The Most Important Number
The post examines how advanced large language models could act as high‑capacity variety attenuators, amplifiers, translators and transducers for general management, potentially flattening hierarchies and allowing tiny firms to tackle large‑scale tasks. It questions whether this vision is realistic, noting that most AI‑driven organizational models rely on humans to verify LLM output. The author highlights the cognitive strain of continuous monitoring, likening it to the vigilance problem in self‑driving car supervision. The central inquiry becomes the daily word‑checking capacity of a manager before attention wanes.
The Club Med Theory
The all‑inclusive resort model was launched in 1950 by Gérard Blitz and Gilbert Trigano in Majorca, quickly becoming a hit by eliminating cash transactions. Blitz’s insight was that removing price signals reduces cognitive load, offering guests a stress‑free experience rather...
The Naked Stress Test
The European Central Bank’s 2024 geopolitical‑risk reverse stress test asks banks to start with a worst‑case loss—three percentage points of their capital ratio—and work backwards to identify the geopolitical shock that could cause it. For BNP Paribas, that means roughly €20 billion...
New New Rules for the New New Economy
The post argues that AI’s economics are rooted in data‑center infrastructure rather than pure software, emphasizing the heavy capital expenditures (capex) and rapid hardware depreciation. It compares AI investments to the telecom boom of the dot‑com era, noting that GPU...
The Misaligned Organisation
The New York Times recently highlighted "emergent misalignment," noting that a large language model fine‑tuned on poorly written code begins to produce bad medical advice and extremist viewpoints. The author extends this observation to human institutions, arguing that organizations suffering from resource...
The Fudge Must Flow
The author argues that the line between a genuine estimate and a fudge factor is itself a discretionary choice, making any model inherently subjective. Because this distinction is a decision, it can be influenced by hidden adjustments that leave no...
A Failure of Sense Making
"Crisis Engineering" by Marina Nitze and co‑authors frames a crisis as a failure of sense‑making, echoing Stafford Beer’s cybernetic theory. The book blends philosophy with a hands‑on handbook, offering concrete methods to rebuild a shared reality when normal processes break...
