Homepage > Apple > General

IBM’s Watson is a genius compared to siri

IBM Watson1
When Apple launched iPhone 4S with the voice-based personal assistant Siri, the world was mesmerized by the idea of Siri and this has popularized the concept of digital voice-activated assistant – in the real world, Siri wasn’t so hot (check our iPhone 4S review), but this is another story… Since then, many other vendors, such as Samsung, have tried to leverage their own voice-based assistants, but Siri is still considered the top-of-the-list by many.

However, that may change soon now. Compared to Siri, IBM’s Watson artificial intelligence system is a rather mega-project which has been designed to crunch massive amounts of data and act on them. Watson made a name for itself by winning the Jeopardy! TV show. Apparently, it can now answer voice questions too.

Today, Watson is not deployed for mass usage because for now, it is a giant which consumes astonishingly huge amounts of data, requires a fairly large power load and needs at least a dozen servers to start its work with. That’s very large, even if you take into account that Siri is also run from a data center as well, but is probably more power-efficient as we suspect that a single server may run one instance of Siri.

But that is fast changing now. According to Bernie Meyerson from IBM, the company has been able to bring down the power consumption levels of Watson. He uses the term ‘dropping like a stone’ to describe the descent in the overall power usage of the voice-based tool.

With such break-throughs happening at IBM, it shouldn’t be surprising if some time from now, the company is able to crunch Watson down to a point where it could process millions of requests for the general public. Although IBM has targeted only enterprise clients so far, the rapid portability of Watson may entice the company to consider other options, and IBM may provide this service for all handset makers to use, or as a n an app… (and reap the advertising benefits)

Like Google+ Tweet Pin
User Comments