Substances Decoded

Browse any health forum and within minutes we will find someone absolutely certain that seed oils are destroying civilization, someone else equally certain they are perfectly safe, and a third person promoting nicotine patches as the nootropic discovery of the decade. Each has studies. Each has mechanisms. Each speaks with the conviction of someone who has found the truth that everyone else has missed. The problem is not that all of them are wrong. The problem is that none of them have systematically scored their confidence against the structural forces corrupting health information. Conviction is cheap. Calibrated confidence costs effort. Here is the effort.

Each substance below is scored using the Health Decode Framework: Corruption Filter (−6 to +6), Evidence Quality (−5 to +9), and Evolutionary Coherence (−4 to +4). Total range: −15 to +19.

Seed Oils

Canola, soybean, corn, sunflower, and other industrially extracted vegetable oils at modern consumption levels.

The claim: "Seed oils are harmful."

Corruption Filter (−2): The food industry has enormous financial interest in defending seed oils, which are cheap to produce and ubiquitous in processed food (−1). Research is mixed in funding and independence (0). Official dietary recommendations have been shifting, with recent acknowledgments that the original pro-seed-oil guidance may have been premature (−1). The corruption pressure leans toward defending seed oils, which means claims against them are swimming upstream against industrial incentives.

Evidence Quality (+3): Three independent inference paths point toward concern: oxidation chemistry (polyunsaturated fats are chemically unstable and produce harmful compounds when heated), epidemiological correlation (rising seed oil consumption tracks with rising inflammatory disease), and mechanistic biology (omega-6 to omega-3 ratio imbalances promote inflammatory pathways) (+2). The mechanism is plausible but not yet fully confirmed (+1). Replication is mixed—some findings replicate, others do not (0). Long-term controlled data is limited (0).

Evolutionary Coherence (−1): Industrial seed oil extraction has existed for approximately one century. No human population consumed these oils at current levels before the twentieth century (−1). There is no traditional population data showing long-term health outcomes because the exposure is too recent (0).

Total: 0 — Low confidence. The concern is plausible but not proven. The prudent response: reduce heated use of seed oils, prefer traditional cooking fats (olive oil, butter, coconut oil, animal fats) where possible, and do not panic. The online discourse around seed oils is far more certain than the evidence warrants.

Artificial Sweeteners

Aspartame, sucralose, saccharin, and similar non-caloric sweetening compounds.

The claim: "Artificial sweeteners are harmful."

Corruption Filter (−3): Both sides of this debate are corrupted. The artificial sweetener industry funds studies showing safety. The sugar industry benefits from doubts about alternatives. Both-sides corruption makes the information landscape particularly unreliable (−1). Research is heavily industry-funded on all sides (−1). Recommendations have shifted repeatedly over the decades—safe, then suspect, then safe again (−1).

Evidence Quality (+2): Two to three inference paths exist: gut microbiome disruption (some evidence that sweeteners alter microbial composition), insulin response (some evidence of paradoxical insulin signaling despite zero calories), and appetite dysregulation (possible disconnection between sweet taste and caloric intake) (+1). Mechanisms are plausible but contested (+1). Replication is mixed (0). Long-term data is confounded by the fact that people who use artificial sweeteners often have other dietary patterns that affect health outcomes (0).

Evolutionary Coherence (−2): These compounds never existed in nature before industrial chemistry created them (−2). No traditional population data exists (0).

Total: −3 — Very low confidence. Artificial sweeteners are probably not as harmful as the sugar they replace. They are probably not as benign as water. The honest assessment: a reduced-harm alternative for people trying to cut sugar, not a health food. If we are choosing between a sugary soda and a diet soda, the diet version is likely less bad. If we are choosing between a diet soda and water, there is no contest.

Caffeine

As consumed in coffee, tea, and supplemental form.

The claim: "Moderate caffeine consumption is beneficial."

Corruption Filter (+2): Funding is mixed but a large body of independent research exists (0 to +1). Recommendations have been remarkably stable over decades—moderate consumption is consistently described as safe and potentially beneficial (+1). No single industry dominates the research landscape in a way that would systematically distort findings.

Evidence Quality (+9): Four or more independent inference paths converge: cardiovascular epidemiology, cognitive performance research, metabolic studies, and neurodegenerative disease research all point toward benefit at moderate doses (+3). The mechanism is crystal clear—caffeine is an adenosine receptor antagonist (a compound that blocks the brain's sleep-promoting signals), and the downstream effects on alertness, metabolism, and neuroprotection are well-mapped (+2). Replication is highly consistent across populations, methodologies, and decades (+2). Long-term data spans over sixty years of systematic study (+2).

Evolutionary Coherence (+2): Coffee and tea have been consumed for centuries across diverse cultures (+1). Populations with high caffeine consumption (e.g., Scandinavian countries with the world's highest per-capita coffee intake) show no population-level harm and often show favorable health metrics (+1).

Total: +13 — High confidence. Moderate caffeine consumption (roughly 2–4 cups of coffee daily for most adults) is well-supported by convergent evidence. Individual variation exists—some people metabolize caffeine slowly due to genetic variants in CYP1A2 (the liver enzyme responsible for caffeine breakdown) and experience anxiety or sleep disruption at lower doses. Caffeine consumed too late in the day disrupts sleep architecture regardless of perceived tolerance. But for most adults, moderate consumption is one of the better-supported health interventions available.

