Self-operating vehicles have a huge potential to upset travel and shipping and blur the distinction between transit and last-mile.

To say nothing of the generic bipedal utility robot. CV is the big roadblock there. [1]

No, these things aren't going to happen tomorrow. But they have a fair chance of being delivered before we could actually put a human on Mars or eradicate poverty in Africa, even if massive societal projects were undertaken with those goals in mind.

And, conveniently, CV won't require those massive societal projects. It'll get them, eventually. To enable the really cool stuff. [2] Sort of like cars changed everything even before we had a massive highway system that unleashed further potential.

[1] Though power density is a big challenge, I don't think anyone's going to mind if version 1 of their autonomous maid/handyman only works in one-hour bursts. Not so long as it can actually do the housework. If someone's keeping up with things day-in, day-out, it's rare that more than an hour or two's-worth of work needs to be done in a day.

[2] e.g. Going from self-driving cars and trucks to self-driving transport pods that autonomously link to tracks and ad-hoc caravans for long-distance high-speed travel, then de-link and queue for self-operating cranes to be load them onto self-piloting ships to deliver them to foreign ports where they operate the same dance in reverse.

Self-operating vehicles will be the biggest disruption of the next decade. This is an area where the government will need to get involved (since so much regulation will need to change), and they now have an opportunity to move things forward tremendously or hold progress back. It will be interesting to see how this turns out.

We already have "self-driving cars", as well as "self-driving transport pods" that can do all of the things that you describe. Assuming, that is, that you widen the definition of "self-driving" to include having a human being at the steering wheel.

Self-driving cars would make my commute wonderful, but won't be materially different than having a good bus route. The only real difference I can see autonomous vehicles making is to remove a certain class of jobs and creating another class of jobs.

Likewise, you could hire a "generic bipedal utility robot" pretty easily today. It's expensive, but that's mostly a political problem. (For the humor-impaired, that last clause was a joke.)

> "Self-driving cars ... won't be materially different than having a good bus route."

By this definition a cell phone is not materially different than having a personal land-line available regardless of where you live, work, eat, play, etc. That seems like a fairly silly thing to say when you compare the two. One because it points out the absurdity of the notion of perfect coverage/perfect availability of a 'good bus route' and two because there's obviously much more you can do with a cell phone because it's everywhere, that not even a wired phone at every destination could achieve. (e.g. the entirety of pocket computing)

Similarly, you see no societal advance in bringing something only the very wealthy can afford (personal servants) to the middle class? (initially, and then trending down with commoditization)

That's nearly the definition of massive social change: making the quality of life of royalty affordable to the middle class.

What do you mean by that. If you mean progress, then that has occurred this year, basically the object recognition problem looks like it can be solved by neural nets (google brain cat detector and the latest imagenet challenge in pascal voc).

If applications, then they are literally immense. Robots that can perform any manual labor, thus increase in mnaufacturing ability (plus ability to conduct scientific experiments) by an immense amount. Marshall Brain has good articles about what computer vision technology can enable, including the short story Manna. Computer Vision is easily the most disruptive technology since the invention of the steam engine (it multiplies the work capacity of humanity by that much). It's a new gold rush bigger than the internet (robocars and humanoid robots including Baxter from Rethink Robotics are signs of that beginning). Startups should be all over it, web 2.0 and social networking are miniscule in comparison.

Where I come from 16% accuracy does not mean the problem is solved. Yeah, its a big step up from where we were, but the cat detector neural nets by no means solved the object recognition problem.

It was 85% accuracy on imagenet of pascal voc, using just a couple of computers with nnets vs 1000 for that google brain. And both of these were just experiments run by grad students, a proper industrial effort would solve it.