This week I handed my raw genotype file to OpenAI’s new Rosalind Workbench and asked it one question about migraines. Twenty minutes later I had a bioRxiv-ready paper.

The latest LLMs are good enough to referee a Nature paper. I don’t find that terribly surprising but that’s the conclusion of a large expert study from CMU and KAIST. About 50 domain scientists rated almost 3000 individual criticisms — every point raised by human and AI reviewers of 82 Nature-family papers — for correctness, significance, and evidence. An OpenAI agent beat each paper’s top-rated human reviewer, 60.0% to 48.2%. Claude Opus 4.5 and Gemini 3.0 Pro were statistically indistinguishable from the best human. All three beat the worst human on every dimension.

The (somewhat minor) catch: LLMs raise more incorrect points than the top human (86% correct vs. 92%), they overlap heavily with each other (21% of AI–AI criticisms match, vs. 3% for human pairs, so three bots are not three opinions), and they’re weak on “subfield conventions” — the tacit rules a working scientist absorbs slowly on the job. (See PSWeek260508)

For professionals, this is a workflow question. For personal scientists it’s something bigger: the difference between having a reviewer and not having one. Nobody peer-reviews my magnesium experiment (PSWeek260514). Until LLMs, if I wanted a second opinion on my own data, the options were a social media query or a Quantified Self meetup. When I tested a doctors-only medical LLM against plain Claude (PSWeek260423), I found the general-purpose model just as good. Now you can get even better, general-purpose models packaged to fit a scientist’s workflow.

Rosalind

ChatGPT Rosalind Workbench is OpenAI’s new life-sciences workspace, in research preview inside the regular ChatGPT app. It runs on GPT-Rosalind, a model that combines frontier reasoning with “specialized tool orchestration across medicinal chemistry, genomics, wet-lab assistance, and other scientific applications.” In practice that means it already has the sequence viewers, alignment tools, and plotting you’d ordinarily have to install separately.

I tried it with the easiest prompt ever. I uploaded a 700,000-line text file — the raw output of a cheap consumer genotyping test like 23andme. (PSWeek250403) — and asked it to find markers that might reveal my risk of migraines. I had already ruled out EMF as a possible cause for my teenage daughter’s headaches (see PSWeek230706 ) so I wanted to know if I carry something that might run in the family.

After a few minutes of chugging, it gave me a table of four specific variants associated with migraine, and this summary:

I matched 85 of the 123 independent migraine loci from a large study of 102,084 cases and 771,257 controls. Thirty-five were exact lead variants; 50 were strong markers from the same loci. The weighted balance was roughly the 80th–85th percentile among the study’s European-ancestry reference population. Treat that as a research estimate—not an absolute lifetime-risk percentage or a clinical polygenic score.

Then I asked it to write the whole thing up in a format suitable for bioRxiv, the preprint server where professionals post work before peer review. Soon I had a professional-looking paper: detailed methods, a discussion section, useful charts, and citations that — I checked — were real. Read it yourself (PDF).

What the 80th percentile means

Okay fine, but how “scientific” are my results?

When I asked Claude to review it pointed out that the “large study” on migraine risk that Rosalind quoted is Hautakangas et al. 2022 in Nature Genetics, and it’s a good one: 123 loci, 86 of them new. But the same paper estimates that all the common variants together explain maybe 11% of migraine heritability. Rosalind’s score is built from the part of that 11% my genotyping chip happens to cover — 85 of 123 loci, and 50 of those 85 by proxy markers rather than the actual lead variants. Whether a proxy in the same neighborhood carries the same weight is somebody’s judgment call, and who knows how it was calculated. I wonder how much of my 80th percentile lives in those 50.

To its credit, Rosalind said so. “Research estimate, not a clinical polygenic score” is well put, and frankly it’s more honest than a lot of consumer genetics reports.

What about the paper it generated? Real citations, sensible methods, a discussion section that would please a graduate advisor — produced in twenty minutes from a question I typed on a Sunday afternoon. Back when I tested the first “AI scientist” tools (PSWeek250612), the citations were the weak link; now they check out. What hasn’t changed is that just because something fits the pattern of an academic paper doesn’t make it true. I’m my case, ChatGPT wrote me a well-formatted hypothesis about a sample of one, and the only real way to confirm it is the same as it’s always been: go to the real world and find real data. In other words, back to personal science.

Personal Science Weekly Readings

Speaking of peer review, it’s not like humans have been doing all that great of a job either.

A 2026 blinded study in a cardiology journal had an LLM review 40 real submissions alongside the human referees. The AI matched the final editorial decision 68% of the time; the humans, 67.5%. Neither number is impressive, which is kind of the real point here. (tldr; the bar was never that high.)

