CS373

I just started CS373: Artifical Intelligence for Robotics, which is Sebastian Thrun‘s robot car course on Udacity.

Thrun is an Elon Musk-type, who has been wildly successful in a number of disparate domains — Stanford professor, father of the self-driving car, Udacity CEO. There’s a lot to say about Thrun on another occasion, but here I’ll focus on the Udacity robotics course.

This is the first course I have taken on the Udacity platform, and I am really impressed by what they have put together. The format is a big advance over the lectures I listened to in college.

For one, Thurn has moved way beyond putting his PowerPoint slides into a YouTube video and doing a voiceover. Instead, Thrun is basically doing a very polished whiteboard presentation, specially crafted for the Udacity format. Which means we’re not looking at Thurn standing at a whiteboard, but rather we’re looking at his hand (or that of a hand model), drawing out well-contained lessons.

But the big step forward is the constant quiz and feedback mode. Every 1–2 minutes, Thrun will ask a quiz questions, to verify we’re still following along. Sometimes it’s a multiple-choice question; often it’s a toy programming problem which requires we write 2–5 lines of Python in the context of a larger program that he gives us.

Thrun is very enthusiastic, constantly telling us how amazing and remarkable we are as students, to have so quickly programmed up a toy version of the Google self-driving car localization algorithm.

In reality, I think it is Thrun who has built something quite remarkable.

Originally published at www.davidincalifornia.com on October 5, 2015.

Car Crash Inequality

The Washington Post reports today on the growing inequality between auto fatality rates for the highly-educated and less-educated in America.

The article is itself reporting on an academic paper (gated) that finds:

Adjusted death rates were 15.3 per 100,000 population (95% confidence interval (CI): 10.7, 19.9) higher at the bottom of the education distribution than at the top of the education distribution in 1995, increasing to 17.9 per 100,000 population (95% CI: 14.8, 21.0) by 2010. In relative terms, adjusted death rates were 2.4 (95% CI: 1.7, 3.0) times higher at the bottom of the education distribution than at the top in 1995, increasing to 4.3 times higher (95% CI: 3.4, 5.3) by 2010. Inequality increases were larger in terms of vehicle-miles traveled. Although overall MVA death rates declined during this period, socioeconomic differences in MVA mortality have persisted or worsened over time.

First things first, death rates declined overall, which is great news.

The disparity across educational classes is troubling, but there doesn’t seem to be a solid explanation. Seat belt usage, automobile model year and safety features, drinking, and other behavioral issues are among the possible culprits.

The Post points out that self-driving vehicles will make the disparity even greater in the near term (assuming self-driving cars are safer than human-driven cars). They do not highlight that this, too, is a good thing. Fewer deaths are better, even if the reduction comes at the higher end of the educational distribution.

My hope, though, is that self-driving cars become so ubiquitous so quickly that the disparity goes to zero, sooner rather than later.

H/T Tyler Cowen

Originally published at www.davidincalifornia.com on October 2, 2015.

GM Super-Cruise

S auto sales are booming, which provides money for R&D. Along those lines, GM has just announced that Super-Cruise hands-free driving will appear in Cadillacs next year.

This is a great step forward technologically, although it’s unclear how important this highway-only, handsfree mode will be to consumers.

“It’s going to be a creep, it’s not going to be a mind-bending thing,” said GM’s product development chief Mark Reuss earlier this year. “I don’t think you’re going to see an autonomous vehicle take over the city anytime soon.”

I’m reminded a little bit of the first touch-screen phones. When I worked at mSpot, I managed a few of our products that ported to the Samsung Instinct, which was pretty buggy and not so functional.

Anyone judging the future of smartphones by using the Instinct could have been forgiven for doubting the whole endeavor.

But the iPhone the phones improved rapidly, due to competition and consumer demand, and by 2010 nobody doubted the importance of smartphones.

I wouldn’t be surprised to see a similar story play out with the first autonomous driving systems.

Originally published at www.davidincalifornia.com on October 1, 2015.

Sergey Brin

Yesterday, Sergey Brin made a surprise appearance at a press event for Google’s self-driving car.

USA Today doesn’t report any notable quotes or announcements, but this is still a great sign for self-driving cars.

Brin is famously press-averse, so to see him lavish this much attention both on the automotive group and furthermore on a press event, can only help focus resources on this area.

Originally published at www.davidincalifornia.com on September 30, 2015.

