Raw LLM Responses

It isn't said explicitly in the article, but from what I heard elsewhere part of the reason why it had so much trouble classifying the obstacle, was because it was a woman pushing a bicycle. Obviously this would look a lot like a person riding a bicycle, but behave somewhat differently. So the car thought "oh, there is a bicycle up ahead, but bicycles are fast so it will be out of my way before I get there." and than got confused when it didn't. Obviously when in doubt what exactly is up ahead the car should have slowed down instead of reclassifying it in order to predict what it would do. It seems pretty sure that Uber's cars do know how to deal with that sort of situation nowadays. The main issue in my opinion is not wether self-driving cars are safe. This accident notwithstanding they are going to be safer than humans within a short time if they aren't already. It is the speed at which a number of companies try to get to this point. We are on the brink of something new. Everyone is trying to be the first to being a fleet of selfdriving cars to the market, because in the end there won't be enough room for more than a small handfull. The big companies like Uber and Waymo and Tesla and everyone else all want to be the ones who hire out their self-driving cars to the public. There are billions perhaps trillions to be made there. But nobody will have more than two or three apps on their phone to hail a self-driving car. There is not enough room for the several dozen competitors who are working towards this now. Only the first few will see a return on their investment and the others will lose their money and if they business relies on making cars or serving up ride shared they will lose their entire business too. Large companies will go bankrupt over the next decade because they lost out in the race to be one of the two or three self-driving car companies that survived. In that race to dominate the industry, corners are being cut and safety issues are being ig