Originally Posted by
Maelochs
Based on human history I can pretty much assure everyone that every problem we have imagined, and more, have long since been imagined by the people who ALREADY HAVE WORKING AI CARS ON THE ROAD, and have for hundreds of thousands of miles.
Pretty long way to bashing something with a rock to being able to open a person's chest, remove the heart, and replace it with another. Think maybe there were a few issues to resolve? Nowadays we have surgeons operating remotely via the internet. No biggie.
We seem not to grasp the things we have proven able to do.
As one poster noted, we are using tiny computers made by many different corporations over a span of years and all able to communicate, surf the web (Just think about that---we could never have a billion different computers in a hundred different languages if two hundred different countries all communicate---while in cars" we say while using our cell phones to do just that ... to such a degree that it as actually a hazard we all complain about. Cognizant disconnect, much?)
Are there problems? Difficulties and unresolved issue, to be sure. Is it complicated? People wouldn't be spending billions and years io it if it were simple.
But based on what we have done in just the past few years --DONE--not imagined or are aware of or denied---as Walter notes, the main problems are commercial and social.
People either want, or don't want, or are neutral regarding AI cars .... and all these tens of thousands of words wasted in a supposed 'examination" of the topic is really just a poll with A,B, or C.
Personally I'm not one to be especially optimistic about solving arbitrary technical problems based on doing amazing things in the past. Each problem has its own unique characteristics. Some problems might seem at least achievable but in fact have no practical solution because of inherent complexities, the laws of physics, etc.
OTOH in the case of driverless cars people are not spending huge sums of money on the idea for no good reason. Billionaires have lots of money for a reason and they invariably want their billions to grow and won't casually invest heavily without compelling proof-of-concept. Driverless will not only happen, it's already happening in environments that are more controlled than the public road network. Getting from where they are now to that is a matter of additional refinement to existing technology and the economies of scale that make it affordable. Inevitable IMO.
I'm personally a software engineer since the 1970s with lots of experience developing microprocessor controlled embedded systems. Driverless cars are certainly among the most complex of those but the issues are dropping away quickly. When I look at the state of the technology today, I see technology that is well beyond a prototype, not ready for deployment, but definitely headed that way.