[Info-vax] Intel junk...Kernel-memory-leaking Intel processor design flaw forces Linux, Windows redesign

[JF Mezei]

On 2018-01-06 18:38, nospam wrote:

> an autonomous vehicle only has to be a better driver than a human,
> which is sadly, not that difficult. 


Through busy small town main street. car ahead of you is slow. Human
sees the driver ahead using his phone for mapping instructions, knows
that the driver ahead is potentially erratic, might make abrupt moves so
you keep bigger distance behind him than you would normally at that
speed.  It will be a while before car driving software has that.

Similarly, you drive thyrough same town, and know that most of the
traffic is slow because they are looking for parking and may make abrupt
stop if they find free parking, or turn righT/left to get into a
designated parking lot off the main street.

A human who travels this often knows it. A computer wouldn't.

Same with construction. When an interchange is reconfigured, with exits
closed, moved to other side of road, or in some cases, moved up by as
much as 1km, the maps may not reflect this. And if the computer in car
misses the exiut, by the time it rings the alarm and asks the passenger
to drive, it will be too late for passenger to 1- get situational
awareness 2- find out what the best way is (since by then they have
passed and did not noticed the sidns for the new exits).

I suspect governments may have to devide standards for beacons places a
few km ahead of construction to get the passenger to take over from
computer. Or provide standard wireless data at the construction sites
top advise car's computer how to navigate the road through construction.


Until Google et all have research tests here on our roads, I would not
trust these cars.  And one reason they do not do tests here is that they
know their software isn't ready.