The Cylon (TOS) Design Fallacy

Just to be clear, I'm talking about these Cylons...

... not these Cylons.

In the pilot of the original Battlestar Galactica, Apollo and Starbuck steal a Cylon fighter. They're able to do this because those ships, like all of the others depicted in the show, appear to be designed for human use. They have manual controls, display screens, life support systems, and voice communications—all things that were either unnecessary or horribly inefficient for technology designed to be used by a race of robots.

It's best not to try to think too deeply about that 1978 show. For one thing, we know the reason that Cylons looked like they did. It was because the producer, Glen Larson, blatantly ripped off Star Wars and got sued for it (Larson was one of the few people in Hollywood who couldn't get along with the famously well-liked James Garner—to the extent that Garner once punched Larson in the mouth over plagiarized Rockford scripts) . Cylon centurions looked the way they did because they were supposed to remind people of Imperial stormtroopers.

The Cylon design fallacy is that robots should be designed to interface with systems that were designed to interact with humans and systems designed to work with robots should interface with them in the same ways they would interact with humans.

The standard argument for humanoid bipedal robots, put forward by people like Elon Musk, is that they can do anything that humans can do physically. That's a dubious claim, but even if they could back it up, it's really not that impressive a boast. Most of what we do requires tools, which raises an interesting question: since robots are tools, why pick a design that is far more expensive and less efficient just so it can use tools that only look the way they do because they were designed for humans to use?

Bipedal humanoid robots are inevitably top-heavy and unstable compared to other designs, badly suited for lifting, carrying, and stacking heavy objects—particularly those where the load can shift while moving (think of a half-full container of liquid or a box with loose objects that can slide around). It's true that a humanoid robot could operate equipment such as a forklift, but if you're going to go to that expense, you could get an autonomous forklift that would do the same things at far less than the combined price—with the added advantage of being more compact, since it would not need space for a driver.

As for the argument that the same robot could go from working in a factory to cleaning your house, why would we want it to? You’ll be paying for functionality you don’t need and compromising on functionality that you do. And even if you insist on a one-size-fits-all model, what reason—other than the naivest of naive biomimicry—is there to believe that the sci-fi movie robot is the design you’d want to go with?