The beginning of the calendar year means that the California DMV releases each company’s annual report on disengagements in its autonomous vehicle program. A “disengagement” is an instance in which a human safety driver takes control of the vehicle because the autonomous systems fails.

The disengagement reports are difficult to compare across companies, because different companies have different standards, such as under what circumstances a safety driver should intervene. That doesn’t stop the comparisons, though.

Waymo has the lowest disengagement rate: “1 disengage per 11,017 miles self-driven.” Cruise follows with a rate of, “once every 5,205 miles.” On top of that, Cruise likes to emphasize that it drives in San Francisco, a more challenging environment than most of Waymo’s suburban miles.

Apple and Tesla came in for shaming, although for different reasons. Apple published the worst disengagement rate of any company: once per mile. Tesla reports no self-driving miles, at all, which raises questions about the validity of its “full self-driving” feature set.

All of the disengagement data, though, elides an important fact, which is that disengagements require a human behind the wheel. That means that, almost a year after Waymo tried out fully driverless vehicles in Arizona, and several months after winning approval to go driverless in California, it’s not yet past the disengagement stage.

This is was widely reported a few months ago, so it’s not news. But it’s a reminder that there’s still a ways to go yet.

A whole bunch of autonomous vehicle companies have raised huge amounts of money in the last few days.

Aurora raised $530 million in a Series B (!!) round.

Nuro raised a $92 million Series A round.

I used to think self-driving trucks were an under-explored space, and maybe some other people thought that, too, and decided to start self-driving truck companies.

Ike was so far under the radar that it’s website doesn’t even come up when I search “Ike self-driving trucks”. They raised $52 million last week.

TuSimple, which has been around for a bit, just raised $92 million for its self-driving trucks.

Autonomous shuttle company May Mobility raised $22 million, a sum that seems relatively calm by comparison.

One thing I have not seen a good take on is whether this money is being raised to fund capital or labor. Basically, is all of this cash going to super highly-compensated robotics engineers, or is it going to purchase massive fleets of vehicles. I assume the former, because the latter might be better funded with collateralized debt. But it would be interesting to get a read on the business plans.

On Thursday, February 28th, I’ll be talking about self-driving cars with BMW at the BMW Research and Innovation Center in Munich! 🏎

Sign up here: https://www.eventbrite.de/e/connected-autonomous-driving-with-bmw-and-udacity-tickets-56111365605

There are 300 spots and we want to fill every seat!

Here’s the deal: for every 50 people who attend, I’m going to give away one piece of Udacity swag. You might win a mug! A Thermos! You won’t know unless you’re there 😉

See you in 🇩🇪

My hometown of Reston, Virginia, is getting self-driving cars!

Optimus Ride has struck a deal with a giant property development firm to offer self-driving shuttles at a new mixed-use development down the road from my elementary school.

The proposal seems similar to projects managed by May Mobility, in that the autonomous vehicles will run fixed shuttle routes to and from parking garages. The service will also apparently be limited to tenants of the development. So it won’t be a service open to the general public.

But starting small, self-driving cars are heading everywhere 🙂

Autonocast is a terrific podcast hosted by three of the most prominent personalities in the self-driving car ecosystem: Kirsten Korosec (TechCrunch), Alex Roy (Geotegic Consulting), and Ed Niedermeyer (The Drive).

Last month, they interviewed my former boss, Oliver Cameron, who is now the founder and CEO of Voyage. For an hour, Oliver is remarkably forthcoming about the Voyage story and where they are going. It’s awesome.

https://player.fm/series/autonocast/ep-121-oliver-cameron-of-voyage

My latest post on Forbes.com puts some context around the recent news of Arizona residents attacking Waymo self-driving cars.

“On the one hand, this kind of intimidation is bananas. On the other hand, it raises the question: how often does this type of thing happen to regular citizens?”

The type of behavior that the The Arizona Republic and The New York Times would be totally unacceptable if they were directed against human drivers. But they’re no better when the targets are self-driving cars.

Road rage is just bad.

And besides, there are human safety operators in these vehicles!

“Reading the stories of Arizona residents using weapons to assault Waymo self-driving vans, it doesn’t seem that far-fetched that a person could get hurt or seriously killed.”

I hope we can all take a deep breath and remember that violence is not the answer.

At the beginning of the year I make predictions about how self-driving cars will progress over the coming 12 months. At the end of the year, I score those predictions.

In 2018 that went surprisingly well.

Here is what I think 2019 has in store.

100% Certain

No Level 5 self-driving cars will be deployed anywhere in the world.

90% Certain

Level 4 autonomous vehicles will be on the road, at least in test mode, somewhere in the US.
Deep learning will remain the dominant tool for image classification.
Human drivers will be permitted on all public roads in the US.
No car for sale anywhere in the world will include vehicle-to-traffic-light communication.
C++ will be the dominant programming language for autonomous vehicles.
Autonomous drone delivery will be available commercially somewhere in the world.

80% Certain

Level 4 self-driving cars will be available to the general public (with or without a safety operator) somewhere in the US.
Waymo will have recorded more autonomously-driven miles (all-time) than any other company.
Level 4 vehicles will operate, at least in test mode, without a safety operator, somewhere in the US.
No vehicle available for sale to the general public will come with OEM-installed lidar.
No dominant technique will emerge for urban motion planning.

