Cars could be our next wave of prosperity.

One of the interesting things I learned in China was that their political leadership believes the 1950s in the U.S. with the development of Interstate highway system, big advances in car ownership, and dispersion of population out of crowded cities was key to economic growth and success and the advancement of the U.S. to superpower. So China wants to emulate this. Charlie Wilson once reduced the entire U.S. economy to “what’s good for GM is good for the U.S.”

So I believe a new generation of cars could revitalize the U.S. economy. Now I don’t mean tinkering around the edges and bailing out the sclerotic auto industry and dumping money in Tesla or adding some new styling or a tiny bit of hybrid drive. I want something way bigger than that.

This article could my attention that California is moving toward self-driving cars, which largely don’t really exist. But I think self-driving cars could be just the ticket. I think we’re in about the same place, in the technology curve, with self-driving cars as we were with rocketry just before announcing the moon program in the 60s. We more or less know what to do but our efforts thus far are too puny to really make a difference. But this is the kind of technology where you can light a fuse and it takes off and it pays off.

It doesn’t have to be some top-heavy government program; in fact, I’d expect the private sector, with appropriate regulation, to do most of it. The government’s role is simple – vision. Today in private industry the planning horizons are totally short-term, rarely beyond the current quarter. When I started in high-tech, 2, 3, even 5 years development (not even R programs like Xerox Parc which was decades) were common. Today almost nothing is done longer than a year (except of necessity in fields like pharmaceuticals). I’ve been critical of companies like Facebook that they’re not actually “high tech”, because they just cobble together existing bits to create relatively trivial products. Show me any transistor in anything any VC funded company has done for decades.

So no one does hard and expensive and long-term research. Yet that is where the huge payoffs are, perhaps not to a single enterprise, but to an economy. So it’s simple, the U.S. simply sets a mandate that by 2030 no human-driven car will be allowed on any federally funded road. And it sets a second program of spending to: a) upgrade existing federally funded roads to be compatible with self-driving vehicles, and, b) it funds some additional basic research in a few critical areas (like the mission DARPA had that worked well). And it does NOT do any subsidies. No, set the goal, do a little of the longer-term and riskier research, let enterprise entirely do the D where it must succeed through competition not political favors.

Now I believe this simply translates into a rebirth of some areas of R&D, a pile of patents if we get there first we can charge China for, and a complete turnover of the automobile and trucking fleets (probably want retrofits for trucks) which is bound to create a boatload of jobs and profits for the innovators. Along the line we set drastically higher CAFE standards (which will do more for energy independence that more drilling could ever do, Romney is a lousy accountant if he believes there is 60% of our daily oil requirements just sitting in a few wildlife refuges), we set higher pollution standards, and we set way higher safety standards (the carnage on our highways is unforgivable). All that pushes technology really hard, but I believe within its range to accomplish. Ambitious national goals not only get a job done and make business a lot of money, it restores our national confidence and can-do attitude that can spill into other areas.

And we pay for all this with a steep carbon tax, at least the equivalent of 10$/gallon. Obviously that clobbers working poor and some others, but it all gets rebated back against payroll taxes, even negative tax if needed. But people should pay at the pump even if it costs them nothing over the year because it is that immediate shock that gets attention and gets their interest in alternatives moving. I love my Prius every time gas prices go up, but I think the Prius technology is too complex (I actually think Volt is smarter design, but I didn’t buy Prius until 3rd generation and I wouldn’t buy Volt until then either, hopefully it will exist and be perfected by the time the Prius is worn out). But I think we can do even better than Volt or Prius as the basic platform and a lot better in the control system.

And I think self-driving has to go further than just software and a few lidars. To deal with traffic congestion we also need networked vehicles. And not the top-down command-and-control model, but the nearly flat and distributed model of the Net, each car talking merely to a couple of nearby peers but then able to act in concert with larger-scale conditions. I know that congestion is caused by human being reacting incorrectly to traffic situation. We have a lag, like any control system, and that lag creates various feedback effects. Remove that lag and traffic density, per minute, can rise. I believe it would be possible to have a flow of traffic, with tight spacing, at 120mph, but never with human drivers. It’s like the modern fighter planes, they are dynamically unstable and statically unstable – a human being can not fly a B2 (or the Space Shuttle), BUT, a human being can give high level commands to it.