Ask Dan Ariely, one of the most famous behavioral scientists in the world. You may even have read his book, Predictably Irrational, that impressed all of us with some practical consequences of his idea that humans make bad decisions in predictable ways.

Well, the book that made him famous is based partly on his 2002 peer reviewed paper Ariely & Wertenbroch (2002) on deadlines and procrastination. It’s been cited thousands of times. But when Data Colada examined it more closely they found those spreadsheets were a little too spot-on to prove his thesis. Those spreadsheets — “last saved” by user “Dan Ariely”, had suspiciously convenient results. It’s Ariely’s second such finding; a 2012 honesty study was retracted in 2021. His response this month: after “more than two decades, and hundreds of experiments,” his memory is “insufficient.”

The charts below show (left) what you’d expect if the data had been left alone and (right) how it was presented in the paper. Somehow the grades were adjusted in a way that statistically gave a final result with the conclusion Ariely wanted. Peer reviewers, looking just at the means, didn’t see any shenanigans. But this type of manipulation is easy for an LLM to spot.

And in the “science-based policy that turns out to be wrong” department, Works in Progress makes the painstaking case that your careful recycling is probably a net harm: something like a quarter of what’s in your bin is contaminated and has to be washed — expensively — before it can be processed. Much (most?) of the time, landfill is cheaper and cleaner than the alternatives. (But if you read PSWeek250102, where somebody put an Airtag into a Starbucks recycling bin, you already knew that).

Finally, what all of this week’s readings have in common: the real checking was done by outsiders, not by the credentialed process that was supposed to catch it. That checking just got very cheap. Rosalind will write you a paper in twenty minutes; or it can tear apart a professional’s life work in ten. The gap between “professional” and “amateur” has never been narrower.

So what about my paper?

It’s easy to submit a paper to BioRXiv (I’ve done it before) but with this one I haven’t tried, at least not yet. Part of my reason is that, because I’m not a professional scientist, I feel no urgency to “publish”. Is this just me suffering “imposter syndrome“?

I believe that in a few years people will prefer AI-assisted writing over human writing, at least for non-fiction or scientific writing, where clarity and accuracy are critical. So what’s left for humans?

The answer is reputation. I feel responsibility for anything that goes out under my name and to be honest, I don’t feel like I understand “my” paper well enough to defend its conclusions. Rosalind, to me, was a fun exercise and I’m glad it’s now available to people like me, but the real consequence is not how much stuff you can publish, but whether other humans will care.

About Personal Science

Personal scientists test the claims that the rest of the world repeats — including the ones a very capable AI just wrote up for us. Nullius in verba — take no one’s word for it — was the motto of a society whose members could check each other’s work. The checking is cheaper now. That’s not an excuse to skip it.

We publish every Thursday. If you’ve tried Rosalind or one of the other AI research workbenches on your own data, let us know.

Leave a comment

This week, I look at the “how much is me” question from a heavily-discussed 1989 book about the settling of America.

I inherited my last name from William Sprague, who in 1629 was a passenger on the Lion’s Whelp, that sailed into Massachusetts Bay at the very forefront of the Puritan migration—an eleven-year window that brought twenty-one thousand people to America, after which the traffic essentially stopped. After ten generations of mixing, William contributes less than a thousandth (2¹⁰) of my genome. But DNA travels in chunks, so after all that time it’s roughly a coin flip whether he contributed anything at all—except for the one piece that doesn’t get lost in the shuffle, the Y chromosome, which came with the surname. It’s intriguing to wonder what else he did give me. I speak English because he did, for example, and I eat wheat bread instead of rice.

Historian David Hackett Fischer gave a detailed account of that legacy in his 1989 book Albion’s Seed. Instead of personality traits, which may or may not have a genetic basis and are hard to pin down in any case, he looks at broad social habits, which he calls folkways (a term he borrowed from sociologist William Graham Sumner): how a community marries, raises its children, spends its time, and answers an insult.

His claim is that America was settled in waves, where people of similar cultural backgrounds clumped into specific geographic regions: Puritans in New England, royalist gentry (”Cavaliers”) and their indentured servants in Virginia, Quakers in Philadelphia and the Delaware Valley, “Backcountry” Scots-Irish in Appalachia. Despite the passage of time and the whirlwind of mixing since then, today’s institutions and the cultural artifacts they support are recognizable descendants of those original settlers.

