Cake
  • Log In
  • Sign Up
    • Google started their self-driving car project in 2009 and up until lately were projecting a limited but fully autonomous taxi service by the end of 2018. I'm seeing more doubters now because the problem is so hard to solve. For example:

      On the other hand, Voyage seems smart to go for big retirement communities where the roads are more predictable and the speeds are slower. It seems like Voyage is introducing their service at the same time as Google, but tackling a much more solvable problem, but still a big one.

      Little companies like Microsoft beat big ones like IBM by tackling the easy problems first to get a foothold. Any predictions?

    • I think there's definitely something to be said for constraining the scope of the problem in order to achieve usefulness quicker. But constraining scope also limits opportunity, and what the big names in the self-driving space (Waymo, Uber, Tesla) really want is to be the first to claim their slice of the pie, because whoever gets there first gets the largest slice.

      But full self-driving is such a complex problem that it seems to suffer from Zeno's paradox: the effort required to solve it scales non-linearly as we get closer to a solution, so the closer we get the slower we go, and the solution never quite seems to be within reach.

      I'm skeptical that non-constrained full self-driving can be achieved within the next ten or even twenty years. But I'd love to be wrong!

    • Yeah, Tesla's anti-lidar stance is really bold given that literally everyone else is using it. But I think Elon is right: if you can't get self-driving right with vision alone, it seems unlikely you'll be able to solve it with lidar either.

      Lidar is a shortcut that can get you close to a solution more quickly than vision will, but there are still plenty of problems it can't solve for you, and solving those remaining problems will still require being good at computer vision.

      Elon's bet is that by doing the hard work to get good at computer vision up front, Tesla will be more likely to ultimately achieve reliable full self-driving before its competitors even if the competition initially gets off to a faster start by relying on lidar. A marathon approach vs. a sprint approach.

      πŸπŸ’πŸ‡

    • I could be coming from a totally wrong direction, but has anyone even thought "why even bother?" Because if self driving is the goal, the public roads infrastructure needs to be totally overhauled, in my opinion, in addition to the vehicles and creating this monstrous AI. So again, I want to understand what is the point of giving up the freedom, and yes for some even exhilaration & pleasure of driving a personal vehicle? Just so someone can get from A to B while pecking on their mobile or reading or sleeping? Won't efforts be better spent toward modernizing infrastructure so that public, elegant and comfortable transport would become so ubiquitous to make personal vehicle driving merely a hobby?

    • I think infrastructure is unlikely to change until it's forced to.

      Cars had existed for years before anyone bothered to start paving roads, building gas stations, or designing cities with cars in mind. And I think self-driving cars will also have to exist for years before infrastructure begins to adapt to them.

      So again, I want to understand what is the point of giving up the freedom, and yes for some even exhilaration & pleasure of driving a personal vehicle? Just so someone can get from A to B while pecking on their mobile or reading or sleeping?

      I think self-driving cars will offer more freedom, not less. The idea isn't necessarily that every self-driving car will only work in self-driving mode. Some will work this way, sure, but you'll still be able to buy a car with a steering wheel and pedals that you can drive yourself.

      The main difference will be that you won't have to drive the car yourself if you don't want to. So when you feel like driving, you can. But when you just need to get from A to B and you're not in the mood? Let the car do the work.

    • This appears great, on the surface. But when I really think about all the random factors that exist out there, in the real world on the roads, I am not so sure I'd want one. I had hit a deer once and only survived out of pure luck, but ever since, I had avoided by a large margin several other similar occasions, based on the experience I had gained on how to ride defensively and be more observant on what's going on way out off the road surface, in the bush or trees. That is just an example of negative aspect I see. Another will be a quick monopolization of decision powers being taken away from actual owner and yielded to some higher authorities, such as insurances for example. I think there is allot more to it than meets the eye, and do not think the enthusiasm alone is what's funneling these efforts, but perhaps I am being too cynical.

    • People who live where Waymo tests their cars often say they were initially afraid of them. After a few months however many of them say they feel safer because Waymos don't text, don't drink, don't drive aggressively and don't argue with their their kids while driving.

      When Waymo can show their cars have less accidents per 100,000 miles than human driven cars, I predict a big shift in attitude. People will wish parents got their teens a self-driving car. If they show Waymo taxis are safer than human driven Ubers, I know which ride service I'll hail.

      Here is what it is like to be in a Waymo taxi. The complaint is how conservatively they drive.

    • The technology behind is really awesome. But it's also interesting to note that common modern transportation is being brought up as part of a modern city, and to my mind it's a pity relatively young cities don't get better at that. It doesn't take a rocket scientist to noticed the gridlock, and see that good public transport really is the key to all the road congestion issues, and ultimately making our predecessors life easier with less time and stress in getting from point A to B in the future, which self driven cars won't help resolve. If anyone has ever traveled in Switzerland would understand what a smooth and precise rail and bus system can do.

    You've been invited!