Raw LLM Responses

>What will those cars do if those lines are obscured by snow cover, or will self-driving cars only be used in states that don't have winter? The issue highlighted in this article really only holds true if car-manufacturers insist on using guidance-systems that rely solely on *visible-spectrum-only* guidance-sensors/cameras. It stands to reason that most manufacturers will favor the solution that costs them the least to manufacture, however, I suspect they'll be using something similar to the ol' [Fight Club cost-of-business calculation](https://en.wikipedia.org/wiki/Talk%3AProduct_recall). With redundant *overlapping* sensor-arrays (which really aren't that expensive to manufacture these days, relatively speaking), a vehicle that has trouble "seeing the way" on crappy roads could get its guidance-information bolstered **substantially** by GPS, for example, among many other technologies already essentially commodified. While it is true that current *civilian* [GPS capability](https://en.wikipedia.org/wiki/GPS_Block_IIIA) is typically limited to an accuracy of approximately 10-13 meters (approx. 32-42 feet), plans have already long been in-motion to [upgrade civilian-accessible GPS-capabilities](https://en.wikipedia.org/wiki/GNSS_enhancement) to 20-30 centimeters (approx. 8-12 inches) of so-called "absolute" accuracy. Coupled with now-inexpensive, highly-available multi-axis inertial and other kinematic sensors (like those now found in numerous "drone-type" aircraft these days), the resultant positional accuracy can be boosted to less than 10 centimeters (approx. 4 inches). Going further, automotive "vision-systems" can be equipped with hyper-spectral or multi-spectral sensors (which are also on their way to becoming a commodity-item), enabling computer-systems in vehicles to easily see through even the most problematic of environmental conditions. Multi-spectral and hyper-spectral "camera-sensors" can "see" right through dense f