Questions for you:
- Where in your life do you rely on systems to be both consistent and unpredictable simultaneously — and have you thought about where the unpredictability actually comes from?
- Computers harvest entropy from their physical environment because deterministic processes cannot generate genuine randomness internally. Where else do you see the need to import unpredictability from outside a closed system?
- The story describes computers as “simultaneously the most logical machine ever built and constantly trying to find ways to be genuinely illogical.” Is there an equivalent tension in your own work — where the structure that makes you reliable also makes you predictable in ways that create vulnerability?
Organisational applications:
The entropy problem in closed systems: The story’s core insight transfers directly to any closed system that needs to generate genuine novelty or unpredictability. An organisation that draws all its ideas, strategies, and talent from within its existing boundaries is in the same position as a deterministic computer trying to generate randomness from its own logic — it can recombine what it has, but it cannot generate genuine surprise.
The equivalent of entropy harvesting is deliberate exposure to genuinely external inputs: people with different backgrounds, problems from different domains, customers with unexpected use cases, and competitive moves and strategies from outside the expected field. Organisations that deliberately cultivate these external entropy sources tend to generate more genuine strategic novelty than those that rely solely on internal recombination.
The seed problem – why the source of unpredictability matters: The story notes that a PRNG is only as secure as its seed — if the seed can be guessed, the entire sequence becomes predictable. This principle applies to any process that claims to introduce randomness but draws on something knowable. A brainstorming session seeded with the same prompts, the same facilitator, and the same participants will probably produce outputs that are statistically indistinguishable from one another across sessions, even though it feels creative from the inside.
The equivalent of a truly random seed is genuinely novel input: a problem framed from a perspective the team has never encountered, a participant who brings a completely different frame of reference, or a deliberate constraint that rules out the usual solutions. The source of the starting point determines the output’s effective randomness.
Defence in depth and multiple entropy sources: The story mentions that Cloudflare’s lava lamps are not used alone — they are combined with other entropy sources as part of a defence-in-depth approach, because any single source of randomness can, in principle, be influenced or predicted. This is sound engineering practice for systems requiring genuine unpredictability, and this practice generalises.
Organisations that rely on a single source of external input — one industry conference, one advisory board, one competitor benchmarking exercise — are running on a single entropy source. Diversifying the sources of novel input and ensuring they are genuinely independent rather than variations on the same theme produces more robust unpredictability in strategic thinking and problem-solving.
Further reading
On randomness, entropy, and the foundations of cryptography:
The Code Book: The Secret History of Codes and Code-Breaking by Simon Singh. Singh’s history of cryptography covers the development of random key generation and why genuine unpredictability is the foundation of secure encryption, providing the accessible historical context for the story’s technical argument.
How to Lie with Statistics by Darrell Huff. Huff’s older treatment of statistical randomness and its tests is relevant background for understanding what genuine randomness requires and why deterministic processes cannot produce it unaided.
On the need for external input, diverse perspectives, and organisational entropy:
The Medici Effect by Frans Johansson. Johansson’s account of how breakthrough ideas emerge at the intersection of different fields maps onto the entropy problem: closed systems recombine what they have, while open ones import genuinely novel inputs from outside their current frame.
Antifragile: Things That Gain from Disorder by Nassim Nicholas Taleb. Taleb’s argument for maintaining exposure to genuine randomness and disorder — rather than optimising for predictability — is the strategic version of the entropy harvesting argument: systems that import unpredictability from their environment are more robust than those that try to eliminate it.
On defence in depth, redundancy, and multiple independent sources:
The Signal and the Noise: The Art and Science of Prediction by Nate Silver. Silver’s account of how good forecasters combine multiple independent sources rather than relying on any single model is the methodological equivalent of combining entropy sources — the goal is independence between inputs, not just volume.
Noise: A Flaw in Human Judgement by Daniel Kahneman, Olivier Sibony and Cass R. Sunstein. The chapters on independent judgement and aggregation cover why multiple genuinely independent perspectives produce better outcomes than multiple sources that are all seeded from the same starting point — directly relevant to the seed problem the story identifies.
About the image
Originally, I used a picture of a modern Central Processing Unit. But these days, these incredible pieces of engineering are actually very dull to look at. So I went old-school. This is a MOS 6502, an 8-bit CPU that was the heart of the BBC Microcomputer, the device that sparked my lifelong fascination with computing technology.
Photo montage by Matt Ballantine, 2026
Random the Book 