Forget what you know about robots (if you learned it from TV)

I participate in a group where we critique each other’s novels. A few weeks ago, my book SuSAn (experimental rewrite) went through review. One of the guys wrote in the margin, “For research on robotics, I’d recommend you watch the awesome television series ‘Humans’.”

I agree with him that the TV series was awesome, and I can’t wait to watch the next season. However, an education in robotics it is not. What do you learn from ‘Humans’? It drops a few hints about the Singularity, and some oblique mentions of Asimov’s laws. (Both of those concepts are fiction, not real science.) Robots are filled with bright blue fluid that acts something like blood when it suits the drama. They can take on several days worth of charge using a cable no heavier than USB. And their software has all kinds of weird loopholes.

I have a PhD in AI and have worked on real robots. In writing SuSAn, I worked out the hardware and software in enough detail to be self-consistent, perhaps even scientifically possible. The story stays true to that technical design, but a novel is not the right place for a lecture on robotics. I ask the reader to trust they’re in good hands and let the clues form a new picture in their mind.

Do I stretch things? Yes, a little. For example, Susan runs on only 300 watts at rest. Nowadays, the biggest supercomputers run at about 15 to 20 megawatts, and they’re probably still too small to simulate a human mind. The only way Susan will be possible is a radical breakthrough in neuromorphic technology.

Another show that’s incredibly wrong is Extant. The child android Ethan runs on two batteries about the size of D cells. And he is so efficient that he actually needs a heater to warm his skin for more natural touch. OMG, people who write this stuff have no idea about the energy cost of computation, or what is remotely possible with battery technology.

Leave a Reply