That was a great article. It explained so many topics so clearly I even enjoyed reading the definitions of those I already knew about like stalling. The author is basically pushing for augmentation over automation. Those in IT or INFOSEC are probably well-aware that augmentation seems to be winning given how often automation fails on corner cases. Over a decade in high-assurance INFOSEC has taught me we probably won’t automated it due to all the contextual stuff and effects of domain knowledge. Instead, we continue to automate away as much tedious or unnecessary stuff as possible while letting humans use their brains on whatever level they choose to look at. Seems to be getting best results.
Extra props for the plane and car examples given I think that will really drive the risk home for laypersons. Especially this: “what chance would you or I have when the computer in our car says, “Automatic mode disengaged” and we look up from our smartphone screen to see a bus careening towards us?” Given time a context switch takes, they’ll have just enough time to realize they’re dead before trying not to be in a fraction of a second remaining. Most will probably get hit or hit someone else.
What I found interesting about this argument is that many pro-automated cars is that there will be less accidents in total. In fact this is an argument for solving the liability problem: fewer accidents means insurance can cover the costs and no-one will be to blame. But it could be that automated cars have fewer accidents but of significantly higher magnitude, which would be bad. I don’t think there is any way to stop this so, it will be interesting to see. Luckily I haven’t been in a car in over a year so hopefully I’ll survive the automated apocalypse :)
“solving the liability problem: fewer accidents means insurance can cover the costs and no-one will be to blame. ”
I like where you’re going with this…
“fewer accidents but of significantly higher magnitude”
…then reality steps in to ruin the plan. It’s like a law of nature. Murphy’s Law of Liability. ;)
“ Luckily I haven’t been in a car in over a year so hopefully I’ll survive the automated apocalypse :)”
Do people drive cars near you? That brings me to the next issue with your nice idea of liability. That is that we never get in a situation where everyone does something. There’s always holdouts or even opponents. So, we’ll have self-driving cars, cars with some driver assistance (eg radar/cameras), safe cars with no assistance, and 18 wheelers with sleepy drivers plowing along at 100+mph since they get paid by the mile. This has already been a significant problem with Google’s cars as human drivers expecting they’ll act as dumb or aggressive have nearly caused accidents many times. Enough automated and non-automated cars on the road will magnify this. I’m sure ice and rain… hydroplaning… will magnify it more. The overall effect might drop down to manageable levels once people get used to being careful around any new car. Could take time, though.