Lean thinking minimizes invested capital and maximizes return. It enables closed-world learning — tight feedback loops, known failure modes, predictable recovery times. Fat thinking deploys more capital than necessary and accepts lower returns. It pays for open-ended learning — unpredictable recovery times, unknown fruits, and the possibility of discoveries that rewrite the rules.
The slack in a fat system is not waste. It is the antifragility budget.
Simple Picture
A lean bridge uses exactly the minimum material to hold the expected load. It is efficient, elegant, and collapses when the load exceeds the engineer’s prediction. A fat bridge uses more material than necessary. It is “wasteful” — until the day an unexpected load arrives, and the fat bridge holds while the lean one fails. The extra material was not waste. It was insurance against the limits of the engineer’s imagination.
Leak-Before-Failure
The engineering concept of leak-before-failure is designing systems with controlled, non-critical failures that prevent catastrophic ones. A pressure vessel designed to leak at a weld seam before it explodes has deliberate weakness in one place to create safety everywhere else. The leak is a designed inefficiency — it costs performance under normal conditions and saves lives under extreme ones.
This applies to organizations: the Theory of Constraints reveals that for a system to flow, non-constraint resources must have excess capacity. That excess capacity is the organizational equivalent of leak-before-failure — it looks like waste but is actually the buffer that absorbs shocks, enables learning, and prevents the brittle efficiency of full utilization from cascading into systemic failure.
The Systems Bible principle: a temporary patch will very likely be permanent. Lean systems cannot afford patches because there is no slack to absorb them. Fat systems can — which means fat systems evolve while lean systems optimize. Evolution requires waste. Optimization eliminates it. And the system that eliminates all waste has also eliminated its capacity to adapt.
Closed vs Open Learning
Lean learning is fast, focused, and bounded. You know the question. You know what kind of answer you are looking for. Recovery from failure is quick because the failure modes are anticipated. This is the domain of simplify and accelerate — once you know what works, make it leaner.
Fat learning is slow, broad, and unbounded. You do not know the question. The fruits may take years to materialize. Recovery times are unpredictable because you are exploring territory that has no map. This is the domain of exploration — the messy, expensive, seemingly wasteful phase that must precede exploitation.
The ergodicity connection: lean systems optimize expected return per unit capital, which is the ensemble-average strategy. Fat systems optimize survival probability, which is the time-average strategy. The lean system that maximizes return goes to zero when the unexpected happens. The fat system that wastes capital survives — and the survival is what makes long-term compounding possible.
Economies of Variety
Economies of scale come from repetition. Economies of scope come from market aggregation. Both are closed-ended learning — fast-followers can skip the early high-risk phase. Economies of variety come from variation, with benefits realized as distributed innovation capacity. This is open-ended learning, and it is the only kind that produces genuinely new knowledge.
Some companies appear to “beat Darwin” through variability selection — hardware (organizational structure, culture, slack) that can handle more variety and survive more experiments. These are fat companies. They look inefficient to lean analysts. They also produce the innovations that lean companies then copy.
An economy full of lean companies has stability but little dynamism. An economy full of fat companies has high churn but many innovations. Lean economies have growth reflected in numbers but are zero-sum. Fat economies have unaccounted learnings in nooks and crannies but are non-zero-sum.
The play framework maps: play is significance alchemy that transforms the meaningless into the meaningful. Fat thinking is play applied to organizations — the deliberate creation of slack where unscripted discoveries can occur. peopleware confirms: people perform better when trying something new, and novelty requires the fat that lean systems have eliminated.
Dimwit / Midwit / Better Take
The dimwit take is “efficiency is everything — cut all waste.”
The midwit take is “some slack is useful, but you need to measure and justify it.”
The better take is that the question is not whether to be lean or fat but where to put the fat. Leak-before-failure is about figuring out where to allocate the antifragility budget — which parts of the system should be over-designed so that the rest can fail safely. A company should strive to capture enough value to survive while creating surplus that spills over. The spillover is not leakage — it is the open-ended learning that makes the company worth running.
Main Payoff
The deepest insight: companies should not try to maximize value capture. They should try to capture enough to survive and let the rest create positive externalities — learning, innovation, capability development — that cannot be captured or measured but that make the entire ecosystem richer. The lean company that captures 100% of the value it creates has nothing left to learn from. The fat company that captures 60% and spills the rest is funding the future — its own and everyone else’s.
References:
- Venkatesh Rao, Fat Thinking, Ribbonfarm
- Luca Dellanna, Fragilization