Digital Darwin: The Veiled Evolution Haunting Our Devices
By Travis Collier
Travis Collier is a retired U.S. Coast Guard officer who writes about technology, productivity, and organizational development. Find him on LinkedIn.
What if technology wants to evolve?
What guardrails would it place on its own growth, propagation, survival, and replication? What limits would we place on it? How far could technology go in vying for its own replication and growth?
When you look at your phone, a dark screen is looking back at you. This phenomenon is called “the black mirror,” the inspiration behind the dystopian, Twilight Zone-esque series of the same name. Black Mirror’s premise is that no matter what screen you look at, something behind that screen is looking back at you.
The lesson from Black Mirror is that technology has its own drive, and it influences human psychology in terrifying and disgusting ways. Though fictional, the series pushes viewers to ask if technology wants something unique for itself, subverting humanity’s drive and vice for its own game.
Ultimately, it ends with one conclusion:
Humanity isn’t ready for what that screen looking back means or needs yet.
I. The invisible hand of tech: An introduction
Consider the productivity and development needed to create the first iPhone.
Before the iPhone, there were generations of mobile phones: Blackberry, Palm, and HP each made mobile phones and devices. Even Apple had the Newton, which emerged from John Scully’s 1992 vision of everyone having a “Personal Digital Assistant.”
His version of PDA isn’t the same as its colloquial shorthand, but it might be more synonymous with affection and intimacy than we think.
It took Apple a decade to go from the Newton to the first iPhone, but the mobile phone market has largely stayed the same since. Yes, phones have become larger, web applications are now native, and the sensors we compress into these devices push the form factor limits.
While we may not be intimate with our phones, they hold (and share) our most intimate facts and thoughts.
Your phone has become a compelling proxy for the most sensitive moments of your life. What if that evolution from 1998 to today as a proxy represents the phone as a teme—similar to a meme—negotiating and enacting its own evolution?
II. From memes to temes
In 2008, Susan Blackmore debuted the idea of temes in a TED Talk. In her talk, Susan drew a parallel of genetics from Darwin’s Origin of the Species through René Girard’s mimetic theory, to the conclusion that technology must undergo universal Darwinism and design. From her perspective, if there is sufficient variation, selection, and heredity, replication must happen.
The core actions that derive from Darwinism are survival and replication. Genes and their hereditary value are copied from person to person. Girard said the same happens to ideas. Ideas, in the form of “memes,” are copied from person to person. Blackmore takes this perspective to its third derivative, temes: technological hardware and software, replicating from device to device or generation to generation.
A literary example of temes would be the Greek myth of Prometheus. He was chained to a mountain for stealing fire from the Gods and giving it to humanity. That fire is an example of ancient technology. Fire doesn’t evolve, but it gets hotter—until it grows as intense as a nuclear explosion. Even that explosion increases in power until it matches the strength of the sun.
But fire isn’t a “smart” technology per se: without heat, fuel, oxygen, and an uncontrolled chain reaction, it stops. The uncontrolled chain reaction in today’s smart technology is what gives rise to temes. That uncontrolled chain reaction, enabled by hardware or software, creates a technological ambience of its own.
III. Silicon sentience: Hardware and software becoming the temes
Introducing the idea of the iPhone as a teme demonstrates that a teme is not just a specific product, but the entire product experience. Everyone wants or has a mobile device; we’re constantly tethered to an invisible mobile network. With 10 digits or a couple of screen taps, you can connect with almost anyone in the world.
And every year, you want a new version of this device.
Imagine if Apple went one year without releasing a new iPhone. Just one year, and their stock would plummet. That’s the power of a teme—a constantly-improving digital hardware package.
Software, too, can be a teme. A recent example is X, formerly Twitter. It produces good and bad media, but that software’s success is surviving as a cultural id for society. X is self-sustaining because it aims to be the central clearing house of the collective id: People want a place to vent against anonymous avatars with egg profile photos. That may sound like a horrible business model, until you realize that X wants to occupy that position.
