3 ms·
In a recent Ted Talk[0], they detail that their goal does not align with that of their competing manufacturers. Manufacturers go for a quick-to-market, good-eno
by espadrine 11y ago
In a recent Ted Talk[0], they detail that their goal does not align with that of their competing manufacturers. Manufacturers go for a quick-to-market, good-enough autopilot that requires human supervision.
Google's driverless car won't have a wheel, and therefore has a significantly smaller margin of error. Their reasoning is that rebuilding trust is a lot harder than building trust; catastrophes will make the news, unless avoided altogether. In light of recent events (the VW scandal), it is far from absurd.
[0]: http://www.ted.com/talks/chris_urmson_how_a_driverless_car_sees_the_road http://www.ted.com/talks/chris_urmson_how_a_driverless_car_s...
- qq66 11y agoUnfortunately, the most rushed-to-market self-driving car can ruin trust in them for everyone, which would cost many lives. This is where regulation is critically important to make sure that any self-driving cars allowed to be commercially sold are thoroughly validated.
- TeMPOraL 11y ago> Their reasoning is that rebuilding trust is a lot harder than building trust; catastrophes will make the news, unless avoided altogether. In light of recent events (the VW scandal), it is far from absurd. I think this only works if you're a new player, and even then not always. You bring up VW scandal, but do you honestly expect VW sales to drop in any meaningful way? For most of people, company ethics are at the bottom of the list of concerns when buying something - that is assuming that they even know that the company did something wrong. Traditional automakers have a pretty good shot at dominating the self-driving market; they may launch a somewhat buggy system on expensive models, and then iron out the kinks. By the time your average Joe will be driving such a car, they'll could have most bugs fixed, with the added benefit of being recognized by general population as a well-established self-driving cars provider with years of experience. A good bet for Google, I think, would be to design a perfect autonomous rig and then licence that tech to existing manufacturers.