The Vehicular Turing Test

CNET has a mostly speculative article reporting that in certain cases, Google is training it’s self-driving software to behave more like human drivers.

But which rules of the road is Google prepared to break and which ones will be all too much for its righteous soul? It will now cross double-yellow lines to avoid a car that’s, say, double-parked and blocking its path.

This is an interesting variant on the Turing Test. And do we even want self-driving cars to pass?

After all, to drive in a manner indistinguishable from a human probably means allowing for some unnecessary probability of fatal accident.

There is a wide range of driving ability among humans. If self-driving cars are indistinguishable from the best human drivers, how much less safe are they than if they are programmed to drive perfectly?

Originally published at www.davidincalifornia.com on September 29, 2015.

When Will We Have Self-Driving Cars?

In the past few days I’ve seen predictions that we will have self-driving cars by October, 3 years, 5 years, and probably a few more dates.

The question depends on what we mean by “self-driving”.

Elon Musk was just reported to have predicted fully-autonomous cars in 3 years, although in the quote he hedges a little bit, “My guess for when we’ll have full autonomy is about three years, approximately three years.”

Another interesting question is how widespread driving autonomy will be, when it launches.

Both the iPhone and the fully electric car, for example, were luxury goods when they launched, but decidedly mass-market luxury goods. Not $100,000 toys that only the super-rich could afford.

Tesla, however, has yet produce to a mass-market luxury vehicle (let’s define that as a car in the $40,000 range). So far, Tesla’s strategy has worked, and it has moved steadily down-market from the $100,000+ Roadster. But it will be interesting to see where driving autonomy hits on the price scale.

Originally published at www.davidincalifornia.com on September 28, 2015.

Silicon Valley vs. Detroit

USA Today ran an article entitled, “Autonomous car future will demand tech company and automaker collaboration”.

Amusingly, the text of the article does not mention any such collaboration, or why the future will demand it.

Still, it’s worth thinking about the friend and foe relationships.

A helpful analogy is to the mobile phone industry.

On the Android side, the software itself is owned by Google, but the phone manufacturers are separate entities, who try to add value through a combination of superior hardware, distribution, pricing, and some proprietary software.

On the iPhone side, almost everything is branded and owned by Apple, except for the manufacturing, which is done by a variety of largely-anonymous Asian firms. Perhaps the most famous of these is Foxconn.

Dieter Zetsche, the CEO of Daimler, recently said, “We do not plan to become the Foxconn of Apple.”

How do they feel about becoming the Samsung of Google?

Originally published at www.davidincalifornia.com on September 24, 2015.

Self-Driving [Fill in the Blank]

Clearpath Robotics just announced the name of an autonomous vehicle whose sole purpose is to move inventory around warehouses.

I am curious whether we will see a lot more of this, and whether these more focused efforts will launch faster than more generalized autonomous vehicles.

Lots of industries have specific purposes for autonomous vehicles — shipping, manufacturing, agriculture, defense, aviation. And presumably it is easier to develop automation within the more narrow confines of a niche industry.

Against that, there is the weight of the auto industry — both incumbents and new entrants — who see an enormous market for self-driving cars.

My guess is that we will see the industries develop in parallel, with successful innovations from one field being cross-purposed in other fields. That warms my heart and makes me thing that self-driving cars are coming even faster.

Originally published at www.davidincalifornia.com on September 23, 2015.

Electric Cars vs. Self-Driving Cars

The New York Times reports that Apple is still deciding whether to develop an electric car, a self-driving car, or both.

It is not unusual for Apple to work on several prototypes of a product at the same time, as it did with the iPhone and the iPad.

I am decidedly in the self-driving car camp.

It’s the vitamin vs. aspirin dichotomy I first heard Mark Leslie explain.

Electric cars are vitamins. They’re probably good for us and we should use them.

Self-driving cars are aspirin. We are in pain and we need them now. Driving is (often) excruciatingly painful. Self-driving cars are the solution.

Originally published at www.davidincalifornia.com on September 22, 2015.

Cruise

The self-driving car revolution is mostly being led by big companies — Google, Uber, Tesla. Outside of the tech companies, there are traditional auto manufacturers like Audi and BMW, as well as a host of suppliers ranging from nVidia to Mobileye to Continental Automotive.

One of the few small startups that is emerging in the driverless car space is Cruise. They just raised another $12MM+ to fund hardware that turns a normal car into an autonomous vehicle.

That’s pretty awesome, and I wish them luck.

Originally published at www.davidincalifornia.com on September 21, 2015.