Is Tesla's Most Recent Autopilot Announcement An Improvement Or A Diversion From The Real Issue?

Tesla Motors Inc.’s entrance into the race to build a fully self-driving car has the potential to scramble an already contentious debate about whether autonomous vehicles should be allowed to operate alongside traditional autos on U.S. roads.

The company said last week that it will begin equipping all of its cars with hardware that will allow for fully autonomous driving, despite the fact that its existing “Autopilot” feature already has been involved in a crash that’s believed to be the first fatality of a driver using a car driving in a semi-autonomous mode.

Read Article

TheSteve - 10/28/2016 12:14:26 PM

+1 Boost

Autonomous cars will likely become a consumer reality in the not too distant future, maybe in less than a decade. Google, Tesla, and others believe this, and they are already well into the research, development, and iterative testing cycles.

Tesla's errors were two-fold:

(1) Releasing an early Alpha-version (that's pre-Beta) of their autonomous driving product to the public. They knew their product was nowhere close to being fully autonomous, but it could do some cool stuff NOW, such as *appearing* to be autonomous under some conditions, for a time. Tesla believed that by making consumers Alpha-testers, they could gather data on their product and bring it to a full production (fully autonomous) state much quicker than by doing all the testing themselves in secret. Tesla clearly warned users that they must be in control of the vehicle at all times, but some users did not heed that warning and got into trouble. Some even assumed the system was fully autonomous, and they lost their life because of that.

(2) Tesla chose to use their autonomous product's name -- Autopilot -- when they released the early Alpha version to consumers (AKA the "guinea pigs"). As a fully autonomous system, Autopilot is a solid name, but for an early Alpha version that can only reliably provide a number of "driver assist" services, this name is arguably inappropriate and misleading to consumers.

So when Tesla announced Autopilot will one day be fully autonomous, that is a sort of half-truth. The not spoken half of the story is that Autopilot was intended to be fully autonomous from Day-1; Autopilot was released to consumers in an early Alpha state, and Tesla drivers using it are acting as unpaid guinea pigs for Tesla's R&D department.

nguyenvuminh - 10/28/2016 1:48:02 PM

+2 Boost

Improvement or Diversion? A little bit of both I imagine.

HenryN - 10/28/2016 2:58:44 PM

+3 Boost

Diversion is a 2AM tweet by Musk hinting the next "killer" feature, or press releases by MB, BWM, ... announcing the next "Tesla Killer" by 2020 (or 2025, so to speak).

Spending thousands of dollars on hardware on every car being made today is obviously NOT a diversion. It would be an awful waste of money if it's just a diversion since a free tweet would be just as effective.

mpls - 10/28/2016 4:09:32 PM

+2 Boost

2 deaths ? I thought there was one in china too..

SanJoseDriver - 10/28/2016 4:34:23 PM

+5 Boost

I only know of the Florida one where Autopilot use during the time of the accident was confirmed. Also, as stated many times that was not a freeway where the system was allowed to be used. Deaths where instructions were followed = 0.

There are now well over 300 million miles of autopilot logged with 1.5 million more each day. On average there is 1 death every 60 million miles for manual driving. Even if there were 3 it would still be twice as safe as manual driving. If you follow instructions, then then you can't quite calculate the stat since you would be diving by 0. Probably won't stay that way forever, but the longer we wait on self-driving tech, the more people will needlessly die.