Google’s acquisition of DeepMind for $US400 – $US500 million is another symptom that the next big war in tech is for true Artificial Intelligence that will make Google’s current core product, “search,” look like a primitive offering that we’ll all laugh about in the same way we chuckle over the fact that the original Apple II computer came with a cassette tape for data storage, decades ago.
This war will be fierce. There are signs that companies like Yahoo, Facebook, Apple and Microsoft all regard Google’s core search offering as a potential long-term weakness if it doesn’t evolve fast enough.
When people think about Google and search, they don’t think “weakness.” Google dominates search, consistently fielding around 70% of all searches on the web.
The problem is that those searches are on the web, and increasingly humans need to find things that are outside the web or shielded from it.
There are a bunch of new things that Google is actually quite bad at searching, such as Facebook, Twitter, app stores, and data that lives inside apps. Those are not small areas of the online world — they’re arguably the future of it.
One way to think about search is to regard it the way that humans will interact with machines in the future. Now, we need to find some information, so we punch keywords into Google’s search box. A lot of people in tech, however, regard this as a feeble way to interact with machines. They look at old episodes of Star Trek, in which the characters talk with and interrogate the star ship’s computer, and think that maybe most people would rather interact with machines like that: in normal, conversational language.
More than that, an intelligent machine would be able to anticipate your needs before you request them — making the overt act of searching obsolete because the machine has already recognised the semantics, or meaning, of your behaviour and figured out what information to present to you, when you need it.
- Microsoft’s Bing team has done a ton or research on voice search using “deep neural networks,” reducing the error rate and speeding up the response time to vocalized requests.
- Yahoo recently launched Aviate, its mobile app search and organisation product that anticipates what apps you’re going to need on your phone so that you don’t have to swipe through several screens of apps that you rarely use.
- Google has launched a product that can search the inside of apps — which are currently off limits to traditional search.
- Facebook is developing an AI/speech recognition unit that hopes to use its vast database of Facebook posts to create smarter more intuitive responses to requests.
- Apple has Siri, of course, its much derided personal assistant. (Google has Google Now, a competing product on Android.)
- And there are a bunch of smaller companies that have all advanced search technology beyond keyword matching and link ranking: Foursquare has the ability to handle complex, natural language queries, according to Techcrunch; Disconnect has a “secret” search product that hides your data while you search; DuckDuckGo does something similar and has risen in popularity since the NSA domestic spying scandal; Jelly has a Q&A based search product that lets you crowdsource the answers you’re looking for from other humans.
All of these developments are taking place in such a way that threatens to peel off users from Google’s core search product, or create new ways of doing things that bypass Google. If people stop needing to type in keywords to perform searches, then Google is screwed.
Google isn’t stupid, of course. It’s reacting and building its own alternative products, too. It recently updated its core search algorithm to handle natural language requests. Underestimate Google at your peril.
And we don’t know what Google will use DeepMind for, of course. Google has a million things that AI could be applied to — from the giant industrial robots of Boston Dynamics, to driverless cars, to new phones from Motorola.
But one thing is obvious: Google is looking way beyond a world in which everyone continues to type words into a little box to get what they want.