Elon Musk's Dumb Lie About Smart Cars

I have too much respect for Elon Musk to think that he really believes that, because there’s not the slightest sign that we are anywhere close to that level of automated driving. Therefore I must very reluctantly conclude that he’s intentionally misleading the public.

I’ll speculate on his reasons for surfacing this whopper at the end of this column. Meanwhile, Musk isn’t the only one who’s overplaying his hand on full automation. It’s endemic to the automobile industry, most of which has adopted this “sliding scale” model of how automation will come about:

0. No automation. Driver does everything.

1. Driver assistance. Standard cruise control.

2. Partial automation. Cruise control with lane changing, ability to parallel park and other easily defined, predictable driving behaviors.

3. Conditional automation. Self-driving; system hands control to human when needed.

4. High automation. Self-driving; system hands control to human when needed but overrides dumb human decisions.

5. Full automation. You can bunk out in the back seat when you leave, and wake up when you arrive.

Described in this way, full automation seems like merely an extension of technologies that are already working. We’re now at stage 2 and moving to stage 3, it follows that eventually we’ll make the human driver redundant.

However, when you get outside the bubble of AI hucksters and talk to experts in automation and transportation, a different story emerges. A recent ThinkProgress article provides some excellent examples:

  • “Taking me from Cambridge to Logan Airport with no driver in any Boston weather or traffic condition — that might not be in my lifetime.” — John Leonard, VP for automated driving research at the Toyota Research Institute.

  • “Recent Uber and Tesla autonomous vehicle deaths show general use of real self-driving is a decade away. The tech still needs orders of magnitude improvement.” — Michael Liebreich, the former chair of Bloomberg New Energy Finance (BNEF)

And most damning of all:

  • “Despite several decades of automation in aviation, airliners will have human pilots for the foreseeable future. Streets and highways are much more variable and unpredictable than airways, and predictions that the streets will be filled with large numbers of autonomous vehicles within a few years are ignoring not only the lessons of automation history, but also the numerous additional challenges that will be faced on the ground.” — Christopher Hart, former chair of the National Transportation Safety Board (NTSB)

That last quote is the big buzz-kill because proponents of self-driving cars frequently and loudly cite the example of auto-piloted airplanes as evidence that driverless cars are  practical.

In fact, as Hart points out, the current state of avionics automation argues the opposite–that a human pilot is still necessary even when traveling mostly involves traversing a vast empty space.

Rather than a sliding scale of incremental improvement, car automation is better represented by a chasm that’s yet to bridged, like so:

0. No automation. Driver does everything.

1. Driver assistance. Standard cruise control.

2. Partial automation. Cruise control with lane changing, ability to parallel park and other easily defined, predictable driving behaviors.

3. Automation. Self-driving; system hands control to human when needed.

4. High automation. Self-driving; system hands control to human when needed but overrides a dumb human decision.

5. Full automation. You can bunk out in the back seat when you leave and wake up when you arrive.

The myth that driverless cars are just around the corner would just be annoying hype were it not for the fact that the hype is influencing public policy and infrastructure investment. 

Similarly, truck drivers and teamsters are now being told that they’ll soon be replaced by driverless vehicles. Believing this, they’re likely to focus on simply keeping their own jobs rather than working to change the real danger, which is the erosion and elimination of compensation and benefits by an unregulated, non-union gig economy.

Which lead me to why I think Elon Musk is making a prediction that he probably knows to be untrue. Simply put, he’s trying to convince people to buy more Teslas so that they can be on the cutting edge of a driving revolution that will never take place.

And that’s dumb because eventually people will notice that driverless cars aren’t happening. Even worse, this driverless car nonsense is distracting consumers from the real reason to buy an electronic car, which is that the internal combustion engine is helping make the planet uninhabitable.

As I see it, Musk’s vision for the future has a lot going for it. A car that runs on stored solar power would be a huge boon to humankind. Musk doesn’t need to spout fictions about full automation that’s just not going to happen.

Paul Rosen