This week - can machines control our brains; a reality check for AI; Waymo car gets confused and tries to run away; why robots are animals, not humans; and more!
More Than A Human
With the advancements in brain implants and brain-computer interfaces, some people are afraid this will enable true mind control via a chip inside the head. But the real science and research say that the biological obstacles of mind hacking are far greater than the technological challenges, as this article from Quanta Magazine explains.
Adding a robotic third thumb to your hand not only changes your hand but also your brain, a new study has shown. This is another example of how plastic our brains are and how they can adapt to incorporate a new body part or sense into itself and make use of it.
Researchers from Intel made an AI that can transform GTA V to look photorealistic. Who knows, maybe this will find its way into the game dev toolkit to make photorealistic games much easier.
In this blog post, Filip Piekniewski dumps a bucket of cold water on the entire self-driving industry. AI community as a whole also gets its fair share of a reality check. The main conclusion of the article is that self-driving cars were overhyped and people are starting to notice this. Companies that previously were focusing on self-driving tech are getting rid of their self-driving divisions, full self-driving is nowhere close to the promises and deep learning hype deflates. But not all is bleak and as Piekniewski states, the real revolution is happening not with more data and compute, but with cheap and fast electronics.
Researchers from ETH Zurich propose a new neuromorphic chip that could bring the low power always-on learning chips to edge computing, robotics or even to be used in brain implants.
It's interesting how a line of orange cones confused one of the most advanced self-driving car out there to the point it got blocked in the traffic. And when the roadside assistance team arrived, the vehicle started to drive away before pulling over and completely blocking a three-lane road.
Researchers from Agility Robotics have successfully trained their bipedal robot Cassie to climb stairs without any input from LIDAR or cameras. In other words, it learned how to climb stairs (and also want on rough terrain) while being totally blind.
We are obsessed with comparing robots to ourselves. But Kate Darling makes an argument that robots are closer to animals than to humans. She also makes an argument that our relationship with them is way closer to that we have with animals than we think. This change in how we see robots could shift how we see robots - not as a replacement but as a helper.