Mythos, Quantum Emergence and the Case for Human‑Centred Design

I always advocate that for every system we design, we must keep the human at the centre rather than merely “in the loop”. This is especially true for quantum technologies, the deeper I go into this. AI is an engineered abstraction; quantum systems are nature itself. When we build quantum technologies, we are not just constructing computational tools. We are interfacing directly with the physical substrate of reality – superposition, entanglement, decoherence and phase transitions that emerge only at scale.

This is where Mythos becomes essential. Mythos describes how intelligence in large‑scale AI systems emerges from architecture, connectivity, and dynamic interaction rather than explicit rules. As models scale, new behaviours appear abruptly – a phase transition in capability. The Perception Game illustrates this perfectly: as more cards accumulate, the system stops adding information and starts reconfiguring meaning, mirroring the non‑linear jumps seen in both cortical networks and frontier AI models.

Quantum systems exhibit the same structural behaviour, but at the level of physics rather than computation. A single qubit is trivial; a million qubits form a high‑dimensional Hilbert space whose interference structure encodes patterns no classical system can represent. At scale, quantum systems stop behaving like components and start behaving like collective phenomena. This is why quantum is fundamentally different from AI: AI simulates intelligence; quantum systems instantiate the laws that make intelligence possible.

For risk management, this distinction matters. Emergent systems – neural, artificial, or quantum – cannot be governed with linear tools or static supervisory frameworks. We need human‑centred design, new algorithms, and new governance models that recognise when a system has crossed from predictable behaviour into non‑linear emergence.

Back in the day, I spent years as a quant fund manager, allocating millions using algorithms and C++ signals long before AI entered the markets. Even then, I learned a truth that Mythos now formalises: large systems don’t behave smoothly – they jump. Markets jump. Models jump. And now intelligence jumps.

This was when I was still in the hot seat: Long‑Term Capital Management (LTCM) proved that large financial systems don’t move in smooth curves – they jump. For two years, LTCM’s models (designed by Noble laureates) assumed stable correlations and predictable spread behaviour. Then, after Russia’s 1998 default, markets phase‑shifted: correlations snapped to 1, liquidity vanished, and spreads exploded by hundreds of basis points. A $4.7bn fund lost $4.6bn in six weeks. Nothing in the models predicted it. The system didn’t drift – it reorganised itself. This was an early demonstration of the same phenomenon Mythos now formalises: in large, interconnected systems, risk emerges non‑linearly, all at once.

Mythos explains why. As AI systems scale, they undergo phase transitions in capability: emergent behaviours, new failure modes, and non‑linear dynamics that traditional financial models cannot anticipate. This is model risk on steroids. Stress tests built on continuity assumptions will miss systems that suddenly reorganise themselves.

Concentration risk compounds this. When a handful of firms control frontier‑scale models, the entire financial system becomes exposed to correlated errors, shared blind spots, and simultaneous failures — the systemic‑institution problem, amplified.

As AI begins to allocate capital, price risk, optimise liquidity, and execute trades, Mythos warns that these systems will not behave like incremental improvements. They will behave like non‑linear agents, capable of generating new feedback loops and flash‑instability.

Quantum systems push this further. Quantum is not engineered abstraction — it is nature. When intelligence emerges from physical laws, not code, humans must remain in the loop.

Because the real danger is not autonomy. The real danger is emergence.

Photo: My team playing the Perception Game which I invented to experience Mythos.

To download “Mythos and the Perception Game” click on this link.