Getty
At TechCrunch Disrupt on Monday, Kittlaus finally demonstrated Viv to the public - and it makes Siri look antiquated in comparison.
Siri's problem
Let's first get one thing out of the way: Siri is pretty awful to use. Much of the time, asking Siri to do anything feels like you're just speaking into a Google search. It only has a few functions it can do well, and feels closed.
The main goal of Viv is to give you the opposite feeling: to make you feel like Viv connects to every service in your life.
The war
Kittlaus says we are the middle of a war over who will build the best universal AI assistant. And here are the three key factors he thinks put Viv in a good position to win:
- It's personalized to you.
- It's available in any device.
- "Most important": It's powered by every service.
This last point is where Kittlaus says Viv will really shine. Siri-style assistants can do 20 and 30 things (arguably) well, but Kittlaus wants Viv to eventually do hundreds of thousands. He wants Viv to connect to services like Uber, ecommerce, weather, and so on. In short: any service that wants to integrate with Viv should be able to.
Software writing software
Kittlaus says Viv will push toward universal integrations by using "dynamic program generation": basically software that writes itself. Developers will be able to "describe" or "model" what they want their Viv integration to do, and the computer will do the rest. This makes it much easier and faster to integrate with Viv than with other AI assistants, Kittlaus says. You don't have to hire a whole developer team to build and maintain your app.
TechCrunch Behind the scenes.
Your personal shopper
Kittlaus gave a demonstration of Viv's powers with a few e-commerce integrations on Monday. This is what happened:
- Send Adam 20 bucks for drinks last night pulled up Venmo.
- Send my mom some flowers for her birthday pulled up Pro Flowers.
- Get me a nice room in Palm Springs for Labor Day weekend pulled up some options from Hotels.com.
It was impressive.
Business Insider Viv's weather integration.
Beyond an app
Kittlaus' vision is to have Viv live everywhere: your phone, your car, your smart fridge. It will follow you wherever you go to whatever device you happen to be on. And you'll never have to download new apps. Once Uber has built something for Viv, for example, anyone using Viv will automatically connect to it when they ask for an Uber. It lives in the cloud.
The way Viv is set up feels much more open than Siri, and closer to the way Amazon's Alexa assistant is set up, specifically with the potential that services can program an integration in the way they would make an app - and that doing so would be even easier from a technical perspective.
The power of the default
But there is one thing that should worry some companies, and it was abundantly clear in the demo: the power of being the default. When Kittlaus talked to Viv, it consistently surfaced something specific from his general commands.
Sending money tapped Venmo, get a car tapped Uber, and sending flowers tapped Pro Flowers.
And Kittlaus called these companies - Venmo, and so on - "our friends" at XYZ. No doubt when referring to these companies as friends, Kittlaus was just being polite to partners in a beta program, but it raises an interesting question.
TechCrunch
Viv defaults to Venmo.
Who decided that Uber was what Kittlaus meant when he asked for a car?
This debate has a long history with regards to things like Google search rankings, but feels more pressing when we are talking about AI assistants. Part of the reason assistants are useful is that they make a bunch of small decisions for you (like which car service to book). And I'd hazard a guess that in many situations, people will just go for the default, for whichever service Viv brings up. Viv would be a platform with a much larger percentage of "I'm-feeling-luckys."
And that would give Viv, potentially, a lot of power over the services it draws from.