On 7 May 2016, Joshua Brown was killed as his Tesla drove underneath an 18-wheeler on US 27-A highway in Florida.  The truck was making a left turn from the westbound lanes across the eastbound lanes when the eastbound Tesla Model S struck it.  

The Model S was in Autopilot mode at the time of the crash.  In a blog post, Tesla Motors explained the incident:

Neither Autopilot nor the driver noticed the white side of the tractor trailer against a brightly lit sky, so the brake was not applied. The high ride height of the trailer combined with its positioning across the road and the extremely rare circumstances of the impact caused the Model S to pass under the trailer, with the bottom of the trailer impacting the windshield of the Model S.

Tesla downloaded data from the car's computers, which is evidently their basis for concluding that the car did not "notice" the side of the trailer.  How they concluded what the driver did or did not notice is unclear.  Other reports suggest that the driver may have been watching a Harry Potter movie on a portable DVD player at the time.

(Cue the witty remark about Voldemort working for General Motors.)

Among other things, the incident illustrates a dilemma of progress.  Tesla's permissive strategy for innovation involves marketing features early, gathering data, and improving them on the fly.  As Tesla's blog post states, its Autopilot software is in a "public beta phase", with customers as the guinea pigs.  In so doing, Tesla has gained 130 million vehicle miles of data on the performance of its system.  Presumably, it will work on fixing the issue that led to this crash.

By contrast, Google has been following a precautionary strategy.  That is, it has been testing a fully self-driving vehicle under expert supervision, which has not been released for public use.  As a result, it has obtained "only" 1.5 million vehicle miles of data.  It may be several years yet before their car is ready for market.

Johsua Brown was reportedly a great Tesla enthusiast and, it seems, approved of that company's approach.  The crash certainly throws the dilemma into relief: Which approach is better in this case?

Blog topics

  1. 2018 (17)
    1. December (2)
    2. October (2)
    3. August (2)
    4. July (3)
    5. June (3)
    6. May (2)
    7. March (2)
    8. January (1)
  2. 2017 (44)
    1. December (1)
    2. August (3)
    3. July (4)
    4. June (6)
    5. May (5)
    6. April (3)
    7. March (6)
    8. February (8)
    9. January (8)
  3. 2016 (95)
    1. December (7)
    2. November (13)
    3. October (13)
    4. September (15)
    5. August (20)
    6. July (18)
    7. June (9)