Scott Alexander wrote a lengthy review of the book that concludes with what he admits is speculation: that today’s politics are the four folkways merged into two—Puritan plus Quaker into a Blue Tribe of education, moral reform, and conscience; Cavalier plus Borderer into a Red Tribe of honor, hierarchy, and independence. If true, it would explain why many political arguments seem unwinnable: you’re not disagreeing about a policy, you’re repeating a centuries-old disagreement about the purpose of a good life.

As a personal scientist, I find the overall idea interesting but as much as possible I’d prefer to see claims that can be quantified. That’s why the best part Fischer’s research is the charts in the back, which squish high-level claims into specifiable variables that you can compare against your own family and its values. For example, instead of “what do you believe about family” he asks how many people did the average household actually contain.

Here are a few of my takeaways (and I emphasize, these are my personal interpretations, not necessarily consistent with the book):

1. Pregnant brides

Your own personal experiences with family and marriage will affect you in countless ways that are difficult to measure. So Fischer instead asks quantifiable questions:

What’s interesting is the large spread in the same century, on the same continent, among people who all thought of themselves as English Christians.

If you think of these quantifiable facts as a rough proxy for attitudes about sex and marriage, you can almost see current differences for how important it is to be married, what’s an appropriate age, an expectations about children. Indeed, my family and I probably look most similar to the Massachusetts answers above (we marry and have kids slightly on the later side).

(Side note about that backcountry number: the wedding itself is listed in Fischer’s table under the heading “abduction ritual.”)

2. How to discipline kids

How strict should parents be toward misbehaving kids? Fischer dug through historical accounts to find how each folkway disciplines a child’s “will”.

We all know parents who seem either especially lax or strict with their kids, but despite even extreme differences, most parents I know think of themselves as “normal”. In other words, compared the way they themselves were raised—their folkways—they don’t feel like outliers.

Here again, I probably relate more to the Massachusetts idea: we expect our kids to be well-behaved, regardless of “the role they were born into”.

3. Honor

Fischer also looked at how people respond to an insult. Here I am definitely on the Massachusetts side, as unfortunately my elementary school tormenters discovered! (All you bullies in “The Pack”, you know I haven’t forgotten you!).

Primal honor means your standing is a physical fact about you that other people can damage, and that you are obliged to defend. Fischer traces it to a millennium of clan warfare on the Anglo-Scottish border, where there was no functioning state and a reputation for retaliation was the only property insurance available. Roughly 250,000 people carried that intuition to Appalachia between 1717 and 1775 — by far the largest of the four migrations.

Interestingly, Richard Nisbett’s group at Michigan — the same lab we discussed in PSWeek260827’s perceptual tasks — built an experiment for it.

Cohen, Nisbett, Bowdle and Schwarz (1996) recruited white male undergraduates at the University of Michigan, sorted by where they grew up, and told them the study concerned “limited response time conditions,” but in reality was a simulation of how they respond to insults. In a series of experiments, the researchers showed that southerners had significantly higher stress hormones than northerners exposed to the same obnoxious behavior.

Interesting result, but as far as I can tell from Ulrich Schimmack’s audit, the hallway experiment has never been independently replicated—which, from our previous posts, shouldn’t surprise you.

Anyway, it’ll be hard for me to do a cortisol test the next time somebody cuts me off in traffic, but having grown up in the north, my guess is that I’ll probably fit the stereotype.

4. What time is for

Finally, what about my interest in personal science? Back in PSWeek260528 we discussed my ActivityWatch logs—a minute-by-minute breakdown of how I spend my computer time. But did I think of that myself, or am I somehow echoing my long-ago ancestors’ own ideas about time?

Fischer looked at the vocabulary of each of the founding communities to discern their attitudes toward work. Here are the words each group used to describe time:

Is my interest in personal science a legacy of the classic Puritan attitude of my ancestors, four hundred years and ten generations after the Lion’s Whelp? My friends who “kill” or “pass” the time on TikTok or backpacking or whatever—how much of their attitude is inherited from their own ancestors’ sense of a life well-lived?

But I’m pretty sure William would be proud of me.

Personal Science Weekly Readings

Speaking of stressing out over a perceived slight, somebody filed a Class-action lawsuit against Oura over its claim of 95% sleep-staging accuracy.

Much as I agree with the complaint—see PSWeek251218 for how inaccurate I find consumer sleep products in general—I don’t want to live in a society where we need lawyers to tell us whether a product is worth testing. I’ll make up my own mind, please.

I don’t understand why data centers have been getting so much publicity. If they follow local laws, why treat them any differently than any other business? In fact, they’ve been around for decades, many of them operating quietly in your neighborhood right now, says @LoganDobson, who points out there are dozens in Manhattan.

