Comment by mrandish

Like red light cameras, this is one of those genuinely good ideas that is likely to end up being much less effective than it could have been for two reasons:

1. The regulators who implement it will set the detection times and thresholds shorter and tighter than what would have worked best. Why? Because Pareto says the worst ~20% of serial offenders are causing >80% of the serious accidents. So, the optimal settings would not trigger for brief detections or even for occasional longer detections. They would instead minimize interruptions, inconvenience and false positives for the majority of drivers and only trigger when long lapses start occurring frequently. Just targeting only the worst ~20% would save countless lives with so little disruption, it would be widely lauded.

Unfortunately, the well-intentioned bureaucrats won't be able to reason past the implied moral hazard (and potential political blowback) of being responsible for permitting any lapse which might result in an accident. But any near-zero tolerance threshold forces a useful but inherently imperfect technology into failure modes that will cause resentment, resistance and demand for workarounds.

2. Because the inevitable deterrent fines will be used to pay for enforcement, it will become its own sub-bureaucracy inside the system with its own staff, budgets and performance metrics, all with a vested interest in 'saving more lives' by increasing staff which can be made 'revenue neutral' by increasing fines. This is what happened with red light cameras in many municipalities. The 'free money' from mailed out 'photo tickets' was so good they eliminated reasonable grace periods and even shortened the yellow light time, so irate citizens got them banned in many cities.

Personally, I'm not worried about someone who slides through an intersection a half second after the light turns red. It's vanishingly unlikely that person is going to cause an accident. Where red light cameras could (and should) be saving lives is >5 seconds after the opposite light has turned green and cars are in the intersection. I barely avoided a lady last year coming straight toward me at ~50 mph more 10 seconds after her light had turned red and multiple cars were stopped in all three other lanes next to her. If a car had been stopped at the light in her lane, she'd have hit and killed them. Despite every car at the intersection leaning on their horns she never slowed and never swerved. There's no sign she ever saw the light or the intersection at all. But someone had to fuck with the yellow times to make money.

And I'm hardly 'soft' on distracted driving. My wife and daughter were hit by a distracted driver 9 weeks ago and both received serious concussions. Fortunately, they'll both fully recover but my wife is still in physical therapy and the broken rear axle on her car still isn't done being repaired. They were T-boned at a perfect 90 degrees by a 16-year-old boy going the opposite direction who was turning left across their side of the road into a side street. Here's the thing: it's a straight, extra wide, well-marked, divided residential boulevard with perfect visibility, zero obstructions and bright streetlights. It was 10p and no other cars were anywhere in sight. He wasn't intoxicated and the police just rolled their eyes when he repeatedly denied being on his phone. He was in a dedicated, uncontrolled turn lane, where you stop, check there's no oncoming traffic and then make your left turn. And he drove straight into literally the only other car on the road. We later timed it and he had to be looking away for more than THREE seconds to not see the big, bright, well-lit car approaching under the working streetlights. If my wife hadn't watched him the whole way and swerved when he didn't stop, it could easily have been a potentially deadly near-head-on collision. Until he grows up a bit, he's the ~20% worst-case, repeat offender that this tech could and should stop. Not someone who glances down at their phone while stopped at a light or looks away for a split second to turn down the volume on their radio.