Why human (tech) systems are copying nature – my #SuperNova18 story
Now that I’ve had time to process the explosion (DYSWIDT) of future trends and insights at Supernova last week, I wanted to share some of the patters that I saw emerge during the talks. Those that I found most intriguing were the ones that manifested themselves that at the intersection of nature, human systems and technology.
The accepted discourse about the “race” between human and technological competence is that tech beats us at efficiency. And that we, “soft” humans are better at emotions and creativity. First of all, there’s a lot of creative software out there already, like Google’s Magenta project. Second, Supernova taught me that nature in general, and the human body in specific are still a lot more efficient than most of our hi-tech tools.
Nature is a lot more efficient than humans and their tools
Yes, it’s true that, on a repetitive cognitive level, algorithms tend to function more effectively than humans, especially when it comes to analyzing at scale. But humans still excel at energy efficiency in a lot of domains: the cost of our actions is in most cases lower than their returns, while that is often not the case in technological environments. Several Supernova speakers tackled this significant challenge.
Bram Vanderborght - also a frequent speaker at our innovation bootcamps – showed us how the smart design of our bodies makes us a lot more energy-efficient at for instance walking than the most sophisticated robots. Imec, in its turn is working on neuromorphic chips that imitate the complexity of the brain in a ‘deep neural network’, but in a way that is a lot more energy-efficient than most of the existing solutions. According to Pieter Abbeel (not at Supernova) - a frequent speaker at our tours – it would take about 5,000 dollars per hour of cloud computing power to create a human-level AI: that’s a lot of money, and a lot of power. Not very cost-efficient. Tim Urban (Also not at Supernova, sorry. But his input is really relevant here, you guys.) too claimed that one of the world’s fastest supercomputers – China’s Tianhe-2 – does have the power and the speed to equal a brain’s raw computing capacity. However, “Tianhe-2 is also a dick” (not my words; Tim's): it takes up 720 square meters of space, uses 24 megawatts of power (the brain runs on just 20 watts), and costs $390 million to build. This lack of energy economy is also still one of the biggest challenges of the blockchain, by the way: the amount of processing per day in the Bitcoin blockchain is equal to the energy consumption of the country of Ireland. Neat. Just try for instance to extrapolate that to the ocean of data from the IoT and you’ll surely conclude that we’re not there yet.
There are many exciting technologies, that can do a lot of exciting things. But for them to really change our world, we need to fix their cost-inefficiency. Tim Harford (Sidenote: Another Tim! I just realized that I have rarely met a stupid Tim. If you’re expecting a baby boy, you know what to do.), who opened Supernova, perfectly illustrated this conundrum with what he called "the toilet paper principle": once a technology is cheap enough to wipe your bottom with, it’s cheap enough to change the world. But for the moment us humans are still a lot more (cost-)efficient, and thus cheaper at a lot of tasks.
Human systems are inherently flawed
I have always been intrigued by this paradox: though our human biological system is beautifully efficient in all its complexity, our race seems so very unskilled at building our own systems. I think that no one can argue that a lot of our economic, social, technological and company systems are quite sub-optimal. So it should not come as a surprise that a lot of us are turning to the smartest kid in class when it comes to systems organization: good old Mother Nature. Just think about how efficiently a flock of sparrows, our ecosystem, an ant colony, or even the biological system is organized. There’s even an entire interdisciplinary field devoted to imitating the models, systems, and elements of nature for the purpose of solving complex human problems: biomimetics or biomimicry. Nature is highly resource efficient, both in energy and in materials, as described in Biomimicry 3.8’s Life’s Principles. And it’s just a lot better than us at managing at scale. Mickey McManus brilliantly explained how we ought to learn from nature if we want to manage the scale of the IoT “waking up” with intelligence:
This morning, I interviewed the fabulous Leen Gorissen (stay tuned in the coming days for the interview here) about biomimicry and she summarized this perfectly: humans, and their systems, degenerate the planet, while nature is a regenerative system. In other words: humans extract value, while nature adds value to the system.
"The third Industrial Revolution will be about life" - Nina Tendon
No surprise, that at this very pivotal time, we saw quite a lot of biomimetics examples at Supernova. Nina Tandon explained how EpiBone is the world’s first company growing living human bones for skeletal reconstruction. “We’re just mimicking a natural process, using the cells that grow the bones every day in our body to do it [outside]”, she explained. While competing companies manufacture inert devices or cells without structure, EpiBone does both. That’s why Tandon believes that the third Industrial Revolution will be about life (the first being about machines, and the second about data). There were so many examples at the intersection of nature and tech: imec’s MEA chip, Jan Rabaey’s brain computer interface and human intranet, Charlotte D’Hulst’s biological nose on a chip, Angelo Vermeulen’s interstellar Spaceship on an asteroid: all of them are mimicking nature in their own way to make technology more (cost-)efficient.
Bram Vanderborght also told us how his team is developing soft robots out of flexible, soft material, making them suitable for applications in uncertain, dynamic task environments, including safe human-robot interactions. The really cool thing about tese is that they consist of self-healing elastomers that 'fix' microscopic and macroscopic damage on the basis of healing capacities found in nature. Another thing tech is so inefficient at: the cost of fixing appliances is often so high that we just throw them away and buy new ones. (Unless they are extremely expensive like Tianhe-2. You do not throw away Tianhe-2 when it malfunctions, if only because you’d need an enormous bin.) What if our appliances would be self-healing? I can imagine that the good people at HP and Brother would NOT be amused by this.
The most visible example of the marriage between humans, technology and nature was the very colourful (pun so very much intended) Neil Harbisson, the first person in the world with an antenna implanted in his skull and the first to be legally recognized as a cyborg by a government. Though I was intrigued by his crusade to “become technology”, I only really understood the possible impact of us becoming cyborgs when he claimed that “our species would need to evolve from changing and designing our planet to our needs so we could survive to (re)designing ourselves for that purpose”. Night vision would allow us to cut down on electricity. Regulating our own body temperature instead of that of our house, would do the very same. Modyifying our skin could help us better withstand radiation (that one actually came from Angelo Vermeulen). Yes, it does sound very sci-fi, but if you free your mind of prejudices, you’ll come to see that adapting to a system is a lot more efficient than having the system adapt to you, which is what we have been doing for decades and which will destroy us if we keep doing that. We should be less afraid of AI stealing our jobs – which still does need a place on our "to do" list obviously – than about us destroying our home. Last time I checked, we haven’t got a Planet B yet. (Yes, yes, Elon is working on that, but we’re definitely not there yet).
I like this blurring of the lines between nature and technology. Though one challenge to overcome is surely that we do not yet understand everything in nature, and thus cannot copy it on all levels. Like AI, for instance: we have deep neural networks copying the functioning of the brain (Robovision seems to be doing great work in that field, as Jonathan Berte explained at Supernova) but, on the other hand, we have yet to completely understand how the brain works. But that’s another story.
One of the answers to our inefficiency problem – on the technological level but also on the level of human-designed systems - might be that we should start thinking more holistically about the humans, nature and technology ecosystem. We once made the mistake of viewing of the mind and body as separate - we have Aristotle and Christianity to thank for that – and now a lot us are still doing that with the nature-tech duality. If we are to believe the good people at Supernova, this gap might be bridged very soon.
In 2010, nexxworks Partner Peter Hinssen wrote that technology was becoming normal in his bestseller 'The New Normal'. I believe that we are on the verge of tech becoming natural, and in the process - hopefully – adapting (and helping humans adapt) to its natural environment instead of the other way around. To quote Leen Gorissen: "let's stop extracting value from the system and start adding it."