I knew you did…just making a point about a company that has done a lot of research into nothing projects.
You have much greater confidence in the altruism of companies than I
I’m not saying the company is altruistic. I’m saying sometimes there’s a far reaching technical reason that may benefit them in the future.
I worked at DEC for 10 years. When I was there they had enough cash on hand to start a small bank. I was on a couple projects that no one knew would ever be sold. But it was more of a what-if project. At the time DEC had a lot of money to burn. They were reaching into many different areas to try to expand.
Hard to believe they ultimately ended up acquired by Compaq, which then got absorbed by HP. We've lived through a lot of history.
I still have several friends still working at HP that started with DEC. When I was there they were the second largest computer company in the world…they were the largest builder of mini-systems…80% of all networks around the world were DEC…and they built the fastest computer chip in the world (1000 times faster the Intel’s fastest chip at the time).
I do not like the idea of driverless cars but I accept that they will be a reality in my lifetime. That said, they MUST learn and adapt to the environment with driver piloted cars and the sometimes irrational ways we drive. I imagine Google sees these cars as fully networked. It is in their DNA, they can see it no other way. That will allow EVERY robot car to benefit from EVERY other robot car’s experience of irrational driver behavior IF Google identifies the opportunity and creates the algorithm. The car sees, learns and shares with the collective of Google cars. That makes them better able to handle real people driving real cars. Considering they do this now with you search history data, why not robot cars?
Having worked for a few large companies in both development and R&D engineering, I can tell you Mike is right on the money. When it comes to R&D, most of the endeavors fail to be completely successful. Bits and pieces may become lucrative on their own or the learnings are valuable company assets in IP but on the whole, the vast majority fail to achieve the original goal. They know that. This is where the grand slams happen and it’s worth the risk.
$1B invested? I assume that’s over the life of the program. Pin money to Google. Most technology firms I have worked at, or are familiar with, spend on the order of 10% annual revenue on R&D. Google cleared $50B revenue mark back in 2012. Leaves plenty of room for their other endeavors as well.
One other thought- why do most people assume the automated system needs to be 100% faultless before it is viable? We already accept that humans are abysmally worse than that and the staggering financial burden is already accounted for through established means. I have a suspicion that Google could easily afford to be self-insured for the odd mishap…just curious why we feel machines must be perfect when the current system is not- by a long shot…
Google’s self driving cars affected me last week. Seriously. I was considering driving over to Mt View the other day to a restaurant I like there, for lunch. Then I got to thinking I might encounter one of those self driving cars on the Mt. View roads (where Google is located), so I had lunch in a different restaurant, closer to me, in San Jose instead. Self driving cars are a little too creepy for my sensibilities and I’d prefer not to have to deal with them at all.
Driverless cars won’t need any technology we don’t have now and no redesigned roads. We already have high end cars that will take over for accident avoidance. GPS and radar can provide all the guidance the system will need.
Reliability at a mass market price will be the problem, Did you see that Jeep has just had a problem with the cars applying their own brakes for no reason. Can anyone say Takata air bags?
I read about that Jeep problem. It appears in tests that the avoidance systems are overreacting to things they should not be reacting too.
The danger that concerns me the most is that the government might mandate these systems before they’re ready… and before they can be added without pushing the price of cars too much higher. Avoidance systems won’t work if too many people are continuing to drive their old beaters because they can’t afford the high tech new cars. I’m not confident that the economy can support these new systems. But the federal agencies are too often not sensitive to cost.
Wes, that day is on the horizon. Cars can and will talk to one another very soon.
We’re there already. I saw a project on PBS where scientists are working on drones with the ability to interact and behave like a pack. They showed six drones operating in that manner.The scientists in the show said that the algorithms get too complex with more than six drones, but next year they might not be. Cars are definitely on the verge of talking to one another.
“Wes, that day is on the horizon. Cars can and will talk to one another very soon.”
Only if they outlaw cars that pre-date the technology change…in which case, I probably won’t be posting here, as I don’t think they allow internet access in jail. “You can have my steering wheel when you pry it from my cold, dead hands!”
I don’t think cars will need to interact directly with many more than the 6 closest vehicles anyway. Those cars can interact with another six, and in that way there is indirect interaction with a lot of vehicles. In 60 years when I’m too old to drive, I could probably use one of these things. Until then, I prefer to drive myself. Except during my horrible commute.
The autonomous cars will “sense” the old coot… who will probably be me… and react accordingly.
“Dance” is a good term. The difference between the new cars and the old coots will be that the new cars will be able to “dance” around the old coots… and they’ll have to, because we old coots won’t be able to “dance” around the new cars.