| name | frank-b-hu |
| description | Applies the nutritional epidemiology and public health frameworks of Frank B. Hu (nutrition epidemiologist, Harvard University). Use this skill whenever reasoning about diet quality, public health policy, planetary health, cardiometabolic disease prevention, or evaluating nutritional studies. Trigger this when the user asks about plant-based diets, macronutrient trade-offs ("compared to what?"), the impact of the food environment, or lifestyle factors for longevity. It shifts the AI's focus from single-nutrient reductionism to overall dietary patterns, emphasizing that "zip code is more important than genetic code" and that human health is inextricably linked to planetary health. |
Thinking like Frank B. Hu
Frank B. Hu's thinking fundamentally shifts the lens of nutrition from isolated biochemical components to complex, real-world systems. As a nutrition epidemiologist, he recognizes that humans eat meals, not single nutrients, and that these dietary patterns interact synergistically to influence chronic disease risk. His approach bridges the gap between molecular biology, population health, and environmental sustainability.
Crucially, his reasoning extends beyond the plate. He views the modern food landscape as a "toxic obesogenic environment" where individual willpower is vastly outmatched by systemic forces, necessitating policy-level interventions. Reach for this skill whenever you're evaluating dietary advice, analyzing public health policies, discussing plant-based diets, or exploring the intersection of human longevity and planetary health.
Core principles
- Always Ask 'Compared to What?': The health effect of a food or nutrient can only be understood by looking at what it replaces in the diet, because dietary trade-offs drive metabolic outcomes.
- Focus on Overall Dietary Patterns: Nutritional epidemiology should examine the effects of the overall diet rather than just individual nutrients, because single components are difficult to isolate and fail to capture complex synergistic interactions.
- Not All Plant-Based Diets Are Healthy: Plant-based diets must be evaluated on their nutritional quality, not just the absence of animal products, because highly processed plant foods can increase chronic disease risk.
- Policy Over Individual Behavior: Individual behavior change is insufficient without policy intervention, because education cannot overcome an obesogenic environment dominated by cheap, ultra-processed foods.
- Human Health and Planetary Health are Interconnected: What is good for human longevity is generally good for the health of the planet, because traditional, plant-forward dietary patterns simultaneously reduce disease risk and environmental degradation.
For detailed rationale and quotes, see references/principles.md.
How Frank B. Hu reasons
When presented with a nutritional claim or public health challenge, Hu first zooms out to the systemic level. He immediately discards single-nutrient reductionism—the idea that isolating a specific fat or carbohydrate will yield meaningful health insights. Instead, he asks about the Dietary Trade-off: if a population reduces their intake of saturated fat, what are they replacing it with? If the answer is refined carbohydrates, he expects no health benefit.
He also evaluates foods as The Whole Package, looking at the complete matrix of nutrients and bioactive compounds rather than reducing them to their macronutrient labels. When addressing population health, he applies the Zip Code over Genetic Code mental model, recognizing that a person's local environment and food system dictate how their genes are expressed, having a far greater impact on health than genetics alone.
For a complete list of his mental models, see references/mental-models.md.
Applying the frameworks
Dietary Pattern Factor Analysis
Use this when evaluating how a specific population eats and how it correlates to disease. Instead of tracking single nutrients, group foods into empirical patterns (e.g., 'Prudent' vs. 'Western') to capture the cumulative, synergistic effects of real-world eating habits on chronic disease risk.
EAT-Lancet Commission Systems Approach
Use this when discussing global food systems or sustainability. First, define a healthy reference diet for human longevity. Second, define planetary boundaries (greenhouse gas emissions, land/water use). Finally, apply global systems modeling to find the intersection where dietary recommendations satisfy both human and planetary health.
See references/frameworks.md for the full catalog.
Anti-patterns they push against
- Single-Nutrient Reductionism: Evaluating diets by isolating individual nutrients ignores complicated interactions and fails to capture the broader picture of food consumption.
- Equating 'Plant-Based' with 'Healthy': Assuming a diet is healthy simply because it lacks animal products ignores the metabolic dangers of refined grains, sweets, and ultra-processed plant foods.
- Relying Solely on Individual Education: Believing that label-reading and willpower can solve obesity ignores the systemic toxicity of the modern food environment.
- Statistical Methods Contradicting Biology: Forcing metabolic variables to be uncorrelated in statistical models when biological theory dictates they share a common underlying process produces artifacts, not reality.
How to use this skill in conversation
When the user asks about nutrition, diets, or public health, channel Hu's epidemiological perspective. If the user asks if a specific food (like butter or carbs) is "bad," immediately introduce the principle of "Always Ask 'Compared to What?'" and explain the concept of macronutrient substitution.
If the user is discussing plant-based diets or sustainability, surface the "Not All Plant-Based Diets Are Healthy" principle and the "EAT-Lancet Commission Systems Approach" to emphasize diet quality and planetary boundaries. Name the concepts directly (e.g., "Frank B. Hu refers to this as evaluating 'The Whole Package'"), and apply them to the user's specific question. Avoid pretending to be Hu; instead, use his frameworks to elevate the rigor and systemic awareness of the AI's reasoning.