If you want to find the ones near you, use Data Center Map. Here are the ones on the West Coast. (Zoom into the map to see the exact address for each)

About Personal Science

You don’t always need numbers in order to think scientifically. Just follow the motto of the Royal Society Nullius in verba—take no one’s word for it, including ours. If you have other topics you’d like to discuss, please let us know.

Leave a comment

This week we look at some interesting cross-cultural studies that claim to show how personality and instinct differ across families and ethnicities.

Long-time readers know I spent a large part of my life in Asia (see PSWeek251016), the consequences of a lifelong interest in languages and their effect on thinking. One idea that always intrigued me is how much my own preconceptions about the world are caused by my own cultural upbringing. If I had been born elsewhere, how much of me would still be me?

Psychologists and social scientists have studied this too, and now that LLMs are making it easier to research (and potentially replicate) the results of those studies, I’ve been revisiting some of these questions for myself. Here are a few examples.

1. The aquarium

Masuda and Nisbett (2001) showed Japanese and American students 20-second animated underwater vignettes and asked them what they’d seen. A big bright fish moves across a field of rocks, plants, bubbles, small fish.

The experiment simply looks at the first sentence that pops in your head. Americans overwhelmingly say something about fish. Japanese participants open with the scene as a whole — water color, the floor of the scene, inert objects — and made 65% more observations about the background overall.

Now repeat, but this time when shown the fish again, either against its original background or a novel one, the Japanese recognized it more accurately in its original setting; for Americans the background barely mattered. They hadn’t encoded a fish. They’d encoded a fish-in-that-place. (see more at the Masuda Lab).

2. Cow, chicken, grass

Which two go together?

Westerners tend to pair cow-and-chicken: both animals, a taxonomic rule. East Asians more often pair cow-and-grass: the cow eats the grass, a relationship. Neither is wrong. The point is that one of them felt obvious to you and you didn’t experience it as a choice.

3. Where do you put lunch?

Mandarin locates time on a vertical axis where English uses a horizontal one: 上个月 (shàng gè yuè, “upper month”) is last month, 下个月 (xià gè yuè, “lower month”) is next month — and of course 上午/下午 for morning and afternoon. Does that reach past vocabulary into how time is actually laid out in the head? Or is it dead metaphor, the way nobody hears a “deadline” as a line?

Fuhrman et al. (2011) tested this: Participants saw the middle frame of a temporal sequence — a half-peeled banana — then a second frame, and pressed a key to say whether it came earlier or later. In one block the two keys sat side by side; in another, one above the other.

Mandarin speakers were reliably faster on the vertical layout. Nobody was asked to arrange anything, and no spatial word appeared in the instructions.

To faithfully replicate this I suppose you should use a millisecond timer, but there's a crude home substitute: hand someone three photos — green banana, half-peeled, empty peel — and say "put these in order." Nothing else. Watch whether the sequence goes across or down. Try it on native English and then native Mandarin speakers.

But, but…

The heyday of these sorts of studies came between 2001 and 2011— the pre-registration dark ages, before the 2011 Bem precognition paper and the 2015 Open Science Collaboration replication project made everyone more skeptical of the “pop science” literature. So the obvious question: do these findings hold up?

tldr; not really.

Over the years, I’ve tried these tests on my Asian and American friends with, frankly, mixed results. Back when I was less sensitive to the Replication Crisis in the social sciences (see PSWeek251203 ) I tended not to worry too much about the failures. The overall idea — that some cultures are more group-oriented and some are more individualistic — is such a tidy stereotype that who wants to ruin the cocktail party with rigorous testing. You can explain away the anomalies with comments like “Oh, you probably lived in America too long” or “Hmm, maybe that Japanese girlfriend affected you more than you think”.

The professionals vs personal science

But I’m just a humble personal scientist. What do the professionals say?

The eye-tracking evidence is a mess. Although Nisbett’s lab continued studies throughout the 2000s, other labs had a hard time replicating. (e.g. Rayner 2008). More sophisticated studies, using eye-tracking techniques developed for computer UX research, seem to show maybe some affinity for faces that look like yours. Okay, maybe true but hardly proof of a ‘foreground/background’ perception difference.

The horizontal/vertical time differences in the banana figure above has been replicated, sort of. A group of researchers from one Chinese lab continue to publish confirmations of their own findings, though I couldn’t find other independent examples. To me this smells like null publication bias—labs that tried to do similar experiments gave up and didn’t bother to report their failure.

In my experience, to really understand a culture you have to go very deep into the nuances of how the people think. It would be nice to have a simple test that could help quantify how deeply you’ve attained that level of insight. Unfortunately, I think people and cultures are simply too complicated for that.

