The Relocation of Bounds
The relocation thesis is this volume's central analytical claim: AI has relaxed three of the four constraints that produced bounded rationality — information, computation, and time — while leaving the fourth, attention, intact. The result is not the transcendence of bounded rationality but the relocation of its binding constraint. In the old world, the binding constraint was generation: the builder could not produce alternatives fast enough. Information was scarce, computation was expensive, implementation took months. In the new world, the binding constraint is evaluation: the builder cannot assess alternatives wisely enough. Generation is cheap. Evaluation remains as expensive as it always was, because evaluation requires attention, and attention is a property of consciousness rather than of the information environment. The relocation produces the specific asymmetry of AI-augmented work — unbounded generation meeting bounded evaluation — that drives every behavioral pattern the volume analyzes.
In the AI Story
The relocation is invisible from the generation side. The builder experiences expanded capability: more information, more computational support, more speed. The expansion is real and consequential; it produces the twenty-fold productivity multipliers that The Orange Pill documents. The experience of unbinding is authentic. It is also partial.
The relocation is visible only from the evaluation side. The builder who examines her own cognitive state after a day of AI-augmented work notices that she has generated extensively and evaluated shallowly. The alternatives she rejected were rejected quickly. The alternatives she accepted were accepted without the depth of scrutiny that the old generation-bottlenecked workflow naturally imposed. The evaluative bottleneck is less dramatic than the generative unbinding, but it is more consequential for the quality of what gets built.
The thesis connects Simon's framework to every other volume in the Orange Pill cycle. Edo Segal's ascending friction describes the same phenomenon phenomenologically: the difficulty that AI eliminates at the implementation level reappears at the judgment level. Simon's contribution is the structural explanation: the bounds have moved, because bounded rationality was never about a single constraint but about the cognitive architecture of finite minds, and that architecture has not changed.
The practical implications are specific. Organizational structures designed for the old bounds — specialist silos, sequential handoffs, functional decomposition by generation domain — are becoming obsolete because they optimize for the constraint that has been relaxed. New structures are emerging — vector pods, judgment-focused roles, evaluation-centric workflows — that optimize for the constraint that remains.
Origin
The relocation thesis is not Simon's own. He died in 2001, before large language models existed. The thesis is a straightforward extension of his framework: if bounded rationality has four constraints and a technology relaxes three of them, the fourth becomes binding by process of elimination.
The specific framing — relocation rather than removal — matters because it preserves the architectural insight at the heart of Simon's work. Bounded rationality was never about specific constraints that could be eliminated. It was about the structural fact of finitude. That fact has not been changed by AI. It has been made more severe, because the finitude now operates in an environment that actively overwhelms it.
Key Ideas
Three bounds relaxed. Information, computation, and time constraints have been substantially expanded by AI — not to infinity, but enough to constitute a qualitative change in what individual builders can generate.
One bound intact. Attention has not been expanded, and cannot be expanded without fundamentally altering what it means to be a conscious agent.
The binding constraint has moved. What previously limited what could be built now limits how well what is built can be evaluated.
Organizational structures lag. The hierarchies, handoffs, and decompositions designed for the old bounds continue to exist, becoming overhead rather than support.
Architectural response required. The relocation demands new structures — educational, organizational, personal — that optimize for evaluation rather than generation.
Appears in the Orange Pill Cycle
Further reading
- Simon, The Sciences of the Artificial (1969, 1996)
- Simon, 'Designing Organizations for an Information-Rich World' (1971)
- Edo Segal, The Orange Pill (2026)
- Daniel Kahneman, Thinking, Fast and Slow (2011)