monumental task, Google can’t simply algorithm around the Internet for keywords; it needs to possess information and also understand and process difficult concepts such as context and human language.
Here’s the kicker: A Star Trek-esque Google search engine could happen in as little as three years. Or at least, it would pass the “toothbrush test” of being able to correctly give users answers at least twice a day. Considering what such a search engine needs to comprehend to make that work, that feat sounds incredible.