By Justin Mullins Innovation is our regular column that highlights emerging technological ideas and where they may lead When the search engine Wolfram Alpha launched earlier this year, the interest was huge. Enticed by a well-oiled publicity machine, web users swamped the site and its servers were overwhelmed. Then everything went quiet – so quiet that it was easy to imagine that Alpha would follow countless Google wannabes to the great search engine directory in the sky. That was to reckon without Stephen Wolfram, a physicist famous for creating and selling the mathematics software Mathematica and for his pioneering but controversial work on cellular automata. In the past few months, Wolfram has been plotting the future of Alpha. Last Thursday, we saw the first stage of his plan with the link-up between Alpha and Microsoft’s search engine Bing. That’s just the beginning, however. Wolfram’s plans for Alpha are much more ambitious. Alpha is not a conventional search engine. It aims to provide factual answers to questions using information from databases around the web. The goal is to provide expert-level knowledge to everyone on the planet. So type in bismuth and you get the element’s atomic number and weight, and much more. Wolfram also included advanced mathematical capabilities. Alpha will plot complicated functions with ease, even in 3D. It can compare stocks or the inflation rate in two countries. By comparison, Google can do high-school calculations but cannot answer factual queries unless a webpage somewhere explicitly answers them. Google appeared spooked by Alpha’s launch, and rushed out a version of a tool that anybody can use to present their own data visually. It also stepped up attempts to index the “dark web”, things like databases that are not directly accessible to search engines. But soon after the launch, people began to wonder whether Alpha could live up to its own hype. Reviews were mixed. Although clearly powerful, users complained that the engine often didn’t understand the questions it was asked and that finding information was only possible if you knew it was there already. Consequently, public interest in Alpha waned. According to the web analytics site compete.com, Alpha had over 1.5m unique visitors in May but only 260,000 in October. But Alpha may yet confound the sceptics. Last month, Wolfram released an application programming interface, or API, that allows anybody to build software or websites that use Alpha’s abilities. “Alpha is a technology platform that allows one to inject computable knowledge into any application or computer system,” says Wolfram. The deal with Microsoft is the first sign that this plan is working. From today, Bing will carry answers from Alpha in response to certain types of query. Bing will use Alpha’s mathematical plotting capability, for example, and its body-mass-index calculator. The service is being opened to the public in the US first. Letting others build on Alpha’s abilities may give the service the boost it needs. Wolfram claims there is more to come. Since its launch, Wolfram says Alpha has been exposed to more utterances than a typical child would hear in learning a new language, allowing it to get smarter at understanding how people phrase their requests. My own experience is that Alpha seems able to answer queries today that it could not parse just a month or two ago. It feels like Alpha has become smarter. Ultimately Wolfram wants Alpha to generate knowledge of its own. “One of things I imagine will happen with Wolfram Alpha is that you’ll be able to ask it a question, and instead of it using knowledge that came out of a method invented 50 years ago it will invent a new method on the fly to answer it.” Just how a knowledge engine will do this kind of work isn’t clear – at least, Wolfram isn’t any more forthcoming. Neither is it clear when we’ll see this capability. It would be easy to dismiss it as the wild-eyed hyperbole of a fanatic, but that would be doing Wolfram disservice. His track record in business and in academia means that his predictions cannot be lightly dismissed. And his latest moves to open up Alpha show that it will still be troubling the established search engines for a while yet. Read previous Innovation columns: Can technology persuade us to stop trashing the planet?, Ultimate jukebox is next step in net music, You Facebook, you Tweet, now lifelog, The psychology of Google Wave, Inside Sony’s broadcast lab, Classic computers on the danger list, Are we ready for the Autonomous Age?, Why do users fawn over Twitter’s failings?