Why am I so easily annoyed by articles in the MIT Tech Review about autonomous vehicles? Example. Example.

Here is another article that I’m finding exasperating. This one is about my favorite self-driving vehicle company, Otto. It is my favorite because they have clearly realized the same thing I have realized and that is that the best challenge to undertake in the field of self driving vehicles involves long haul freeway driving, ideal for trucks. Here are some reasons I believe the Otto business model of focusing on long haul trucking is the best.

  • Professionalized setting.

  • Managed fleets.

  • Driver costs are non-trivial.

  • Predictable and repetitive routes.

  • Huge amounts of long simple highway travel.

  • Fuel economy and management are critical.

  • One of the biggest safety problems involves keeping drivers awake.

The article adds to this list by informing us that according to the American Trucking Associations, the U.S. currently has a shortage of about 50,000 drivers and predicts that a total of nearly 900,000 new drivers will be needed over the next eight years. "We have customers calling us up saying they’ll buy 10 new trucks from us if we can provide the drivers, too," says Carl Johan Almqvist, who heads product safety at Volvo Trucks. (Of course if you actually check these facts at the BLS you’ll find that truck driver is not projected to gain or lose with respect to other employment options and the expected 5% growth will require ninety thousand new drivers, not nine hundred thousand. Order of magnitude derp? Whatever.)

Great! Long haul trucking is a perfect problem to solve. Let’s get to work!

But then comes the buzzkill. Although the article says stuff like "the relentless pace of automation" and "we’re not waiting", they quote the Otto product manager stating, "We’re at least a decade away from having trucks with no driver in it."

Seriously? Ten years for this too? I was envisioning this as a relatively easy problem (i.e. way mo easy than a Waymo car defeating the adversarial animatronic human-shaped "yard sale" signs I erect to thwart it, which humans would easily know to ignore). I know I will not live long enough to see Google’s vision for autonomous vehicles. But the simplest unmanned trucking? The actual (absurd) pretext for the original DARPA challenge? I am disappointed.

Although they seem squeamish about an unmanned cab, they seem strangely ok with the driver being present but not driving, e.g. napping. This strikes me as weird since there’s little a human can do; in the event of an accident you’ll just lose a truck driver in addition to whatever other mayhem results. I thought it was a little cringeworthy when the Otto product manager says, "drivers could use the time away from the wheel to…[do lots of worthwhile stuff, including]… learn a second trade…" Don’t worry guys, you’ve apparently got 10 years. If autonomous vehicles don’t provide a real ROI soon, there will be some AI researchers facing that same problem.

Here’s my real pet peeve about Otto. The article mentions a stretch of highway near Columbus, Ohio planned for testing autonomous trucking which makes me really wonder why the heck is Otto’s headquarters in San Francisco? Oh, it’s in the "once-seedy South of Market section"; ah, I see, so it’s really down to earth and sensible for a company focusing on long haul trucking. They explicitly say, "No fancy, shiny offices here." In San Francisco. (For reference, I walked farther from my parking spot to my office this morning than I would from Otto HQ to Fisherman’s Wharf cutting right through downtown SF’s financial district.) I’m sure they know some things about bay area real estate that I don’t. If it wasn’t for their extremely harsh location, I’d love to work for them.

I feel like this whole industry knows something huge is going to happen but they have no idea what or when. The infrastructure, political dynamics, capital allocation, liability, technological equivocation, talent pool, culture, public sentiment, and even benefits, all are huge unknowns. For example the article says, "Volvo is still unsure about social acceptance of the technology. The company sometimes identifies the license plates of passing cars when testing its autonomous trucks, and then tracks the car owners down and surveys them about their perceptions." No, that’s not creepy at all.

Whatever happens with Otto, it’s going to be a long road.