A former pilot on why autonomous vehicles are so risky

I can see it. The other day a self driving Cruise car was extremely aggressive with me, almost feeling aggressive as I tried to cross an intersection on foot. It was stop&#x2F;starting quickly towards me to try and get me to walk faster it feels like.<p>I think that these self driving cars need some kind of feedback system for pedestrians and cyclists. Even an emoji face that shows facial expressions would help. When I see a real driver at an intersection, I feel like we come to some agreement about what is going to happen next based on our facial expressions. Or I can at least guess what may happen and take an action. With the self driving car there is 0 feed back. It&#x27;s just an empty car that may or may not run me over.<p>Additionally, I also almost got run over by a dude head down driving in the Mission the other day, typing on his phone with both hands as his he just felt comfy having his Tesla do all the work.<p>At the moment all of this feels so sketchy. I hate my life being used for beta testing software lol.

This is one of the things that really annoys me. We have <i>so much</i> data and case studies of the effects of automations (of different levels of sophistication) and user interfaces and how those interact with other things including distractions through FAA investigations.<p>To say nothing of other safety critical industries.<p>We know what works ok. We know what’s dangerous. We know what’s lethal. There are lessons to look at.<p>And yet Tesla and a lot of people seem content to re-learn the same lessons through blood over and over again. Like this is 100% incomparable to anything that’s ever happened before.

Sounds like a very good person to have in that position. Being a single pilot in an aircraft, she would not only have years of experience relying on automation, but also risk-reward decision making involving life or death. Self-driving cars are not a &quot;restart the container&quot; situation.

Did any of you read the article? She acknowledged there are some cases where it’s working in well defined environments.<p>But she clearly shows how we are currently in the most dangerous phase “car drives except handing off to human drivers in a crisis”.<p>“ The policy should be that either the computer is driving or you are driving. And by driving I mean steering—people do fine with regular cruise control. The act of keeping your hands on the wheel and guiding the car’s lateral motion is enough to keep your brain engaged. So, no L3 [full self-driving, but the driver must be ready to take the wheel], which is too confusing, and no hands-free L2 [partial self-driving]. I am not against the passing of control per se, but there should just be two modes of operation, with crystal clear feedback about which mode you are in.”<p>FSD has been 5 years away for… 15 years now?

I see a lot of people arguing from their intuition (or experience in other fields) that what Tesla is doing with their driving assistance features is dangerous. There are millions of Teslas on the road, shouldn&#x27;t the first step be to find out check whether they crash more often in practice? If not, then what&#x27;s the big deal?

I&#x27;ve tried &#x27;self driving&#x27; and driving assist. Driving assist is intolerable for me because it can jerk you around even when you have a reason to be close to the line. I don&#x27;t like fighting for control of the car.<p>I am a huge proponent for good self driving but experiencing driving assist and being a passenger in Tesla cars has made me much more cautious of the current capability.

For once, someone with the courage to say « not in my lifetime »

“ Cummings: There is no evidence of mitigation. At NHTSA we couldn’t answer the question that you’re less likely to get in a crash—no data.”<p>Compared to what? I find this very hard to believe. Surely forward collision avoidance systems have stopped a number accidents over the years. To say that there is no evidence for it, no data even, seems to suggest that you might have a data problem. Hard to take them seriously when they don’t have good data, yet maintain their very negative position.

I have a feeling that she has the eyes of the highly trained professional driver (pilot) with good senses and ability to react well.<p>Not something that can be told about the bulk of the drivers on the road.<p>Although I see the point of and agree with the necessity of &#x27;crytal clear&#x27; situation, knowing what exactly can be expected from the assitance, I believe that the average driver should have as much help as possible. The roads are pretty unsafe yet, because of the human drivers.

As with any automation; it’s not the automation that’s the problem, is that there are still some parts with people in it.

‘Former pilot’ seems like a fairly irrelevant part of the resume of a human-autonomous vehicle interaction researcher with a PhD in systems engineering and a series of professorships either side of an NHTSA advisory position.

You know what&#x27;s risky? Letting <i>humans</i> drive. 30,000+ deaths a year in the US alone.

It is difficult to get a man or woman to understand something, when their salary depends on them not understanding it.