While I am certainly interested in the Singularity, and will frequently post ideas about what is coming…I find that I am most often incited to post something about a related but very different topic. Other “Futurists” and how they think (or fail to think).
Lest this come across completely snarky, I want to stress that my emphasis is not on the low quality of the average attempt at Futurism. I don’t wish to call other “self proclaimed futurists” on the carpet or embarrass them. That is not my goal.
My goal is to educate you, dear reader about the perils of lowering your bar on verisimilitude in Futurism and more specifically on the Singularity.
Most often when one calls out a Futurist or one of their articles, it tends to be an effort to reign the person in…bring them back to “reality” after they have overstepped the boundaries of common sense.
In this post I wish to do quite the opposite. This post was incited by an article I read recently about the new RoboCop movie that I saw at the theatre today. The movie was just OK…the article was the worst kind of pseudo-science (the completely non-scientific variety).
In this article, the author supposedly quotes an alleged and self-proclaimed “Electrical Engineer and Neuroscientist” by the name of Professor Charles Higgins. From the good professor’s website, it looks like he is doing some cool things…but the alleged quote in the above referenced article is “right but completely WRONG”.
Alleged Quote: “…unless somebody discovered a battery with 100,000 times the energy density of an iPhone battery in my iPhone, you won’t be seeing it [RoboCop] anytime soon—and certainly not by 2028, which is when the movie takes place.”
Now I want to go on record here that it is VERY unlikely that within the next 14 years we will see the likes of a RoboCop. Forget about the difficulty of hooking a nearly dead brain and pile of organs into a fast-moving robot. There are plenty of reasons it will take longer than 14 years for us to get to this point.
But battery-life is what this guy is going with? That’s why he thinks we wouldn’t be able to build a RoboCop in 14 years???????? Seriously? Battery-life?
Ok, granted — today we don’t have battery technology capable of running a slightly larger than human robot all day. Agreed. That’s where the sciency part ends and I part company with the good Professor.
Since I don’t know if dear miss Watercutter (a nome de plume I can only hope) misunderstood him, or mis-quoted him, or took something completely out of context. I don’t know how this catastrophe went down…but it is a complete fail of a quote.
While blogging has value by reducing the barriers to entry for would-be authors…it also lowers the bar when it comes to things like fact-checking and accurate quoting. Anybody can fancy themselves an author, but it certainly puts a great deal more of a burden on the reader. Gone are the days when easily found writings are likely to be either labeled as fiction or “close enough” to being accurate.
To give just a hint at how lame and exaggerated this “100,000 times energy density” claim is, let’s look at some actual science. Wikipedia has an article about human power, in which it is stated “A trained cyclist can produce about 400 watts of mechanical power for an hour or more”. This is from actual measurement provided in a cited mechanical engineering reference book, so I feel fairly comfortable using it as a ballpark value from which to start.
Let us say that rather than starting from this “heroic” level of human output, we will assume that RoboCop has “the power of ten men”. Then we will stretch it a bit more up to a simple round number of 5 kilowatts. The measured output for a trained cyclist limited the output to one hour, which we all agree would not be adequate for a practical RoboCop, so let’s decide that he needs an hour to run to work, an hour to run home to his charging station, and an 8 hour “tour” on work. That gives us a 10 hour goal. Even if we pretend that this 5 kilowatt output would be sustained relentlessly throughout his strenuous workday, we still wind up with a rather manageable 50 kilowatt hours of energy we need to store.
So that’s our starting point. We need a battery of practical size (something that would fit in and be carryable by a RoboCop sized Borg-ish beast) that can store 50 kilowatt hours of energy on a single charge.
The Tesla Model S electric sports car (again citing a Wikipedia article with fully cited statements of capacity) comes with three different sizes of battery: 40kW-hr, 60kW-hr, and 85kW-hr. So, our current battery technology has the ability to create a battery that can power a RoboCop all day…but it is too big and heavy. The size of the battery is roughly 15 cubic feet and approximately 1000 pounds (depends, of course on which kW-hr size we’re talking about, but we’re just going for very gross approximations here).
I did a substantial amount of googling, but couldn’t find a definitive number for the physical size of the Tesla Model S battery. The best I could find was a guess based on the location of the battery in the car and the approximate dimensions of something that would fit there. That approximation was 15 cubic feet.
According to Wolfram Alpha, the volume of a typical human body is about 2.4 cubic feet. Robo is BIG compared to a human. I’m guessing his volume is on the order of maybe 4 cubic feet. If we say 1/4th of that is taken up by his “human bits”, we have about 3 cubic feet of electro-mechanical robot stuff. If we go with a wild approximation, let’s say that maybe there is room for 1 cubic foot of battery in a practical RoboCop. Perhaps the one shown in the movie had less than that, but we’re talking about something that would fit in a large backpack here. Maybe it wouldn’t look as awesome as the movie version of RoboCop, but I think all would agree that a battery the size and weight of a large back-pack would be practical for Version 1.0 of RoboCop.
