#49 - You can't cheat entropy, even with AI
The law of nature that your AI transformation cannot escape
I’ve been thinking about life, and processes of creation and destruction a lot recently. And trying to connect that to what I see is happening with organisational transformations driven by AI.
I hear from many product leaders who feel they’re being left behind in this race to AI everything, everywhere, all at once. They have huge pressure from their boards to demonstrate what they’re doing with AI — where is it in the product roadmap? How are you using it internally to reduce costs and improve productivity? How is it impacting the bottom line?
These fears aren’t always being voiced in public forums because of the stigma of being the one admitting they don’t know what they’re doing, or they’re not sure about all the hype. Saying ‘I don’t know’ would reveal a precarious position when the default is certainty. So instead we just hear a lot of noise from the AI bros, a distortion that affects our perception. Just look at your Linkedin newsfeed as the worst offender of this.
Allow me to take you on a philosophical detour momentarily. I recently came across Erwin Schrödinger’s 1944 book, What is Life?. Leaving the cat aside for one moment, he explores how life is an ordered system existing in a universe that tends towards disorder. Life completely bucks the trend of natural physical decay. Left for long enough, our universe will eventually burn itself out. Our nearest star (the Sun) will spend all of its hydrogen in nuclear reactions and die. The giant black hole at the centre of our Milky Way — Sagittarius A* — will eventually leak so much energy through Hawking radiation that it will evaporate in a brilliant flash of gamma rays (or so we believe — our universe is not old enough yet to have seen this happen). In the meantime, DNA takes in this chaos and creates order, through four simple building blocks (adenine, thymine, guanine and cytosine).
All cellular life on Earth is built via DNA, so as far as we know, it’s required to create life elsewhere in our universe too. From those four nucleotides come the incredible variety of bacteria, plants, animals, fungi, and more — everything from penicillin to pandas to pumpkins to portobello mushrooms. Incredible, isn’t it?
But this process takes time — evolution happens over thousands if not millions of years. When DNA is replicating itself, a mutation may appear randomly and spontaneously in the genetic code. These mutations may lead to changes in the phenotype or behaviourally that prove advantageous to the organism during its lifespan, and so it has a better chance of reproducing and passing that mutation onto its offspring. Over time, the mutation reaches a critical mass within the population, with more carrying it than not. Those that don’t carry it are less likely to survive, so die, and don’t pass on their non-mutated code to the next generation.
Evolution as a process of creation (emerging new characteristics) is slow. And when we zoom out to look at other natural creative processes — The Big Bang, star formation, plate tectonics, fossilisation, forming coral reefs — these are on a timescale of 1,000’s to 10,000,000,000’s of years. In other words, building order takes a long time, and there’s little hurrying it.
On the flip side, destruction can be instant. A supernova explosion. An asteroid hitting the planet. A volcanic eruption. A rainforest cleared to make way for grazing land. It’s so much easier to effect, and once the process has started, there’s no going back to the previous ordered state in the same instant.
Our universe tends towards entropy, which is the number of ways the parts of a system can be arranged while still looking the same overall. For example, low entropy is a neatly stacked pack of cards, whilst high entropy is a shuffled deck of cards. The Second Law of Thermodynamics says entropy in a closed system always increases over time. This is why heat flows from hot to cold, why things break more easily than they assemble, and why time has a direction.
I’ll bring this back to AI now. Is AI a force of creation, or destruction within your organisation and our society? Many would argue it’s creative: you can build apps, automate workflows, and learn new skills through engaging with it. But it’s also destructive: cognitive skills atrophy, excessive consumption of natural resources, degrading real human relationships. One results directly from the technology use, the other emerges indirectly from how the system surrounding it adapts to it.
If organisations go all-in on AI transformation, it will take time to build new workflows, capabilities and value generation to leverage this technology — much more time than the current rhetoric would lead you to believe. Yet it will also destroy (and we see this happening in real-time) jobs, trust, institutions, truth, ecosystems. It will take much longer to re-build these again in the aftermath. The technology doesn’t care about this asymmetry.
We should be wary of anything that grows rapidly without constraints. In nature, this looks like cancer, pathogens, invasive species, algal blooms — and they tend to destroy the system they’re growing within, and often themselves in the process. Uncontrolled exponential growth is the signature of pathology, not progress.
So if you’re feeling behind even at this stage, you’re not.
You could absolutely burn down your whole organisation tomorrow if you felt that was needed. Fire all your staff and replace them with AI agents. It would be quick and dramatic, but also painful and irreversible.
The truth of building (and rebuilding) is that it takes time — much longer than can be shown in a single quarterly earnings report. A new pattern (random DNA mutation) needs to emerge that is advantageous to the organisation — and nobody knows quite what that looks like yet. So instead this means we need to engage in deliberate experimentation with a closed feedback loop. Test things out to see what works for you, in your organisation, in its environment (industry, sector, market). Look at what others are doing and spot the patterns of success — and failure. Try those patterns out for yourself. Be wary of anything that looks destructive.
The real question to ask is this: is AI breathing life into your organisation, or destroying it?
What's the AI pressure looking like in your organisation right now? I want to know what leaders are actually being asked to deliver, and what's getting pushed back on. Drop a comment or shoot me a message at caroline@expansum.space