CNET también está disponible en español.

Ir a español

Don't show this again

Software

Viv aims to out-Siri Siri in its first public demo

Say hello to the next-generation AI assistant from the creators of Siri. Viv is self-teaching, folds into other apps and crosses platforms. But we haven't heard its voice yet.

When it comes to artificial intelligence and digital assistants, Dag Kittlaus brings a lot of gravitas to the table: he co-created Siri, which Apple later purchased and made the "voice" of iPhone.

And now Kittlaus has co-created what may well be the next generation of Siri: Viv, a new AI startup that aims to replace the app world as we know it with natural language commands.

Kittlaus on Monday provided the first public demo of Viv at New York's TechCrunch Disrupt event, deep within the Brooklyn Cruise terminal in Red Hook. It was the first glimpse of the tech he says will beat Siri, Google Now, Cortana and Amazon Alexa at their own game in a year where bots and AI technology are suddenly all the rage.

Viv had gathered some headlines over the last year or so, but this was the first chance to see Viv in action.

The Viv demo involved talking to a phone and a computer, and at first it performed tasks that Siri, Google and others can do, such describing the weather. Then the questions got harder, like "Will it be warmer than 70 degrees near the Golden Gate Bridge after 5 p.m. the day after tomorrow?"

vivstagedemo.jpg

The Viv stage demo was missing any spoken voice feedback...which lessened the magic considerably.

Dan Ackerman/CNET

Kittlaus ordered flowers, sent money (or simulated it) via Venmo, looked for hotel rooms for a vacation, and ordered a six-person Uber for a ride to Madison Square Garden. Viv pulls up the third-party services seamlessly, making apps unnecessary, Kittlaus said.

The demo received mild applause, but ended up looking not much different than a visualized version of what Amazon Echo can do. Of course, the underpinning of Viv could be a lot more complex. The real magic is dynamic program generation: "When it understands intent of user, it generates a program as needed," Kittlaus said. That makes Viv something that can start building services, in theory. But it's hard to judge how that will actually work during a 20-minute stage demo.

Other things to know:

vivai.jpg

A tree of how Viv parses out spoken information and turns it into information to act on.

Dan Ackerman/CNET

It's multi-platform

In previous stories, Viv seemed different because it's aimed at being a ubiquitous, ever-listening cross-platform type of conversational audio AI. But of course, Google Now runs across many platforms. So does Amazon's Alexa, and Microsoft's Cortana. Siri works across many types of Apple devices, too. Viv wants to be in all things: not just computers, but appliances and other gadgets. In an interview with TechCrunch's Matthew Panzarino, Kittlaus said there are no plans to build hardware, but Viv's talking to hardware makers to find ways to "breathe life" into Viv's execution

It has plug-ins to other things

Like Amazon Echo, Viv works with third-party services, adding app-like skills invisibly to knit capabilities inside. Viv will have a certain number of services by year end, but eventually will aim for thousands, Kittlaus said.

Viv is 'always-learning'

According to Viv's website, the service is "taught by the world, knows more than it is taught, and learns every day." Kittlaus demonstrated on stage that Viv generates software on the fly based on requests, and grows its trees of information over time. But again, that's hard to prove in a short demo.

Now Playing: Watch this: Viv wants to beat Siri at her own game
1:14