Viv is the work-in-progress, superintelligent, self-learning, voice-enabled agent of San Jose, Calif.-based Viv Labs, whose minimalist website describes her as:
“… a global platform that enables developers to plug into and create an intelligent, conversational interface to anything. It is the simplest way to interact with devices, services and things everywhere. Viv is taught by the world, knows more than it is taught, and learns every day.”
Of course, the other intelligent agents aren’t exactly sitting around wherever they hang out, waiting for Viv’s birth.
“I would expect Siri, Google Now, and Cortana to evolve into this type of solution in time,” IDC application development director Al Hilwa told VentureBeat.
Harry Wang, the director of health and mobile product research at industry research firm Parks Associates, echoed that assessment.
“Viv represents maybe the next milestone in the evolution of A.I. and human-machine interface design,” he told VentureBeat, “but not the greatest leap-forward nor the end-game.
“It may be more advanced, more intelligent, and more efficient than Siri, Google Now, or Microsoft’s Cortana, but I am hard pressed to believe it is far ahead in this race, because its overall approach is similar to [its] competitors’ R&D direction, only that it is ahead by a step or two for now.”
The key to Viv’s unique charm, assuming one exists, may be her capability to learn. Based on the limited information available, Hilwa said that this capability might mean she could “use ready databases and knowledge bases rather than require a massive procured knowledge base be built especially” for her.
She certainly has a lot to learn from Viv Lab’s founders — Dag Kittlaus, Adam Cheyer, and Chris Brigham — who were the original team behind Siri. Apple bought Siri for an estimated $200 million in 2010 and gave it to the iPhone.
IBM’s Watson is also known for its capability to learn, so get ready for future Jeopardy shows in which, instead of wiping out human opponents as Watson did, it battles Viv and its generation.
Without breaking a sweat, today’s Siri, Cortana, and Google Now can open your calendar, read your email to you, or make a reservation at a restaurant for five people at 8 p.m.
But a match between Watson and Viv should include the kinds of inferred logic sentences that humans deal with all day long, and that make developers gasp when computers handle them. As noted in Wired, these requests could include:
- “Give me a flight to Dallas with a seat that Shaq could fit in.”
- “On the way to my brother’s house, I need to pick up some cheap wine that goes well with lasagna.”
Once an intelligent agent can actually learn such things, what’s left to do except — you know — take over the world?
Hilwa pointed out that “self-learning is a very ambiguous term.”
“If it means being able to build new assertions, connections between facts and rules of deduction, then, yes, that has been the end-game of [artificial intelligence.]”
“The question,” he told us, “is how much and how fast.”