Google Now is just one example of the increasing need for "smarter" software.[/caption] In late January, Google announced the acquisition of DeepMind for $400 million. Self-billed as a “cutting edge artificial intelligence company” that combines “the best techniques from machine learning and systems neuroscience,” DeepMind specializes in general-purpose learning algorithms for market segments such as games, e-commerce, and virtual simulations. Perhaps the most surprising part of the deal wasn’t in Google’s official announcement; it was the subsequent news reports that suggested the search-engine giant had beaten out Facebook for DeepMind’s assets, after a battle so intense that Google CEO Larry Page personally oversaw negotiations. Why were two tech giants fighting so fiercely over a small artificial-intelligence company? Artificial intelligence holds the potential to become the Next Big Thing. For Google, it’s no longer enough to simply index the world’s data (and scan it to serve up ads); users expect to interact with that information in natural and dynamic ways. Whereas a couple years ago, your typical office-worker or student found it sufficient to type a query into Google’s homepage and receive a page of blue hyperlinks, they increasingly want to ask their mobile device (“Where can I find good Chinese in this neighborhood?”) and see or hear a definitive answer in return. They want their social networks and search engines and mobile apps to work faster, anticipate needs, and constantly iterate. And that requires smarter software. At the moment, natural-language queries (the most intensive focus of many tech companies) are problematic. Ask Apple’s Siri a question, and it’s just as likely to kick you to a Web search as deliver an actual answer. IBM’s Watson, too, is best when questions are phrased in the “right” way. Facebook introduced a similar system with Open Graph, a search engine built to search through that massive network in a granular manner (and understand queries phrased in a “normal” way, such as: “Find friends of friends who live in San Francisco and like pizza”), but that project is very much a work in progress. In any of these systems, if you stray too far from the “path,” the underlying software quickly becomes confused and useless. A system that can dynamically adjust in response to user inputs, and make “human-like” inferences based on the information it already has on-hand, could quickly overcome many of these issues confronting search as it attempts to evolve into a more productive and streamlined form. Software from DeepMind could provide the intelligence needed to help make that goal a reality—which is why Facebook and Google, two of the biggest companies with an interest in search, were so anxious to get their hands on the startup. The question now—and one that will probably end up answered in the affirmative—is whether the DeepMind acquisition will kick off a heated round of acquisitions, as the big firms target more artificial-intelligence startups. Image: Google The Coming War for Artificial-Intelligence Startups
[caption id="attachment_16931" align="aligncenter" width="618"] Google Now is just one example of the increasing need for "smarter" software.[/caption] In late January, Google announced the acquisition of DeepMind for $400 million. Self-billed as a “cutting edge artificial intelligence company” that combines “the best techniques from machine learning and systems neuroscience,” DeepMind specializes in general-purpose learning algorithms for market segments such as games, e-commerce, and virtual simulations. Perhaps the most surprising part of the deal wasn’t in Google’s official announcement; it was the subsequent news reports that suggested the search-engine giant had beaten out Facebook for DeepMind’s assets, after a battle so intense that Google CEO Larry Page personally oversaw negotiations. Why were two tech giants fighting so fiercely over a small artificial-intelligence company? Artificial intelligence holds the potential to become the Next Big Thing. For Google, it’s no longer enough to simply index the world’s data (and scan it to serve up ads); users expect to interact with that information in natural and dynamic ways. Whereas a couple years ago, your typical office-worker or student found it sufficient to type a query into Google’s homepage and receive a page of blue hyperlinks, they increasingly want to ask their mobile device (“Where can I find good Chinese in this neighborhood?”) and see or hear a definitive answer in return. They want their social networks and search engines and mobile apps to work faster, anticipate needs, and constantly iterate. And that requires smarter software. At the moment, natural-language queries (the most intensive focus of many tech companies) are problematic. Ask Apple’s Siri a question, and it’s just as likely to kick you to a Web search as deliver an actual answer. IBM’s Watson, too, is best when questions are phrased in the “right” way. Facebook introduced a similar system with Open Graph, a search engine built to search through that massive network in a granular manner (and understand queries phrased in a “normal” way, such as: “Find friends of friends who live in San Francisco and like pizza”), but that project is very much a work in progress. In any of these systems, if you stray too far from the “path,” the underlying software quickly becomes confused and useless. A system that can dynamically adjust in response to user inputs, and make “human-like” inferences based on the information it already has on-hand, could quickly overcome many of these issues confronting search as it attempts to evolve into a more productive and streamlined form. Software from DeepMind could provide the intelligence needed to help make that goal a reality—which is why Facebook and Google, two of the biggest companies with an interest in search, were so anxious to get their hands on the startup. The question now—and one that will probably end up answered in the affirmative—is whether the DeepMind acquisition will kick off a heated round of acquisitions, as the big firms target more artificial-intelligence startups. Image: Google
Google Now is just one example of the increasing need for "smarter" software.[/caption] In late January, Google announced the acquisition of DeepMind for $400 million. Self-billed as a “cutting edge artificial intelligence company” that combines “the best techniques from machine learning and systems neuroscience,” DeepMind specializes in general-purpose learning algorithms for market segments such as games, e-commerce, and virtual simulations. Perhaps the most surprising part of the deal wasn’t in Google’s official announcement; it was the subsequent news reports that suggested the search-engine giant had beaten out Facebook for DeepMind’s assets, after a battle so intense that Google CEO Larry Page personally oversaw negotiations. Why were two tech giants fighting so fiercely over a small artificial-intelligence company? Artificial intelligence holds the potential to become the Next Big Thing. For Google, it’s no longer enough to simply index the world’s data (and scan it to serve up ads); users expect to interact with that information in natural and dynamic ways. Whereas a couple years ago, your typical office-worker or student found it sufficient to type a query into Google’s homepage and receive a page of blue hyperlinks, they increasingly want to ask their mobile device (“Where can I find good Chinese in this neighborhood?”) and see or hear a definitive answer in return. They want their social networks and search engines and mobile apps to work faster, anticipate needs, and constantly iterate. And that requires smarter software. At the moment, natural-language queries (the most intensive focus of many tech companies) are problematic. Ask Apple’s Siri a question, and it’s just as likely to kick you to a Web search as deliver an actual answer. IBM’s Watson, too, is best when questions are phrased in the “right” way. Facebook introduced a similar system with Open Graph, a search engine built to search through that massive network in a granular manner (and understand queries phrased in a “normal” way, such as: “Find friends of friends who live in San Francisco and like pizza”), but that project is very much a work in progress. In any of these systems, if you stray too far from the “path,” the underlying software quickly becomes confused and useless. A system that can dynamically adjust in response to user inputs, and make “human-like” inferences based on the information it already has on-hand, could quickly overcome many of these issues confronting search as it attempts to evolve into a more productive and streamlined form. Software from DeepMind could provide the intelligence needed to help make that goal a reality—which is why Facebook and Google, two of the biggest companies with an interest in search, were so anxious to get their hands on the startup. The question now—and one that will probably end up answered in the affirmative—is whether the DeepMind acquisition will kick off a heated round of acquisitions, as the big firms target more artificial-intelligence startups. Image: Google 
