Uncommon Descent Serving The Intelligent Design Community

AI: A rational look at self-driving vehicles, and a cautionary marketing tale as well


Further to them being oversold, from researcher Filip Piekniewski at his blog:

When the software fails and e.g. the control system of the vehicle hangs, it is more than likely that the end result of such situation would not be good (anyone working with robots knows how rapidly things escalate when something goes wrong – robots don’t have the natural ability to recover from a deteriorating situation). If that happened on a freeway at high speed, it would easily have lead to a serious crash with either another car or a barrier. If it happened in a dense urban area at small speed it could lead to injuring pedestrians. Either way, note that Waymo only reports the events that fulfill the California definition, i.e. these are actual failures or events threatening traffic safety as concluded by their extensive simulations of each event.

First the progress appears to be plateauing. It is not a very strong trend based on one or two data points, but the progress Waymo and others have made over the year (perhaps excluding Cruise) is not overwhelming. In my personal opinion, the current level of disengagements roughly reflects the rate of tail events – some more or less unusual conditions that either require some common sense or better anticipation of the behavior of other traffic. Behavior in these conditions can no longer be solved by improved sensors and requires better AI.

Secondly, it is hard to support the claim that AV’s are much safer than humans. … More.

This is well worth knowing though one senses that it doesn’t much matter to enthusiasts.

Experimental physicist Rob Sheldon offers a historical take: The 1960s industry enthusiasm for the home picture phone: your phone as unscripted real-time You TV…

Well, of course, self-driving cars are oversold! For the same reasons that Ma Bell thought everyone would want a picture phone in 1966, but thought cell phones were only a niche market for doctors and lawyers and emergency personnel.

The guys building the technology, like the fellow in Florida who died in a Tesla doing 75mph on Autopilot, are “first adopters”. They really want the future to be now. Most of us just want to get to work without the hassle of fighting the traffic. And honestly, it takes a lot of brainpower to shave 5 minutes off the commute.

Why would I use a safety conscious computer that will never bend the traffic rules? The car is a tool, not an end in itself.

The picture phone never took off because then phone calls had to use visual protocols–combing your hair, putting on a shirt, making the bed–which destroyed the whole convenience of an audio-only communication channel in the first place.

Yes. The inability to look sick while pretending to be sick killed the picture phone for sure. (Deselecting or dimming the vision part wouldn’t work because that would only create suspicion.)

Sheldon doesn’t even get into the question of the informal social power that audio-only creates: The other party must pay attention to what is being said, not to how one (or one’s environment) looks.

The lesson here is that technologies are not adopted because they exist. They are adopted if they serve a perceived need. Even today, visually enabled online meetings tend to be “socially scripted” encounters, not “I answered the phone and the whole company saw me in my underwear shouting at my dumb dog. And some joker in Receiving published it on YouTube … ”

Note: Rob Sheldon is the author of Genesis: The Long Ascent

See also: Artificial intelligence: Self-driving cars are oversold, says researcher. Whether or not these problems get straightened out, some of us wonder whether self-driving cars will really take off. Don’t people drive cars because they want to do it themselves? Otherwise, why not take the bus? Note: Anyone who has dealt with an older senior who must give up driving due to cognitive losses will understand what I mean about driving and a sense of independence.

Nonlin.org, please ignore my comment @3. I realize that I misread what you wrote @2. FourFaces
@2: It’s not rocket science and very useful. Useful yes. But it's turning out to be orders of magnitude harder to figure out than rockets. AI experts found out that it takes a lot of common sense and brain power to drive safely. They are not even close to level 4 autonomy. FourFaces
Not the same. Picture phones failed because the technology wasn't there (very expensive until recently), because voice covers most of the communication needs and only lastly due to privacy issues. We all use cruise control and would gladly use some higher level of self-driving if available. It's not rocket science and very useful. Of course, in the end AI is just a dumb tool like a hammer and will never "evolve" into anything like in the movies: http://nonlin.org/ai/ Nonlin.org
news, picture phones have arrived, at base level as smart phones, then onward to video teleconferencing. Of course, decades later and with very different technologies. KF kairosfocus

Leave a Reply