Traffic Meditations

Sitting in Brisbane traffic has its upsides, because it allows for uninterrupted brainstorming. I have some insights worth discussing with you, supposing I prompt some interest.

Tiny Courtesies

The first is the way traffic moves. We live in a society where a host of collaborative processes align to allow us to go about our business more or less peacefully. From time to time I reflect on what a magical thing this is — sort of the opposite of the prisoner’s dilemma, where the balance seems to be on getting along rather than getting ahead. This is all the more remarkable on our roads when you think about how deadly these machines are, and how capable of assertive behaviour.

I have driven all over the world, and until recently lived mostly in the US. You can tell a lot about a people by how you fit into traffic. As an AI practitioner for as long as it has existed, I do not expect autonomous cars to fill a role of social cohesion because driving is the one thing that exercises our social compact in detail. Once you know what to look for, you see thousands of tiny courtesies. This is how it works, by graceful collaboration, largely selfless. Possibly the most obvious are letting cars merge on the highway. Merge lanes here are far shorter than in the US as a fact.

The US also doesn’t have what they call ‘traffic circles’. The reason I believe is that there a permissive tacit collaboration required that Americans just can’t support. It seems a small thing to notice while driving, but I do notice it and believe it reflects a quality of Australia that makes it the most cohesive society I know. Driving gives me joy in this regard, as I approach my swearing in as citizen in a few weeks.

Manufacturing Matters

I will guess that there are about 20 million cars on Australian roads, and that conservatively each averaged over $50,000 in price. On a typical day driving the kids, I see a part of that astonishing $1,000,000,000,000 that is rolling around. Folks, that is a trillion dollars. I venture a guess that I’ll see one car a month that was assembled in Australia. That’s a trillion dollars we send out of the country, basically to never return, other than money we get (and only partly keep in country) by selling the dirt that comprises most of those machines.

I happen to know that the automobile manufacturing sector drives the robotics industry. So if we have no automotive industry, we lose more than half of the drivers for robotic innovation.

I also know something about manufacturing costs and processes, and I don’t buy the argument that Australian labour is too high. It is high, but not notably higher than in any European or North American country. As it happens, the ‘touch labour’ in that business is small and getting smaller. And it isn’t because our right-hand drive are unduly peculiar. Who else drives on the left? India, Japan, Indonesia, Pakistan... more than a third of the world.

I’ll make two guesses as to why we are so broken. The first is that the market is divided not just by model type, but also by manufacturer. Probably the most popular specific model I see is Toyota SUVs, but as a percentage that is a vanishingly small number. I doubt there is a country with more competitors than we have. What that means is while our total market is large, our consumption habits push a variety so rich that no single model would get near the critical mass required to build a plant.

The second is that investors here are unsophisticated, and consequently the most risk averse in the world, despite having massive investment dollars in our super funds.

So, are we doomed manufacturingwise? 

A trillion is a big number, and sending that out of country, forever, makes our submarine escapade seem like kid stuff.

Here’s an idea. When at DARPA, we worked with a major car company (the biggest and most dynamic at the time) and a consortium of robotics companies. We ran some numbers and projected some futures. One reality of mass production today is that you have to make cars in one place and ship to many places. You have to guess right, because costs are such that delivery must be a one way trip. You also have to make them and store them, usually for many months, hoping someone walks by with cash. It costs more to ship and store these things, and build the lavish dealerships, and pay high sales commissions than it does to make the buggers.

We now have systems that can economically set up small local robotic centres that can assemble a car on demand. You and your family can drive a demo car for a physical experience. Then you can work in three d headsets to design your own car with many more options and aftermarket improvements than now possible. Then that design goes to the robots and in our simulations in two weeks you get a bespoke car potentially unlike any other on the road.

The savings in carrying cost to the dealer network is close to 20%. Depending on how efficient your just in time supply chain is, assembly cost savings can be closer to 10%. Robots, my fellow Australians. And get this, the group we were working with were company agnostic. Toyotas and Fords in the same regional assembly centres!

The first folks to put this together can franchise everywhere in the world. Within our shores, this plays to our market tendency for variety, which can increase manifold.

There are a couple more meditations on swarming and micromobility. But I’ll save them for the next journey.