I may have made some bad decisions recently...
I hear there's a spacecraft bound for Jupiter needing a computer..
IBM has said that it wants to throw more than a billion dollars at powerful cognition overlord and Jeopardy!-playing human crusher Watson... so that the supercomputer can become a new business unit. The Watson Group will be headed up by Big Blue's former senior veep of software solutions Michael Rhodin and will try to milk …
Or VASA for short.
Your little gadget, pc, etc, has a voice recognition program running, with a library of commands it can instantly recognise. Once the instructions get unintelligible to your local machine it passes the phrase/sentence to Watson, which then passes back something your machine can work with.
eg. "Where on Earth are my car keys?" Watson replies, "Find car keys" PC then looks through its recent NSA surveillance videos and says "where you left them"
Assuming the same algorithms. AI, even 'good enough for a first-line at a call center' AI as Watson is intended to be, is very hard. Experts have devoted their career to trying to advance the field. The slowdown you notice probably means they sacrificed some speed in order to improve accuracy of answers.
Although the supercomputer fired up the public's imagination with its Jeopardy!-winning antics, IBM has appeared to struggle to find a way to turn the technology into a business.
I'm assuming they didn't try "Watson - What would be a good business plan for a computer which can search through data very quickly, and answer questions with natural language responses?"