Finally For a year and a half, two inflatable white models each equipped with five additional cameras and one palm supercomputer have been quietly roaming around San Francisco. In a city and an era rocked by questions about the capabilities and limitations of artificial intelligence, the startup of modified teslas tries to answer what the value of simple software:
The startup, which is making its operations public for the first time today, is called hyprlabs. Its team of 17 (only eight full-time) are divided between Paris and San Francisco, and the company is releasing the Amazon company in 2018. Ultimately, it plans to build and operate its own robots. “Think of the love child of R2-D2 and Sonic the Hedgehog,” says Kenntley-Klay. “It will define a new category that doesn’t exist right now.”
For now, however, it is starting to announce its own software product called Hyprive, which is phrased as a breakthrough in self-driving cars. These kinds of leaps are in the robotics space, thanks to advances in machine learning that promise to reduce the cost of autonomous vehicle software, and the amount of human labor involved. This emergence of training brings a new movement to the space that has been in the years of “disappointment,” as the technology developers have failed to realize the funding they have been denied to use robots in public places. Now, robotaxis are picking up passengers in many cities, and automakers are making new-fangled promises to bring self-driving cars to consumers.
But using a small, old, and cheap and cheap team to get from “driving well” to “driving more safely than a human” is its long hurdle. “I’m not going to tell you, Unite at heart, that this is going to work,” Kenntley-Klay said. “But what we have built is a really strong signal. It needs to be said.”
Old technology, new tricks
Hyprlabs ‘Technique Training Technique From Other Worlds Robots’ teaching methods to teach their programs to drive themselves.
First, some background: For years, the biggest battle in autonomous vehicles seems to be between those who use cameras for their software! But beneath the surface, there is a huge philosophical difference.
Camera-only fans like Tesla wanted to save money while planning to launch a large fleet of robots; For ten years, CEO Elon Musk’s Plan has been to suddenly switch all of his customers’ cars to self-driving with a push of a software update. Upthiside these companies had a lot of money and a lot of data, as their self-driving cars collected images wherever they drove. This information is compiled into a form called “End-to-End” machine learning with reinforcement. The system takes pictures-a bicycle-Issuing driving orders-Move the steering wheel to the left and go easy on the acceleration to avoid hitting it. “It’s like training a dog,” said Philip Koopman, a self-driving car software and security researcher at Carnegie Mellon University. “In the end, he says, ‘A bad dog,’ or ‘a good dog.’ ‘”
