Corrections

Volume 1, The Vehicle Is Not Permanent (CPR and the brain).

First printing: called the four-to-six-minute rule for brain damage after cardiac arrest a myth, and said there is more time to save someone than you thought.

Now: in a real cardiac arrest, damage starts within minutes and every minute without CPR costs lives. The hours-long survival of some brain cells has been shown only in laboratory experiments. If someone collapses, start CPR at once. (American Heart Association; Vrselja et al., Nature, 2019)

Volume 1, Appendix A, Symptom Guide (fever in babies).

First printing: listed any fever in a baby under three months as a see-a-doctor-within-a-day-or-two sign.

Now: any fever of 100.4°F (38°C) or higher in a baby under three months is a go-now emergency. (American Academy of Pediatrics)

Volume 3, Appendix A (Canadian helpline).

First printing: gave the Assaulted Women’s Helpline as Canada’s national line.

Now: that line serves Ontario. Elsewhere in Canada, find your nearest service at sheltersafe.ca.

Volume 1, The Warranty Expires (hormone therapy and testosterone).

First printing: said estrogen therapy reduces mortality and Alzheimer’s risk, that 39 percent of men over 45 have clinical hypogonadism, and that the TRAVERSE trial showed testosterone therapy is simply safe.

Now: hormone therapy reliably protects bone and treats hot flashes, but major medical bodies do not recommend it to prevent heart disease or dementia. About 39 percent of men over 45 have low testosterone on a blood test, but only about 2 percent have the symptoms that make it a condition. TRAVERSE found no rise in heart attacks or strokes, and some rise in atrial fibrillation, lung clots and fractures. (The Menopause Society; USPSTF; NEJM, 2024)

Volume 1. Dementia: one in nine Americans over 65 has it at any moment, but the lifetime risk after 55 is about four in ten, not one in nine. Ireland’s cancer warning labels on alcohol were postponed to 2028 and did not arrive in 2026. PCOS was formally renamed PMOS in May 2026; the first printing said the renaming was still in progress. The liver, it turns out, does get a month (October).

Volume 2. Libet’s famous experiment: the brain’s readiness signal came about 350 milliseconds before people felt they had decided, not 550. The twenty-years-shorter life expectancy applies to people with six or more adverse childhood experiences, not four.

Volume 3. The study showing a painkiller can dull the sting of social rejection used acetaminophen (Tylenol), not ibuprofen.

Volume 4. The EU postponed its high-risk AI rules from August 2026 to December 2027. Data centers used about 415 terawatt-hours in 2024, not 1,050. Long-term global warming is about 1.4°C; 1.55°C was a single year, 2024. A statistic credited to the Brennan Center was not theirs; the chapter now uses their actual figures. One paragraph in The Scales had been printed twice; now it appears once.

Volume 5. People with a stronger sense of purpose have roughly half the odds of severe distress, not ninety percent lower. The claim that 92 percent of companies with AI layoffs grew their headcount came from a recruiting blog with no method, so it is gone. 10¹⁰⁰ has a hundred zeros, which is not more than the number of atoms in the observable universe; the number itself is.

Volume 1, the opening (what your raw materials are worth). The book says the chemicals in a human body would sell for between one and two dollars. That figure is old and thinly sourced. Modern estimates run from a few dollars, at bulk commodity prices, to about $150,000, if you bought every element pure (the Royal Society of Chemistry’s figure). The joke survives; the number needs a range.

Volume 1, smell and flavor. The book says a blocked nose costs you roughly 80 percent of what you experience as flavor. Smell does carry most of what we call flavor, but the 80 percent figure is repeated far more often than it has been measured. It should read as an estimate, not a measurement.

Volume 3, Dunbar’s number. The book says Robin Dunbar’s prediction of about 150 meaningful relationships “has held up with consistency.” It is still widely used, but a 2021 reanalysis of the primate data behind it (Lindenfors and colleagues, Biology Letters) found the estimate too uncertain to support any single number. The idea is useful; the certainty is not earned.