Self-Driving Trucks are WAY more Exciting (and important) than Self-Driving Cars. Here’s Why.

Note: Republishing this post I shared on the Medium paid subscription a few years ago. This was originally written in October 2015.

Every time I take a road trip, my mind wanders to the highway of the future. What vehicles will be driven? What kind of intelligence will drive them? What will America’s Interstates look like in 20 years?

We’re all excited for self-driving cars (so that we don’t have to look for parking or drive kids to school), but I think we’re missing more enormous applications of self-driving technology...

Autonomous Trucking is going to be insanely cool.

We’re largely blind to the beautiful, sophisticated network of logistics that works every minute of every day, so it’s not the first application that comes to mind — but it might be the most important. Here’s why…

What Self-Driving Trucks might look like

Since we don’t have a driver who needs the space (and comforts) that a human requires, there’s no need for the driver’s cabin... or anything else you see in front of the trailer. The truck of the future could look a lot more like a giant skateboard than anything else.

This is what the platform of a Tesla S looks like without the body:

Tesla Chassis.jpeg

This ‘skateboard’ is the foundation of the car — everything it needs to drive is in this picture. Add headlights and the autonomous rig and you’ve created a fully operational truck bed. Make it big enough to drop a shipping container on and you see the future of trucking.

Logistics Implications (Way Faster)

The first improvement from autonomous trucking is going to be run time. Puny humans need to sleep or they crash. Tight regulations limit how many hours a driver can put in: no more than 60 hours over a 7-day period.

How many hours can software drive? All of them. All 168 hours every week.

That would nearly triple the speed of delivery.

The second-order effect is also significant — trucks will be on the road 3x longer. They’re not in rest stops being used as portable hotel rooms with 100,000 bagels attached to them in the middle of Iowa.

Another implication of software control on trucking is that trucks can drive WAY faster. We can allocate the left lane of enormous long stretches of highway and have trucks all network together, get in a tight formation like a train, and go 100+ mph in a slipstream of aerodynamic efficiency.

Autonomous Trucks might be able to drive 2x the current speed.

The combination of these two effects will have a huge impact on our ability to transport goods quickly. A 24-hour drive from the center of America (even at the speed limit) will get you almost anywhere in the country (Sorry, Seattle).

United States Map.png

Assuming 1.5x the current speed on average (adjusting for traffic, cities, etc.) driverless trucks can get their cargo anywhere in the country in under a day, often significantly less. This improvement will decrease spoilage and decrease inventories. Also, faster deliveries will enable new services that are impossible today.

Efficiency Implications (Way Better)

If these computers turn out to be good at driving (as it looks like they will), the energy efficiency of trucking improves a lot. Currently, the ton-mile efficiency of trucking is trounced by trains — by ~4x. With a good algorithm behind the (now metaphorical) wheel, the efficiency gap will close up a bit.

Networked software can communicate far better than puny humans. They can find out when there’s an accident 10 miles ahead, and prepare. Nothing is a surprise. In the safety of their designated lane, these trucks can get within 6 inches apart from each other and not worry about creating an accident. This synchronized driving will create an aerodynamic efficiency which edges closer to the energy economics of locomotives.

If these trucks end up being electric (and why wouldn’t they be?), there will also be an improvement in energy efficiency. Electric power has roughly double the efficiency of internal combustion engines. Elon Musk talks about that here:

Economic Implications (Way Cheaper)

How much will this matter? Probably a lot. Nearly 70% of all freight transported annually in the US depends on trucks — $671 billion in goods.

The combination of the above factors would (hopefully) result in a massive decrease in the cost of transporting goods, and significantly increase delivery speed, which in turn will reduce costs further.

Massively increased utilization will decrease the total number of required trucks, and that plentiful supply will lead to lower prices. Because shipping is a competitive industry, with low value capture, these gains in cost efficiency will be passed on to the end consumer.

Since almost everything gets shipped at some stage of its supply chain — usually multiple times — this efficiency gain should affect almost all goods to varying degrees. Low-cost retailers like Amazon and Wal-mart will press their advantage and push prices down, forcing others to do the same.

The speed of this new set of technologies will improve the efficiency of warehouses and storerooms, making just-in-time even more ‘just-in-time’ and lowering Inventory Carrying Costs across all industries.

All of these small incremental improvements will add up to significant benefits to society as they overlap and compound. (Doing the math on this level of hypothetical seems dubious, but if you take a stab let me know.)

What About All These Truckers?

Truckers are probably not thrilled about this vision of the future. This is a bigger problem than you might think: there are over 3.5 million truck drivers in America. More than 1% of the population will find themselves suddenly very jobless (Or not so suddenly, if they’re reading this).

This map shows the most common job in each state as of 2014:

This is bad news.

This is bad news.

[As a total aside, there are some cool potential products for about-to-be-replaced truck drivers. They can work on new skills training through audio books and courses while they’re working their last few trucking years. Maybe translations, or sales, or something? An opportunity worth thinking about.]

Another bummer is that the days of CB radios and getting truckers to honk are almost past. On the bright side — a few less puny humans in rest stops.

What Next?

I’m not sure what we should DO about this future possibility. Buy some UPS stock? Get ready to hire some ex-truckers? Try to find, fund, or found the company that is going to build this product and supporting ecosystem?

All I know for sure is I’m not enrolling in trucker school. And that I’m excited for what this might mean for our economy and our lifestyles in the future. In the meantime, I’ll keep dreaming of Optimus Prime trucking along next to me hauling my books and blueberries.

It’s Happening Already

When I talked about this idea and asked friends (thanks Matt and Nauzli) to read this draft, they found examples of this technology starting to roll out.

Walmart is investing heavily in it’s own advanced logistics infrastructure, and its trucks looks super cool.

Stolen from Business Insider. Not sure who they stole it from.

Stolen from Business Insider. Not sure who they stole it from.

But they’ve got nothing on Freightliner, who unveiled its first version of the street-legal autonomous big rig, called The Inspiration Truck. Look at this beast. Maybe Optimus Prime isn’t such a far-off dream at all.

This photo both makes me want one of these trucks, and deeply fear the robot revolution.

This photo both makes me want one of these trucks, and deeply fear the robot revolution.