Nicotine (Isolated)

Nicotine itself, separate from tobacco smoke. As delivered by patches, gum, lozenges, or other smoke-free methods.

The claim: "Isolated nicotine has cognitive benefits and relatively low harm."

Corruption Filter (−4): Both sides are heavily corrupted. The tobacco industry spent decades lying about nicotine's risks, poisoning the entire research landscape (−2). Anti-tobacco advocacy, while justified in its campaign against smoking, has sometimes conflated nicotine (the compound) with tobacco (the delivery mechanism), creating overcorrection in the opposite direction (−1). Recommendations have been highly unstable (−1).

Evidence Quality (+5): Three inference paths support cognitive benefit: direct cognitive enhancement studies (improved attention, working memory, and processing speed), Parkinson's disease epidemiology (smokers have markedly lower Parkinson's rates, suggesting nicotine may be neuroprotective), and appetite-regulation research (+2). The mechanism is very clear—nicotine binds to acetylcholine receptors (the neurotransmitter system most directly involved in attention and learning), and the downstream effects are well-characterized (+2). Some replication exists, particularly for the cognitive effects (+1). But long-term data on isolated nicotine (without tobacco) is limited because the isolation of nicotine from its traditional delivery method is relatively recent (0).

Evolutionary Coherence (+1): Tobacco was used in traditional cultures for thousands of years, primarily in ceremonial or occasional contexts (+1). Modern delivery methods (patches, gum, vaping) are novel, and modern consumption patterns (constant low-dose exposure via patches or frequent vaping) have no traditional precedent (0).

Total: +2 — Low confidence. Nicotine's terrible reputation is likely collateral damage from tobacco. The compound itself, separated from the combustion products that cause cancer and cardiovascular disease, appears to have genuine cognitive effects and lower inherent risk than its association with smoking suggests. Low-dose isolated nicotine (via patch or gum) is probably low-risk for most adults. Vaping adds unknown variables. The honest position: interesting, possibly useful, still uncertain. Not the miracle the nootropics community claims, not the poison the anti-tobacco movement implies.

Alcohol

Specifically the "J-curve" claim: that moderate alcohol consumption is healthier than none.

The claim: "Moderate alcohol consumption is beneficial for health."

Corruption Filter (−4): The alcohol industry has funded an enormous body of research examining the supposed benefits of moderate drinking, and industry-funded studies are significantly more likely to find benefits (−2). Industry funding bias is well-documented in the alcohol research literature (−1). The famous J-curve—the claim that moderate drinkers live longer than both heavy drinkers and abstainers—has been increasingly contested as methodological flaws in the original studies have come to light (−1).

Evidence Quality (−2): One or two inference paths exist, and they are weakening under scrutiny (0). No clear biological mechanism explains why a cellular toxin would improve health at low doses (−1). The J-curve does not survive reanalysis—when studies control for the "sick quitter" effect (the fact that many abstainers are former heavy drinkers who quit due to health problems), the apparent benefit of moderate drinking largely disappears (−1). Confounded data compounds the problem: moderate drinkers tend to be wealthier, better-educated, and more socially connected than abstainers, and those confounders drive health outcomes independent of alcohol (0).

Evolutionary Coherence (+1): Fermented beverages have existed for thousands of years across many cultures (+1). However, traditional consumption was typically lower-dose, less frequent, and socially embedded in ways that differ markedly from modern drinking patterns (0).

Total: −5 — Very low confidence. The claim that moderate alcohol is good for health is likely corrupted. Recent independent meta-analyses, particularly a major 2022 analysis published in The Lancet, suggest there is no safe level of alcohol for overall health. The J-curve appears to be an artifact of confounded study designs that the alcohol industry had little incentive to correct. If we do not drink, the evidence does not support starting. If we drink moderately, the absolute risk increase is small but real. The honest answer: alcohol is a mild toxin that many people enjoy, and the health benefits attributed to it are probably not real.

Summary

• Seed oils (0, Low): Plausible concern. Reduce heated use. Prefer traditional fats. Do not panic.
• Artificial sweeteners (−3, Very low): Less harmful than sugar. Not health food. Use as a stepping stone away from sugar, not as a destination.
• Caffeine (+13, High): Moderate consumption well-supported. Mind the timing (not late in the day) and individual variation.
• Nicotine isolated (+2, Low): Probably over-demonized. Possibly useful for cognition. Still uncertain. Separate the compound from the delivery method.
• Alcohol benefit (−5, Very low): J-curve likely an artifact of corrupted research. No safe level for health. Enjoy it if we choose to, but do not pretend it is medicine.

How to Apply This

When someone claims "X is good for you" or "Y is dangerous," resist the urge to accept or reject based on vibes, authority, or the conviction of the person making the claim. Instead, run the framework. Ask who profits. Ask how many independent paths support the conclusion. Ask whether there is a clear mechanism. Ask what traditional populations show. Calibrate confidence from the scores, not from the volume of the claims. The framework will not always give a definitive answer. But it will always tell us how much certainty is warranted—and that alone is more than most people have.

How This Was Decoded

Applied the Health Decode Framework to five commonly debated substances. Each scored independently for corruption risk, evidence quality, and evolutionary coherence. Cross-referenced against peer-reviewed literature, population studies, and documented industry funding patterns. Validated framework calibration against known high-confidence claims (sleep, exercise) to ensure scoring consistency. Applied DECODER principles of incentive divergence, convergent confidence, and corruption structure analysis to identify where information asymmetries distort public understanding of each substance.