Tesla's autopilot tech has often been criticised for its lack of accuracy and reliability. Many critics don't even consider it as self-driving tech, but Elon Musk has also been very gung-ho about the technology. A couple of months ago, it released a new self-driving beta to select users which have been trained on the Dojo supercomputer.
Now a video has emerged of a Tesla running the full self-driving update in beta which shows how good the system is. It can detect a deer and navigate around it avoiding a collision with it, even though it still needs some level of driver monitoring.
The driver has to enter the destination in the navigation computer and then the car will attempt to autonomously drive to the destination in question. The new system does try to replicate more real-world scenarios and that's why the beta has been deployed to users to add more training data from the road and to deploy a more robust neural network for a wider amount of users.
A video by YouTube Dirty Tesla chrysalises this. In the video, you can even see the driving visualisation of the full self-driving beta as it detects the deer, slows down and plans a new path to avoid it. This is not the first time a Tesla a skirted around Deer when on autopilot but it for the first time it has achieved this with utter confidence.