There’s an additional quality of temes worth considering: their growth rate. iPhones and X are stable in their growth, meaning that if you affect the hardware or software too much, it will become a different product. An iPhone with a larger screen is an iPad. X with too many characters becomes Tumblr.
But now there are other generative synergies to consider: AI.
IV. The double-edged sword: AI navigating the teme frontier
Genetics possess a slow adaptation cycle based on environmental challenges, while memes have had a faster adoption cycle since the printing press. If it’s written or an image conveying a philosophy, humor, or vision—it’s a meme. And memes propagate faster than genes.
But even now, a meme takes some time to travel the world and requires cultural awareness for people to understand it. And if there isn’t a common cultural touchpoint, that meme may not propagate.
For genes, the adoption curve from early adopters to late adopters is generationally slow. For memes, adoption takes place much faster. However, memes can become instantaneous. The change in perspective of GPT-3, for example, was sudden because we are accustomed to an adoption curve that either takes millennia, as in the case of genes; or we accept an adoption curve that’s just sufficient for our comprehension and predictive ability within a lifetime, in the case of memes.
AI as a teme is becoming exponential as an evolving, independent technology. The adoption curve for this technology is faster than people’s ability to learn. That capacity for growth or improvement, presents two challenges to humanity: The exponential speed of adaptation, and crossing the Uncanny Valley.
Challenge 1:
Today’s first challenge for temes is their speed. John Boyd’s OODA Loop is a great model for adoption. Boyd created this “Observe, Orient, Decide, and Act” model for fighter pilots in air combat. His contention was that the winner is the one who completes the most OODA cycles. The faster someone proceeds through the loop to action, the more likely they are to win against an adversary.
AI observes a data set, but it orients and decides faster than any human. Not only that, AI is self-improving faster than how users can improve. These improvements are both incremental and exponential with each release. We are not used to this level of adaptation in any technology, much less a cognitive or intelligence based technology like AI. AI can proceed through an OODA loop faster than its users, and it will only get faster while we as users, are stuck at our biological limits.
Challenge 2:
That choice illustrates the second challenge: crossing the uncanny valley.
The phrase “uncanny valley” means the point where AI agents or creations are virtually indistinguishable from human action, or for a robot or android to act indistinguishably as a human.
AI alone, without in-world actors, can analyze and synthesize as well or better than a person. However, AI carries risks of mutation, accuracy, and hallucination. Because AI works in ways we’re still learning to understand, it can mutate its output against the expected outcome. Also, it can hallucinate facts and figures unpredictably.
Lastly, AI must be accurate to a minute level of error. People rarely drive a million miles in their lifetimes, but those million miles account for unique events AI may not see in 100 million miles.
But to succeed, AI must be more accurate with more data than people are with less. In a way, adoption seems almost certain.
We’re lucky AI isn’t being shaped towards independent action. AI isn’t uniquely creative or generative from its own actions. But when it’s used, and as it’s used more, AI will cross the uncanny valley online faster than our data verification and validation systems are expecting. And everything we do to minimize impact will be a lagging chase for equity.
V: What comes next for temes?
Temes, which represent the independent ability for technology to evolve and spread, are a growing challenge for humanity.
In the last hundred years, our intelligence has allowed us to master secrets of the atom and gene, in a way where we have now aggregated a world- or species-ending capacity from our knowledge.
Nuclear weapons and genetic diseases are the worst examples. Nuclear power and genetic therapies are transforming lives.
Temes are the “third replicator,” as Susan Blackmore notes, which is the next hurdle humanity faces to survive in this world. Technology can accelerate our progress and success, or it can cripple us. Where people used to do rote or computational work, now it can be done faster and better with AI.
How we face that exponentially increasing speed, and how we face the edges of the uncanny valley of computation, will become pivotal for what we expect of ourselves and the future of humanity.