Automobile Conversation Part 1 of 2 ft. Paris Marx

There's no such thing as a self-driving car. There is such a thing as a self-crashing car. If you ever ask someone working on self-driving cars how often they get into accidents, they'll usually answer something like, "Our vehicles have only been in of at-fault crashes in the last two years despite driving of miles." But that "at-fault" part is doing a lot of work. Because, in reality, they would have been in several hundred crashes but were simply found not-at-fault because the vehicle did something that was dangerous and induced an accident but was technically legal. If you then look at how many times the vehicles WANTED to do something that would get them into a crash but the human operator intervened, it's an order of magnitude higher than the crash numbers. If I showed you a person and told you they had gotten in 300 car crashes over the past year and almost gotten into a crash 1,000 more times but someone else in the car stopped them, but were only found at-fault for one or two of them, there are many things you would call such a person but "safe" and "a good driver" are none of them. It's a hammer problem; when all you have is a hammer, all your problems start to look like nails. The IT industry has had as its main application for decades now the creation, collation, and archiving of large data sets for purposes of bureaucracy and projection/prediction. The industry's three greatest tools developed to achieve these ends, fast & efficient indexed databases, heuristically optimized search algorithms, and the neural networks that train and create said algorithms (which I always must point out function nothing like actual neurons), are the pinnacle of the industry, its most profitable endeavor, and the type of software most people in the industry have training in. They are the metaphorical hammer in this case. Tools that were designed for handling databases have been brigaded into driving a car. When you state it like that, the very idea that it would work seems silly.

sudonim2

A significant contributor to increased pedestrian death rates is a corresponding increase in vehicle weight, particularly among trucks and SUVs. 3,000 lbs. compact cars and 6,000 lbs. trucks hit differently at the same speed.