For the long-haul why do we need any human control at all. Many people now program a route into dashboard GPS and slavishly follow that. My Prius knows its gas mileage at minute level resolution and in conjunction with GPS, plus the networked traffic information, can easily program its own gas stops. If the car can memorize tons of seat positions (as a trivial marketing feature of today’s cars), it can learn my bladder habits and whether I got a cup of coffee at the last gas stop and so schedule a biobreak, miles before getting there, so it can do the needed lane switching even at 120mph and with two car-length spacing. With large-scale knowledge my car can be predicting what it needs to do 30 minutes in advance when it’s time to take the exit, just as airplanes don’t take off today until they’ve been assigned a landing slot.

It’s all doable. Not only does it require an entire new fleet of cars it requires, at least on the high speed routes, a lot of new infrastructure (more jobs and profits). I actually first thought of this idea long before the DARPA challenge of building the first self-driving cars (and way long before Google). There is a stretch of the Indian Turnpike that has a lot of sensors because they keep having deer-vehicle collisions. I actually forget how it works, but the sensors are networked into some model and provide some sort of warnings to drivers. Obviously other cars are not the only thing that has to be included in the driving models. So that’s where some of the research comes in. A self-driving car that can go up and down streets to take pictures is one thing, a car that can handle congestion at high speeds, but still with unexpected hazard avoidance is quite another. And a challenge.

And a challenge is what we need. I believe U.S. manufacturing, esp. the high tech stuff can match low-wage stuff so I believe a vigorous car industry that pays good wages can be globally competitive. And if it’s based on technology where the U.S. owns the intellectual property rights then we have that extra cushion of competitiveness to justify the capital investment. U.S. car companies do get it right some of the time, but they are too prone to short-term quarterly decisions to over-invest in fads, esp. in their comfort zone of gas guzzlers. But it doesn’t have to be that way and in fact we just give them a swift kick in the butt to be more aggressive and if they aren’t make sure entrepreneurial car manufacturers can get started, or even, appropriate partnerships with foreign companies.

Twenty years of aggressively trying to make ambitious goals would jumpstart the economy (esp. on something so big as autos) and it would get our mojo back.

About dmill96

old fat (but now getting trim and fit) guy, who used to create software in Silicon Valley (almost before it was called that), who used to go backpacking and bicycling and cross-country skiing and now geodashes, drives AWD in Wyoming, takes pictures, and writes long blog posts and does xizquvjyk.
This entry was posted in opinion and tagged , , . Bookmark the permalink.

2 Responses to Cars could be our next wave of prosperity.

  1. dmill96 says:

    I just found the PBS/Nova documentary on the 2nd DARPA challenge was available on Netflix and so watched it. I’d known several cars completed the race, but the documentary did a good job of creating some suspense between who among the leaders would win (five cars completed the race). It’s interesting that Stanley (the Stanford vehicle) put most of its development emphasis on software, using almost entirely off-the-shelf hardware whereas both CMU vehicles had a lot more hardware development. This actually reflected the bias of their directors (formerly colleagues). Naturally I vote with the software emphasis because the race course itself needed little but a conventional decent SUV. Getting really solid sensors is also part of it, but the adaptive vision system Stanford used definitely seemed to be smarter whereas CMU was a bit more brute force in its approach.

    I actually remember these same kinds of meta-engineering issues in that first MIT/Caltech electric car race. Caltech took the simpler and clean approach, MIT went all out on sophistication. As a result Caltech had lots more testing of their system and it proved to be more reliable and actually completed the race under its own power (MIT “won” but that was mostly with a dead vehicle being towed fast by a Corvette). So I think Stanford, somewhat, followed the Caltech approach and CMU, somewhat followed the MIT approach, which is interesting because I wouldn’t have predicted that.

    IIRC in the next challenge the setting was urban environment and driving in traffic and MIT actually had an entrant in that challenge, but I don’t remember how it turned out.

    Haven’t heard much since the DARPA challenge except for Google’s groundview auto-car, but I wonder what progress has been made.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s