Sometimes I feel like I’m going on and on about the inherent issues associated with semi-automated driver assist systems, and how the capabilities of such systems get overestimated to be actually self-driving. I’ll think about writing about it, but catch myself, because surely everyone gets it, and I don’t want to become some sort of crazed, one-note hack, banging the same drum until the skin rends, comically trapping my hand and drumstick inside. But then I’ll see an article like this CNBC one that boldly claims “self-driving cars are here” when, really, they’re not.
This article is maddening because right in the first paragraph the author, Kelly Evans, mentions another article she wrote back in 2015 that very accurately described one of the big problems of Level 2 semi-autonomy, the conundrum of maintaining driver attention in systems that do much of the driving work, yet still require immediate handoff with minimal to no warnings.
That article was absolutely right in 2015, and is still valid today, as the fundamental “vigilance problem” has not been solved, and if anything, is worse now that driver assist systems like Tesla’s Full Self-Driving (FSD) beta are performing more and more of the driving task. This is exactly the sort of situation that promotes driver inattention, while still keeping the requirements for immediate driver takeovers.
Incredibly, this article doesn’t seem to realize that, and instead focuses on the issue that not enough media attention is being paid to “major, transformative developments in automation”:
Perhaps most telling is actually how little national attention is being paid to the major, transformative developments in automation lately. Again, if you told me Tesla was just going to start rolling out driverless technology nationwide, I’d have bet it would be the single biggest news story for weeks at a time. Instead, it’s barely made a ripple in the news cycle. The biggest Tesla story right now is Elon Musk’s Twitter poll about selling some of his shares.
Look, the reason this isn’t massive news everywhere is because of the little detail that states that for something to be considered Big News, that something has to have, you know, happened. It hasn’t happened. Tesla has very much not started “rolling out driverless technology nationwide” because they do not offer anything that can remotely be considered “driverless.”
It’s not just me being some rock-banging luddite saying this, it’s fundamentally how the system is designed to work. It requires a driver to be in the seat and paying attention, because the system could need to be taken over at any moment.
Hell, Elon himself tweeted about this in far more dire terms earlier this year with this tweet:
He’s talking to Tesla drivers when he suggests paranoia there, because there has to be a driver in place, because, again, the system is not self-driving.
Evans somehow seems to think the name could be an issue here:
I think part of that is because Tesla and its aficionados use the rather wonky-sounding term “FSD” (Full Self-Driving) to refer to the technology. Then there are the software releases, or betas, that are continually being updated. So when you hear “Beta 10.4″ or “Beta 10.3.1,” it’s just referring to which version of self-driving software is being used. And all of this is software that’s being dropped to existing Tesla cars that bought the pricey hardware package to support it. So it’s not like we have a raft of obvious new “self-driving cars” on the roadways; the cars already existed, and now just have new self-driving technology, which few other drivers may even realize.
If anything, saying a car has a system called “Full Self-Driving” sure as hell makes it sound like that system, drives that car, fully and by itself. The name is a problem, but for the exact opposite reason stated in Evan’s article: if anything, it vastly overstates the ability of the software.
Now, there are some actually full self-driving cars out there, being tested and used in restricted environments and being carefully monitored, from companies like Cruise, Waymo, and Argo. They’ve been doing this for a while, and it’s very much still in the experimental and development stages.
But this article focuses mostly on Tesla, and, even worse, says shit like this:
So the moment has arrived. Self-driving cars are here. And I don’t think there’s going to be any stopping them now.
Again, no. The moment hasn’t arrived, and self-driving cars are not yet here, at least not in any real quantity. This is not remotely a solved problem, and suggesting that it is, especially when primarily referencing a system that is at best a Level 2 driver assist system is actively dangerous, as it promotes more confusion and leads to more idiotic behavior by people who vastly overestimate what systems like FSD can do.
If you genuinely believe that there will be safety benefits to automated vehicles, then the last thing you’d want are articles like this proclaiming the “Age of Autonomy” is here when reality shows it’s still a good ways away. Jumping the gun will only lead to more trouble, confusion, and delays.
It’s also frustrating to see this because I know CNBC has other reporters who have been very realistic and careful about how this technology is portrayed.
Don’t believe everything you read online, I guess. Unless I’m telling you something about taillights.