This little phone AI has several wild ideas

Just a year ago I had my first demo Humane Ai Pin – which was also my last, as it turned out. But the next AI gadget at this year’s MWC tries to start, where Humane crashed and burned, and in a way it is even stranger than AI Szpilka. It is a phone that captures a lot of information about you, both the past and the present, and uses them to create your own avatar AI to act as a virtual assistant. This is part of Rabbit R1, partly Gemini’s assistant, partly science fiction. And yes; There is also some blockchain. I told you, it’s wild.

Rear team A new criminal you have is based in Korea and is created by creating Blockchain -based vaccine verification method widely used in the country. At the beginning of my meeting with the founder of YT Kim emphasized for me one thing: “We have never sold cryptocurrency.” They are now clearly aware of the bad vibrations around blockchain.

AI Newnal phone trains the AI ​​model YouSo I need a lot of information. The point is that you go to download the history of personal data from websites such as Meta and Google, as well as things such as medical documentation and financial data, and then transfer them to AI Newwal. Kim claims that the company encrypts this “personal chart of knowledge” and spreads it into pieces stored on many cloud servers of third pages. Theoretically, only a data owner can access it.

Are there huge consequences of privacy in all of the above? Absolutely. For his cases, Newal has published an unusual amount Complementary documentation and technical descriptions On your website to be see-through about what he does. I can’t check if the company’s methods are unthreatening, he says, this is just the first look.

But the rabbit hole is still developing: all this information, as well as your appearance and voice, is used to create a moving, talking version of yourself, which is on the main screen of the phone. You call it by pressing the square button on the side of the phone, and then ask to do things. There are all ordinary things that we are used to in these demonstration versions: it can facilitate buy a novel pair of earrings or developing E -Mail. Kim and his team showed me some of these demonstrations and everyone went as expected – and then we reached the car insurance.

The request was elementary: facilitate me buy a car insurance policy. I saw on the screen because it seemed that the store’s comparable rules. On each of these steps, icons seem to indicate where the model receives specific information. The rules were taken, and the next request was to complete the necessary forms to buy a policy. And so; I saw how it passes through the pages of the form on the Geico website and fill them in.

Kim says he really I bought a used car That they could go through this demo, or bananas. And it certainly looks like it worked; I watched the model complement each step, showing its sources of information along the way, and finally paid for the policy. It was one of the wildest technological versions I’ve ever seen, and I’m still not sure if I believe that my eyes worked.

All this is packed with a elementary, but futuristic petite phone, which according to who he was inspired by the iPhone 5S. It is a petite black rectangle with a separate upper screen on which your artificial intelligence sits. Nowaral claims that he runs the “hybrid” of his own operating system and Android. The versions I saw were still prototypes, but Kim claims that the company plans to launch a phone around the world on May 1 for USD 375. He will send to customers for two months after he enters into the pre -entry.

Time may remind me of the humanitarian contact of AI, but the whole thing gives great vibrations R1. This is an attractive petite gadget with a surprisingly low price that creates huge promises. We all know how it went with the launch of R1 and I am sure that experience contributes to the skepticism that I feel about what Newnal is doing. But unlike when R1 was first exposed, the demos I saw were convincing and the device ended all the demands. But “Can AI complete car insurance forms?” And “Do I feel that I can trust artificial intelligence to complete the car insurance forms?” These are two different questions. The same applies to “Can I train the AI ​​model on all my private data?” And “Should I train the AI ​​model on all my private data? “Yes, on.

Regardless of whether any of them really works except for controlled demonstration verses and whether these are good ideas, I have to admit Kim and novel recognition for an attempt at something brave. Elsewhere, in Mobile World Congress, we place DSLR camera lenses on phones, adding more hinges to the screens that already fold, and trying to make skinny phones even thinner. It doesn’t quite set fire to the world. But if nothing more, Newwal is an idea – regardless of whether it is good or not, it is something that the company will have to prove outside the conference demo. And if Rabbit and Humane showed us anything, that at a technology conference at a technology conference he will not replace the shipping of the product that works.