70% Certain

Level 4 vehicles will be available to the general public somewhere in Europe.
Level 4 vehicles will be available to the general public somewhere in China.
An autonomous shuttle running on public roads will be open to the general public somewhere in the world.
A company will be acquired primarily for its autonomous vehicle capabilities with a valuation above $100M USD.
Grocery delivery via autonomous vehicles, with no safety operator, will be available somewhere in the world.

60% Certain

No Level 4 self-driving cars will be available to the general public, without a safety operator, anywhere in the US.
Tesla will offer the best-performing Advanced Driver Assistance System available to the public.
All publicly available Level 4 vehicles will use lidar.
A member of the public will die in a collision involving a Level 4 autonomous vehicle (including if the autonomous vehicle is not at-fault).
Self-driving cars will be available to the general public somewhere in India.

50% Certain

A Level 3 vehicle will be for sale to the general public somewhere in the world.
Tesla’s full self-driving hardware will include a custom-designed computer.
Amazon will make routine (e.g. non-demonstration) autonomous deliveries using autonomous vehicles.
A company will be acquired primarily for its autonomous vehicle capabilities with a valuation above $1B USD.
Two of the US Big Three and German Big Three (i.e. two of six) will merge.

One year ago, I laid out a number of predictions for the self-driving car world in 2018, along with associated percentages (I lifted the percentages idea from Scott Alexander).

Here are my predictions from one year ago, scored.

100% Certain

✓ No Level 5 self-driving cars will be deployed anywhere in the world.
✓ No GPS or DGPS system will reliably exceed 10cm localization accuracy on all public roads in the US.

90% Certain

✓ Level 4 self-driving cars will be available to the general public, on public roads, somewhere in the world. [Explainer: I count the ongoing Lyft-Aptiv trials in Las Vegas as fulfilling this prediction, even though they include a safety operator. Also, supposedly Drive.AI is operating in Texas, although that’s been surprisingly quiet.]
✓ Deep learning will remain the dominant tool for image classification.
✓ No US road will have a speed limit for autonomous vehicles that is faster than the speed limit for human-driven vehicles.

80% Certain

✓ All Level 4 vehicles available to the general public will use lidar.
✓ Somebody will die in a crash due to a failure of Tesla Autopilot. [Explainer: Mountain View, California, in March.]
✓ Waymo will still have driven more autonomous miles than any other company.
✓ Level 4 self-driving cars will be available to the general public somewhere other than Pittsburgh.
✗ A company will be acquired primarily for its autonomous vehicle capabilities with a valuation above $100M USD. [Explainer: Maybe I’m missing something, but no obvious acquisition jumps out at me. Either the Ford/Autonomic deal or the Hexagon/AutonomouStuff deal might qualify, but neither reported a price. Cruise took some pretty significant minority investment, but was not acquired.]
✓ No dominant technique will emerge for urban motion planning.

70% Certain

✗ Level 4 self-driving cars will be available to the general public in Pittsburgh.
✓ Level 4 self-driving cars will be available to the general public somewhere in China. [Explainer: Pony.AI, although the lack of corroboration makes me suspicious.]
✓ Tesla will sell the most advanced self-driving system available to the general public. [Explainer: IIHS report.]
✓ Deep learning will not be the dominant tool for object classification from point clouds. [Explainer: I don’t have a smoking gun, but my observations indicate that deep learning is still making headway for point clouds, but has not yet taken over.]

60% Certain

✓ 2,000 students will have graduated the Udacity Self-Driving Car Engineer Nanodegree Program.
✗ Level 4 self-driving cars will be available to the general public somewhere in Europe.
✓ Waymo will have exceeded 10 million miles driven. [Explainer: October, 2018.]
✓ Tesla will produce 5,000 Model 3 vehicles in a single calendar week. [Explainer: July, 2018.]
✓ No member of the general public will die in a Level 4 autonomous vehicle. [Explainer: I should have worded this differently, as it fails to capture the Elaine Herzberg fatality, which was probably the most important self-driving car incident of the year. But, as written, this prediction was correct.]

50% Certain

✗ Cruise Automation will open its Level 4 fleet to the general public.
✗ Level 3 self-driving cars will be available for purchase by the general public.
✗ A company will be acquired primarily for its autonomous vehicle capabilities with a valuation above $1B USD.
1,000 Udacity students will have jobs in the autonomous vehicle industry. [Explainer: I’m not going to count this one either way. One thing Udacity has discovered over the last year is how hard it is for us to track students in jobs.]
✓ Self-driving cars will be legal for public use somewhere in India. [Explainer: Startups in India.]

Conclusion

100% of my 100% predictions were correct.

100% of my 90% predictions were correct.

83% of my 80% predictions were correct.

75% of my 70% predictions were correct.

80% of my 60% predictions were correct.

25% of my 50% predictions were correct.

It seems like I did pretty well this year. That I can see I was neither systematically over-confident nor under-confident.

For 2019, I’d like to do a better job identifying issues that are 50/50 uncertainties.

Congratulations on putting 2018 in the books. Here’s looking forward to 2019!