A one-cubic foot battery made up of similar elements to the Li-ion battery found in a Tesla would weigh in at roughly 67 pounds. This seems quite practical for a RoboCop sized robot to lug around with him for an 8 hour “tour” on the beat.
So where does that leave us when comparing a “practical” Robo-battery to a modern-day battery? With the many approximations we’ve made here, largely erring on the side of caution, we are thinking today’s battery would be approximately 15 times too big and/or too heavy.
Our Professor of electrical engineering and neuroscience was supposedly quoted as saying there was a “100,000 times” difference between what we have and what we need for a battery. Ok, 15 isn’t all that different from 100,000, right? What the what?
By the “back of a napkin” calculations of my own electronics engineering expert (happily, my expert is “built-in”, so he is always available for comment and is rarely mis-quoted by me), the good professor is off by a mere 100,000 / 15 or a factor of about 6,666. If the quote was accurate, I think he should turn in his degree and walk away with his tail between his legs.
So then, when can we expect this major battery-life issue to no longer stand between us and making a fully functional and practical RoboCop? I did a little more googling and found a few graphs like this one showing the trend for battery density improvement in recent history. It looks like a rather linear improvement (not exponential by any stretch of the imagination). In broad strokes, the energy density (storage capacity vs. size and weight) seems to be doubling about every decade.
Even given the extremely generous power supply specification I started with of 50 kW-hr, it would take approximately 4 decades of continual improvement similar to that of the past 2 decades to get a practical RoboBattery. That would put us somewhere around 2054 instead of 2028.
In reality, I think there are several reasons to go with a much smaller power specification than the one I used. Surely “the power of ten men” is an exaggeration. Probably a robot with 5 times the power of a man would be adequate for Version 1.0 RoboCop. It is also unlikely that any practicing Cop (Robo or otherwise) puts out their full power continuously during a typical day on the job. Let’s say they average even a still very generous 50 percent of maximum power during a typical day.
With these adjustments to our specifications we have reduced our requirements by half…twice. In other words, we’ve shaved 2 decades off of our expected time to viable deployment. Now we’re at 2034 just by adjusting our expectations for the first version of a practical RoboCop.
Now if you start to look at what drives innovation and technological progress, I would say that there are typically two such primary drivers. One of them is economical in nature, the other has more to do with the things that flow into the technological process.
First the economic aspect of fueling technological progress. It all boils down to how much “activity” there is surrounding a particular aspect of technology. Until very recently, battery technology has been somewhat of a redheaded step-child of technological pursuit. Everybody was annoyed by short battery life and lugging around a heavy lead acid battery…but nothing was actually depending on a solution. We didn’t have a particular model of smartphone selling well or poorly based on its battery life. We weren’t anywhere near the make/break point for replacing 200 year old automotive technology that was polluting the planet and enabling terrorism with…a lowly battery.
Today, compared to even 10 years ago, there is a very strong economic and legislative tail-wind pushing battery technology forward. BIG things happen in technology once a lot of people care about the outcome. We have recently arrived at that point in history where people care about batteries. Future effort will likely dwarf that which has been brought to bear on the problem of energy density. Things will happen more quickly during the next decade than they have in the past decade or two.
Now the other driver of technology is more “sciency”. It has to do with what is at our disposal with which to make some sort of progress. Do we have slide-rules, or do we have computers. Do we have a hammer and a manual lathe, or are we starting to build molecular structures one atom at a time in whatever organization we can dream up? Does the spark of an idea depend on one scientist being introduced to another one at a nerd cocktail party…or do we have a vast network of scientists talking and collaborating across the width and breadth of the planet?
Anything (and everything) that matters to people who have money and want more of it will be the beneficiary of our more modern tools and collaboration going forward. For any number of reasons that have nothing to do with RoboCop, batteries are near the top of this very long list today.
So, when will we be able to make a suitable Robo-Battery? It will, I’m quite certain, be much less than the approximately 16 decades predicted by the absurd 100,000 improvement suggested by our good professor and the decade-long doubling time of the past 20 years.
Given the much more realistic specifications I’ve provided here and the many pressures and resources being applied to the area of battery technology, my guess (and it is just a guess) would be…some time between next Thursday and the year 2020.
That’s my guess. Being a conservative sort, I’m leaning more in the direction of 2020. That’s 8 full years before the 2028 setting of the recent RoboCop movie.