Autonomy. It’s Coming, But Will It Be Any Fun?
There’s a lot of talk in the news and tech world these days about self-driving cars, autonomy, and artificial intelligence (AI) in general. About how all these things are going to make our lives better, or take our jobs, or kill us. But, it doesn’t matter because there’s nothing we can do to stop it. So, buckle up, sit-back, grab a virtual reality headset, and enjoy the ride. Right? Hmmm. Before I fasten my seatbelt, let me make an argument about a missing part of the autonomy/AI policy and investment equation: the part proportional to my fun. I’ll line-up to pay for fun. Efficiency, not so much.
We make all kinds of choices based on data, usually mixed with some emotion, but mostly with some facts in there somewhere. Some of us, including consultants like me, recommend choices based on data for a living. The trickiest part of this game is figuring out what in the data tells you anything about the bigger choice at hand. For example, let’s say my client just loves the concept of future 5G wireless (there’s that emotion I mentioned) and wants to make a bundle off of it. If I had the right access to all the phone and network companies, I could tell him which phone and network designs would have the lowest data packet latency. Does that guarantee his $10 million investment turns into $100 million? Hardly. The real magic lies in the use patterns: what will people do with that latency and what are they lining up to pay for? That requires some art to go with the science.
Yeah, yeah. What’s all that have to do with autonomy? Answer: the missing term in the equation. The one proportional to fun. Costs proportional to fun get people to eagerly line-up.
There are benefits to autonomy in all sectors. However, there are also costs to people—negative terms in the equation that reflect taking away fun and putting in mindless efficiency. This matters because, unlike the industrial revolution, we are close to being able to automate everything we do: food service, transportation, medical advice, stock tips, marketing, even war. So that begs the question, what’s left for people to do? Recreate, educate, atrophy, agitate, friend-list robots? The possibilities are endless (cf., Terminator, WALL-E, Robots and Empire, Transformers, Do Androids Dream of Electric Sheep?).
Luckily, we can go beyond science fiction, look at the autonomy we now know is coming, and make intelligent policy and business decisions. To be concrete, consider the Los Angeles area freeway system. Would Angelinos be happy if they could get a driverless car to autonomously whisk them during rush hour the 30 miles from Sylmar to LAX in 18 minutes? Oh, yes. Autonomy visionaries see platoons of self-driving cars moving at high speed nearly bumper to bumper. There are problems with realizing this: traffic bottle-necks, empty extra-capacity cars in the way, getting from home to the “fast lane”, and human-driven cars. However, there are work-arounds: more smart-road infrastructure, smaller more maneuverable cars, and autonomy-only zones. The current system was built for human drivers and one built for autonomous vehicles would look much different—and would not be very friendly to human drivers if they were allowed at all. If the cost benefit equation only captures ruthless efficiency, the policy argument to build a road system for autonomy only is very strong. Is anyone squatting on mahd.org?[1]
Driving in rush hour isn’t fun, so I may just not be accepting the inevitable. But first, let’s see what we will be missing with all this autonomy on the road. In a world built and optimized for autonomous vehicles, uncertainty is the enemy, and humans are uncertain. To get full efficiency, say good-bye to your Ferrari convertible. Blasting through gears and hitting 60mph before the next intersection[2] disrupts the calculated acceleration curve that gets your auto-electric conveyor safely through while clearing cross traffic by inches. Think that only applies to rich? Well, also say good-bye to your Sunday motorcycle ride. Yes, Yamaha thinks they can build an autonomous bike,[3] but what’s the fun in that. Remember, no wheelies, uncommon actions add complexity and hence uncertainty. Off-roading in your 4x4? If it has a manual mode, it’s probably not street legal. If autonomous transportation-as-a-service fully penetrates the market, average people won’t even own cars. Oh, and two colors, Uber’s and other.
As we go into the future, it’s important to remember the implicit assumptions we’ve made about the value propositions in autonomy. To illustrate business value alternatives to efficiency that value people’s fun, consider how we eat. Today, there is no need for anyone to go out to eat. In fact, going out is a terribly inefficient way to get calories. Just order the groceries or take-out you need at home—soon to be drone-delivered. In the early 1960s, the money people spent on eating out was dwarfed by the amount spent on eating at home. Today, even with online ordering and how-to cooking shows galore, the money spent on eating in and out are about equal, $600 billion each. Patterns of life changed and the eat-out folks offered both convenience and experiences with social and business value. The food sector has a human-focused term in their investment decision.
The take away isn’t to stop autonomy, but to make decisions considering what people lose in addition to what they gain. Yes, I’d like to get to LAX faster, but also consider other things than efficiency gains. Maybe Ferraris and motorcycles are doomed, but what replaces them in the economy and as individuality statements by their owners? Good policy and business questions are: how do we gain efficiencies while simultaneously opening up new possibilities for social and leisure activities (i.e., fun)? Remember the way the world works today: I work; I get money; I spend money on needs and fun; repeat. Forgetting fun and ruthlessly designing for efficiency is the first step down the road of engineering a world without roles for people. That is unstable, and no fun.
[1] Mothers Against Human Drivers (prognosticated)
[2] By the way, there are no more traffic lights.
[3] Note comment on “mandatory” autonomous mode in linked reference.
romaincrisis
baratheonpilled
tangentwithtanner