In The Atlantic Cities, Eric Jaffe shares his experience riding in Google’s self-driving car as it takes to the city streets of Mountain View, California. It’s a fascinating look at the technology underpinning the 700,000 miles driven by Google’s fleet of self-driving cars. The article is also sobering, because putting self-driving cars on the road for consumers will take years, despite Google’s advances. Nissan is aiming for 2020, but robotic car consultant Brad Templeton says that the most advanced of the car manufacturers is Mercedes, whose self-driving car project isn’t even where Google’s was in 2010. This could mean self-driving cars will take much longer to go mainstream, or that the automotive industry is ripe for major disruption. follow link
Open Files with Finder's App Switcher
Say you're in the Finder looking at a file and you want to open it with an application that's already running but which doesn't own that particular document. How? Switch to that app and choose File > Open? Too many steps. Choose Open With from the file's contextual menu? Takes too long, and the app might not be listed. Drag the file to the Dock and drop it onto the app's icon? The icon might be hard to find; worse, you might miss.
In Leopard there's a new solution: use the Command-Tab switcher. Yes, the Command-Tab switcher accepts drag-and-drop! The gesture required is a bit tricky. Start dragging the file in the Finder: move the file, but don't let up on the mouse button. With your other hand, press Command-Tab to summon the switcher, and don't let up on the Command key. Drag the file onto the application's icon in the switcher and let go of the mouse. (Now you can let go of the Command key too.) Extra tip: If you switch to the app beforehand, its icon in the Command-Tab switcher will be easy to find; it will be first (or second).
Google’s Self-Driving Car Takes to City Streets
* Many technologies aren't designed with reliability in mind because it doesn't really matter. No one gets hurt if you have to reboot your smartphone once a week. But other complex engineering efforts like bridges and airplanes are incredibly reliable. And while there are safety recalls, they're generally prompted by a mere handful of actual problems - they're an example of the system working to enhance reliability, not an example of the lack of reliability.
* Self-driving cars are likely to be more efficient than human driven cars, because the algorithms ensure things like smooth braking and steady acceleration, which humans are remarkably bad at. And of course, by the time we see self-driving cars, it's likely - or at least possible - that most automobiles will be electric. While electricity from the grid isn't necessarily clean, it can be.
* As far as "human tasks" go, I think driving is a pretty lousy one. We're not talking a sport here; mostly you sit very still, moving your arms and feet in tiny motions, with the constant threat of injury or death if you make even a small mistake. Cars are one of the leading causes of death in the world; anything society can do to reduce such horrific and unnecessary deaths seems like a good thing to me.
We have seen Auto-pilot features in Airplanes for a long time now and but it doesnt mature as fast we expect. It still needs some amount of human intervention and supervision. I expect a similar aspect from the self-driving cars as well. However, the scenario could be much more complex in case of cars with the traffic on roads.
I don't know about you but that seems like a pretty good track record. I know I have not driven 700,000 miles with only a minor fender bender as a result.
finally none of these incidents resulted in ANY injuries. Over 40,000 people die and countless others are injured every year in car crashes. Humans are not perfect and we can not sense even a fraction of the things that are going on constantly around our vehicle, where a computer can. Isn't it time we eliminate the weak link in this scenario?
There's no way to go from 0% self-driving cars to 100% self-driving cars without passing through all the percentages on the way, and that's absolutely being taken into account by everyone who's working on it. Initially self-driving capabilities will probably be limited to freeway driving, much like we use cruise control today. And slowly they'll be used in more complex situations, but it will take many decades before human drivers are out of the picture.
BTW, thanks for pointing me to this article.