Reflections - 9th of November.
The Power of Uncertainty
There is something special about autumn, as the countryside we are so familiar with changes before our eyes and begins the alchemical processes that will result in spring.
This year, though, it is not just the trees that seem to be changing. There is a distinct shift in tone from politicians and the press, a quiet recognition that the stories we have been taking comfort in are no longer credible.
“Anyone who believes exponential growth can go on forever in a finite world is either a madman or an economist.”
— Kenneth Boulding
Or perhaps a Tesla shareholder.
There was something both fascinating and faintly grotesque about seeing Elon Musk dancing with his robots as his shareholders approved a trillion-dollar payday. It felt like performance art for an age that confuses motion with meaning, the spectacle of progress masking a deeper decay. It is perhaps the clearest sign of how far regulatory capture has gone among the “Magnificent Ten” technology firms, and how much the American model of growth through story rather than substance has become isolated from the world it claims to serve. And as our retail markets accelerate into Christmas, I am reminded of Greg Lake’s I Believe in Father Christmas from 1975:
“And I looked to the sky with excited eyes / ’Till I woke with a yawn in the first light of dawn / And I saw him through his disguise.”
Perhaps we are reaching a kind of Newtonian realisation that for every action there is an equal and opposite reaction. As we chase efficiency, we strip away the fertile inefficiencies where creativity blooms. Our obsession with productivity cannibalises the present, so that when every moment must yield output, we lose the regenerative power of fallow time. Automation atrophies human judgment and craft knowledge, and we lose our relationship with what we create.
In an economic culture that worships marginal gains and compound interest, we rarely see the other side of the ledger, our marginal losses and growing compound disinterest.
Inequality has always existed, but now we seem intent on perfecting it. That same arrogance doubles down when it comes to AI. Geoffrey Hinton, the Nobel Prize-winning scientist behind the neural networks powering modern AI, has offered a warning worth heeding: that the entire economic model for AI depends on eliminating human jobs. In a recent interview, when asked whether the tech giants’ astronomical AI investments could pay off without destroying employment, he was blunt: “I believe that to make money, you’re going to have to replace human labour.”
This is not a critic speaking from the margins but the Godfather of AI himself, acknowledging that the technology he pioneered becomes profitable only when it displaces workers at scale. The issue is not the technology but the story it serves, one that mistakes efficiency for wisdom and growth for renewal.
That does not mean the story is right.
AI seeks to reduce uncertainty. Business hates it, but uncertainty is a curious beast with many facets.
There is aleatoric uncertainty (Alea, as in dice), where we know the possibilities but not the outcome, and epistemic uncertainty, where we lack the knowledge we need. These are the kinds that engineers and algorithms are built to tame. LLMs excel at reducing them.
But the uncertainties that matter most now resist being flattened into data or prediction. Radical uncertainty emerges when we cannot even know what the relevant possibilities are, not because we lack information but because the situation itself is genuinely novel. Dynamic uncertainty arises when the very act of engaging changes the terrain, when feedback loops and emergence make prediction futile.
These are autumn’s uncertainties, fertile, generative, alchemical. They reward not optimisation but adaptability, not efficiency but resilience, not extraction but exploration. Unlike the mechanical uncertainties that machines excel at resolving, radical and dynamic uncertainty demand human judgement, craft, and the courage to act without calculable odds.
Autumn makes visible what summer conceals, the architecture of trees, the exhaustion of growth, and the necessity of letting go. We are entering a season where our societal structures are revealing their skeletal forms. The marginal losses imposed by technology that erode our essential humanity and creativity are like leaves falling, each one seemingly insignificant, but collectively exposing the reality beneath our canopy of progress narratives.
Mycelium works through autumn and winter, not despite decomposition but through it. Every loss in our human systems, a craftsperson replaced here, a local shop closed there, a skill atrophied, a relationship automated, feeds an underground network of consequences. Like mycorrhizal networks linking forest roots, our losses are interconnected, sharing information about systemic failure in ways no metric can capture.
Compound disinterest grows above ground, while what might renew it begins to form below.
Autumn is actively alchemical, breaking down last year’s certainties into next year’s possibilities. Hinton’s warning about AI’s economics is potent: we are being asked to compost our own livelihoods to feed a system that mistakes decomposition for growth. But nature shows us that decomposition without regeneration is simply rot.
I asked Claude to explore the balance between exploiting existing capabilities and exploring new possibilities, and the evidence it surfaced tells a story worth heeding. The rich seam of exploitation is becoming exhausted.
Since the 1950s, R&D spending has increased sixfold, yet productivity returns per research pound have fallen to 85 per cent of 1970 levels. Innovation has shifted from breakthrough to incremental, with most organisations now devoting 85 per cent of resources to optimising existing operations rather than the 50 to 70 per cent balance that yields optimal performance. Rising costs and risk aversion have driven firms and academics alike towards safer, predictable outcomes, even as this strategy produces diminishing returns. The result is a competency trap, organisations running faster to stand still, caught between short-term financial pressure and long-term creative decay.
The jobs the Magnificent Ten seem so eager to eliminate have already been hollowed out. Teachers are forced to teach to the test. Doctors are measured by process rather than healing. Farmers governed by supermarket spreadsheets instead of seasons. Work designed for efficiency, not meaning.
LLMs will not bring us an economic spring. That will come from those who use them to connect ideas, not extract them, to create something new and more inspiring than a megalomaniac and his robots.
Back in 2007, John Robb warned that our interconnected systems were making us both powerful and fragile. His idea of the “super-empowered individual”, someone able to use global networks of finance, communication, and technology to act with the reach of a state, was born in the context of terrorism but now suggests a different context.
In the AI era, the same forces could empower creators, small businesses, or even single artisans to compete with large-scale corporations. The very infrastructure that once enabled global guerrillas now allows a different kind of insurgency, individuals who use AI as leverage not to destroy systems but to build new ones, resilient, decentralised, and human-centred.
At a time when it is easy to feel under siege from technology, bureaucracy, and leaders who hide their deficiencies behind authority, it is worth remembering the season we are in.
Winter is coming, yes, but beneath the frost, ideas are already taking root. What we choose to tend will decide the spring.
Over at the Athanor, we are exploring what it takes to develop and harness our inherent creativity at a time when it looks like it will not only be necessary but welcomed outside the walls of the extractive economy.
Note: My observations on uncertainty are based on
excellent”Strategy in Praxis” Substack.