Raw LLM Responses

There are no self-driving cars right now because the technology is not ready. Eventually it will be ready and lane markings will be unnecessary, except for human drivers. Since the 2004 DARPA grand challenge, which saw multi-million dollar university research projects barely able to navigate open desert, technology has advanced to the point that cars can recognize pedestrian, cyclists, vehicle makes and models, and road conditions. The first vehicle to actually complete and win the grand challenge in 2005 was powered by six Pentium Ms. A single top-of-the-line cellphone with a multi-core CPU and integrated GPU is more powerful than that today. The next ten-to-twelve years will see a leap in capability that is even greater, as sensors and processing power continue to improve and drop in cost, as well as the data sets and AI behind them. More important than all of that is the data gathered by cars. The next checkpoint in the autonomous vehicle race is data collection, and it won't matter if the roads have lines when ten thousand semi-autonomous vehicles have driven on a road and dutifully recorded what their human operators have done. Fully-autonomous cars will be able to replicate their actions with millimeter-precision, communicating their actions to other cars while transmitting and receiving updates on road conditions in real time as they change. People betting against driverless cars are the same people who bet against factory robots, Watson beating Ken Jennings at Jeopardy, self-guided missiles hitting their targets better than missiles aimed by humans, or an AI besting a human in Go.