Will robots be serving us OJ and French Toast in our lifetime?
Most often a writer’s fascination with robots centers around their programming rules or the concept of what it means to be human. Before we get anywhere near those philosophical questions, we must first conquer the challenges of software capable of handling countless unexpected situations and the hardware to drive it all. Our AI software, while making huge leaps this past decade, still has a way to go. However, software might not be the primary problem we face before robotic home workers grace our homes. We first have to solve the issue of powering a big, heavy machine, controlled by power-hungry vision and touch sensors, directing thousands of energy-eating actuators and sensors, in a complicated dance for a full day. Unfortunately, gravity is a harsh mistress. If your Roomba crawls slowly for an hour before needing a recharge, what power would your bicentennial man need?
Science fiction stories, books and scripts are dominated by robots in human form. We want, or perhaps need, to identify with our creations. But only a few of these stories address their robot’s power sources.
Surprisingly, our earliest fiction works detail their power source more than our later stories. We have a crazy early story from the 1800s about a steam-powered mechanical man. A man might have steered it with reins, but Edward Ellis respected the laws of energy in many ways, including “To give the machinery an abundance of room, the steam man was exceedingly corpulent” and “’It won’t run far; the steam is nearly out.’” However it’s a big robot, not one that we could use inside our house. David Keller gets the credit for an early domestic model. The power source? Just plug it into the wall. “She’s made of a combination of springs, levers, acoustic instruments, and by means of tubes such as are used in the radio, she’s very sensitive to sounds. She’s connected to the house current by a long, flexible cord, which supplies her with the necessary energy. To simplify matters, I had the orders put into numbers instead of sentences.”
The Steam Man of the Prairies by Edward S. Ellis (1865) The Psychophonic Nurse by David H. Keller (1928)
Jump forward to the 1950s. We’ve seen what atomic power can do and writer’s begin to… wait a minute, what is this? A reference to atomic power in 1926 by Edmond Hamilton to power 300 foot metal robots. Incredible. “There was no sound to indicate the source of the wheel’s motive power, but Lanier little doubted that Detmold had seized and utilized the same secret of atomic power that had been used by the brain for its own creatures.” Back to the 1950s. Clifford Simak credits atomic power for the rise of all automation, from vehicles to farming to lawn mowers. “No use having gasoline, they tell me, when you have atomic power.” In the real world we are seeing our first nuclear ship and satellite.
The Metal Giants by Edmond Hamilton (1926)
City by Clifford Simak (1952)
Astro Boy (1952)
While we get a glimpse of batteries from Ray Bradbury’s Mechanical Hound in Fahrenheit 451, most of our battery powered domestic robots emerge in the seventies and eighties. However, power is often only mentioned in passing. In 1980 Sladek briefly mentions “[Roderick] hated the little legs, all they were good for was stumping around in the snow until his battery went flat.” C3P0 powers down and Luke plugs in R2D2, but let’s face it, batteries just aren’t as cool as nuclear power. And this lack of coolness or a moral conflict provides two reasons our fiction technology tree avoids exploring the power needed for their creations.
The Bicentennial Man by Isaac Aisimov (1976)
Roderick by John Sladek (1980)
Star Wars (1977) and The Empire Strikes Back (1980)
Batteries might be boring, but fusion energy and other yet to be invented tech gives us limitless power. We settle on technology like this into the 2000s. One reason for avoiding the subject is ironically, readers and watchers want better technical explanations of the often limitless abilities of robots, and batteries just don’t cut it. Kudos to the Iron Man movies, they pay significant homage to the role of power in unlocking future applications of robotics. However, most of our stories simply ignore the power source question and enable our robotic friends to do anything we want.
Iron Man (2008)
Ilium by Dan Simmons (2003)
Ancillary Justice by Ann Leckie (2013)
All Systems Red by Martha Wells (2017)
So which will it be within 50 years?
- Fusion/Future Tech
Let’s eliminate Fusion/Future Tech. Some of you will decry, “We’ve created fusion for a few seconds!” Maybe. But technology must be consistently reproducible, mass produced, and cost effective. If we pull off a joint world Manhattan Project for Fusion, we can update this option.
Time to swing the other way, good ole fashion plug in robots. Honestly, there is no question that this would provide the power we need. It’s well understood, easy, heck, this is the winner. Except that pesky humans won’t accept wires and won’t redo their homes just to accommodate their personal Jeeves. If we were willing to accept this, our bottleneck would move to smarter software.
Nuclear, our old friend that no one is afraid of. Plenty of energy, your home android will have all the Iron Man add ons. There are some promising concepts like Molten Salt Reactors that attempt to reduce meltdown risks if your robot lugs around a chemical factory in its backpack. The smallest nuclear reactor could fit into a building the size of your house, sorta. Unfortunately, nuclear is the kid with the bad rap who won’t receive the investment it needs to overcome these basic character flaws.
That leaves batteries. Boring ole batteries. There is a reason your Roomba vacuum looks the way it does. It is light and therefore requires very little power. Even so, it needs to recharge every 60 minutes. Forget about asking it to climb up stairs, that would require significantly more weight from the parts and power to overcome gravity.
The laptop, mobile phone and now electric car revolution give batteries a fighting chance. First, the cost of batteries has dropped 90% over ten years. Volume, new sources of raw materials, and mass production continues to drive this cost down. Battery sales will cross $100B annually within a few years. More on why this is important in a moment.
We still have a gravity problem. We need more energy stored per ounce. We are ecstatic when our battery performance improves even 10% in a year, far short of our computer processor doubling every year or two. And many predict decades of anemic 1-2% improvements. At that rate our robots might need an hour recharge between cups of coffee. Without creamer. Or sugar.
Fortunately, that large market and volume we reviewed earlier opens up the wallet of our investor community. A big market justifies investment in risky research. Companies like Sila Nanotechnologies have investors fighting over them to invest $1B into new technologies to change that curve. Battery buyers would switch in a heartbeat at a jump of 25% battery performance, making Sila the world’s largest energy seller.
Our personal robot assistant will need about 4KW/h / day for our computer processing, assuming it runs on the equivalent of a gaming PC. Not too bad, but the real killer is the weight of the robot and the batteries themselves. The battery used by the Spot dog by Boston Dynamics currently weighs 10lbs to drive a 60lb dog for 90 minutes. Let’s assume our robot will weigh in at 150lbs without batteries. At today’s storage capacity, to run Robbie for a day, it’s not clear that even 100lbs of batteries would work, as the weight of the batteries themselves would work against our automaton. 100lbs of batteries equals 1/10 the number of batteries in a Tesla 3.
Still, it’s close enough that in fifty years all of that work we are doing on electric car and mobile battery research should cross an inflection point for battery energy density. By that time you will probably be too old to appreciate the change in your routine, and like Gramp in Simak’s City you might prefer to mutter, “Here I am, Mark. Back of the house. Hiding from that dadburned mower.”
Verdict: Science in 50
Leonard Speiser started several technology companies including Clover (sold to First Data) and Bix (sold to Yahoo!) and is an active investor in technology startups. Leonard has previously worked at eBay, Intuit, and CSFB. He graduated from MIT in 1996. Follow him on Twitter: @leonardspeiser or connect to him on LinkedIn: https://www.linkedin.com/in/leonardspeiser/