Risk + Safety + Security + Sciences

This episode unpacks a 2024 paper arguing that intelligence assessments rely on vague confidence labels and intuition rather than rigorous methods. It explains the risk science framework (A', C', Q, K), the Strength of Knowledge (SoK) five-dimensional evaluation, and assumption-deviation risk, using a hands-on case study to show how clearer metrics can change critical decisions. Listeners learn practical takeaways: probe assumptions, judge data relevance, and beware of information overload and layered errors—especially as AI amplifies them.

This episode explores the "authority gradient" — the power imbalance that silences subordinates — through deadly case studies from the rail industry. It shows how everyday deference in meetings or locomotive cabs can override safety procedures, leading to near-misses and fatal derailments.
We examine why investigators often fail to name this root cause, how economic pressures and hierarchy stop workers from using formal safety procedures, and what organisations must do to measure and address authority gradients to prevent future tragedies.

This episode explores David R. Mandel's research showing how intense accountability pressure shapes intelligence analysis: analysts hedge with caveats, use coded ratings, and agencies favour process-based checks over outcome accuracy.
It examines broken tools like verbal probabilities and ACH, the post-9/11 shift to ICD-203, and Sherman Kent’s plea to "elevate the quality of discussion" by prioritising rigorous, honest reasoning over pleasing policymakers.

This episode decodes the IIA's 2026 update to the Three Lines Model, exploring its shift from a defence mindset to value creation, the controversial 12-month cooling-off rule, and the risks of expanding internal audit's power as the "integrator of assurance." We pair the official playbook with an academic critique that highlights accountability gaps, visual-literacy traps in boardroom diagrams, and practical problems for small organisations.

This episode unpacks the Open Risk Culture Standard (ORCS) v8 and shows how a behavioural approach to risk — emphasising leadership, adaptive elasticity, and ethics — enabled companies like Apple and Netflix to survive and thrive in digitised, volatile environments.
Through case studies (Apple, Netflix, Red Lobster, Patagonia) and concepts like prescience, sensemaking, mindful inaction, and dynamic risk equilibrium, the hosts explain the five-level risk culture maturity model and practical steps to transform risk management from reactive to optimising.

This episode dissects classified corporate travel risk briefs from July 2026 to reveal how government travel advisories can lag or contradict real-world threats. From missile strikes in the Gulf and catastrophic typhoons to a localised Ebola outbreak and global heatwaves, the analysts show how simultaneous crises shatter aviation resilience and force corporations to rethink travel as strategic risk management.
Experts explain forecasting failures, the dangers of false precision in probabilities, and the operational measures companies must take to navigate a volatile, interconnected world where the comforting green light on a travel advisory isn't always reliable.

This episode digs into the FAAs 2022 Risk Management Handbook and NTSB data showing roughly 85% of aviation accidents stem from pilot error.We explain key concepts—hazard vs. risk, the five hazardous attitudes, the P.A.V.E. and MSAFE checklists, the risk assessment matrix and personal minimums—and show how pilots use these tools to avoid disaster.A real case study and the Swiss cheese model illustrate threat and error management, expectation bias and automation complacency, plus the 3P model (Perceive, Process, Perform).Finally, we draw practical parallels to everyday life and work: set strict personal minimums, anticipate automation limits, and apply aviation decision frameworks to reduce risk in any high-stakes environment.

This episode walks through my framework for statistical thinking, showing how data is defined, measured, summarised, and used to make inferences — and how each step can mislead if handled poorly.
Listeners will learn about types of data (nominal, ordinal, interval, ratio), measures of centre and spread, distribution shapes (including fat-tailed risks), sampling and non-sampling errors, hypothesis testing and p-values, confounding variables, and the ethical importance of data privacy.
Overall, the episode reframes statistics as a practical toolkit for critical thinking: protect yourself from misleading averages, question causal claims, and always consider uncertainty and ethics when interpreting research.

This episode examines a 2026 paper arguing that precise probability estimates often hide unexamined assumptions in the transformation from knowledge to number. It introduces Strength of Knowledge (SoK), Justification of Transformation (JoT), and Strength of Justification (SoJ) and shows how forcing point estimates or unjustified priors can create false precision.
Using boardroom and event-tree examples, the episode warns that strong data alone doesn’t guarantee trustworthy probabilities and urges analysts and decision-makers to favour honest ranges, interrogate the math behind reported numbers, and treat weakest-link weaknesses as red flags.

This episode deconstructs the corporate buzzword "resilience," explaining three conflicting scientific traditions—engineering (bounce back), ecological (persist in new states), and socio-ecological (adapt and transform)—and shows how vague use of the term creates legal, operational and moral risks for boards and critical infrastructure operators.
Drawing on my July 2026 position paper, the episode stresses the need for cellular-to-system assurance, a software bill of materials, and a six-part specification (system, disturbance, beneficiary/cost bearer, time frame/end state, tradition, frame choice) so organisations can define, maintain and legally justify what they mean by resilience.