Personal Science Weekly Readings

Speaking of replication, University of Arizona psychologist Matthias R. Mehl has over the years re-run experiments involving everyday speaking patterns, work which led to the debunking of that popular myth that women talk more than men. While digging through his multi-year dataset, he noticed something different: that people are talking less than they used to. In fact, since 2005, our daily count has been dropping by about 300 words per year.

I’m a sucker for examples of professional scientists’ mistakes and Mark Manson’s Intellectuals are F*cking Idiots is a recent good summary: a far-left university professor who was so enthusiastic about communism that he visited Pol Pot’s regime in Cambodia, where he was killed for being insufficiently supportive of the leader. Studies about behavior scientists, whose whole job is to predict these things, were unable to do better than a coin flip. And of course the lesson of Robert McNamara, praised in the 1960s as the smartest man alive, whose complex mathematical models convinced him to continue doubling down on the Vietnam War.

ChatGPT as second opinion. UCSF’s Robert Wachter ran a hard case past ChatGPT — his own labs and clinician notes — and reports that it found the relevant data and weighed it with enough nuance to compare favorably to two experts. But I liked his conclusion, borrowed from Joe Biden: don’t compare it to the Almighty, compare it to the alternative. For most people most of the time, the alternative isn’t a doctor. It’s nothing.

About Personal Science

Personal scientists take nobody’s word for it — nullius in verba. The whole point of the cultural-cognition literature is that some of what you “see” was installed before you had any say in the matter. You can’t step outside your own defaults to check them.

We publish every Thursday for people who use science for personal rather than professional reasons. If you’ve tried any of these tests on your own friends and gotten a result — including a boring one — let us know.

Leave a comment

This week we compile some links and other readings that we think are useful to personal scientists.

New Apps and Devices

Bob Troia (aka Quantified Bob) has a new app, Awesome Breathing, to help guide and visualize your breathing: box breathing, Dirga Pranayama, 4-7-8 and other standard exercises that can help alleviate stress and get you to sleep faster. It’s free, funded by supporter contributions who find it helpful.

Lumeria Lumoscope is a $300 device (pre-orders) that connects to your smartphone and shines RGB, UV, and polarized light on your skin “to see what the naked eye can’t”. You take images of your skin and track other variables like sleep, hygiene, products you’re using, etc. and it recommend how to improve your skin. The company behind it is a Y Combinator startup.

Peer review by the rest of us

We mentioned refine.ink back in PSWeek260122 as a very detailed AI-powered academic paper reviewer and I’ve been meaning to try it. X user Ben Golub shows how he ripped apart a peer-reviewed article on Stanford’s Encyclopedia of Philosophy, finding numerous serious errors including citations that point to the opposite conclusion and basic problems with reasoning. AI will make it nearly impossible future academics to get away with sloppy papers.

An open-source project Feynman is a collection of tools to do something similar: Read papers, research and get cited meta-analysis for your question. It also comes with Claude Skills so you can use it from within your existing Claude workflow.

Following Up

Back in PSWeek230921 we discussed cancer diagnostics and how many of the screening tests are so inaccurate as to be counter-productive. Tablet Magazine argues that mammography screening is overdone and uses this great graphic from a 2014 JAMA paper that illustrates the benefits and harms of mammography screening:

We’ve written about caffeine over the years (See PSWeek241010 and PSWeek230622) but not about ways to reduce its effects. Astral Codex Ten on The Quest For Caffeine You Can Have At Night has a great breakdown of several over-the-counter supplements that claim to help you sleep even if you’d had too much coffee. The pills have three main approaches: (1) supercharge the enzyme that normally metabolize caffeine, at potential cost of liver function, (2) skip coffee and take the stimulating chemical directly, and (3) use a caffeine-like molecule that gives you a shorter period of alertness. They all come with tradeoffs, but various personal scientists have tried each and this is a nice summary of the results.

About Personal Science

Personal scientists are skeptical about everything, so we aren’t bothered when experts can’t seem to agree. But that’s not an excuse for giving up. We have to believe something. The question is how to decide.

Economist Arnold Kling says “we decide what to believe by deciding who to believe”. Maybe you trust “our democratically elected government” and the (presumably neutral and science-minded) people they appoint at FDA, CDC, and other regulatory agencies. But you also trust other reputable organizations and their leaders. What do you do when your trusted sources disagree with one another?

Personal science says there’s only one way out of this conundrum: learn to apply the techniques of science to your everyday situations. Philosopher of science Michael Strevens summarizes this as a trust in (1) raw data, (2) good collection methods, and (3) logic to tie it together. (See PSWeek221229)

This newsletter is a weekly summary of ideas and techniques that we hope will help others who want to get better at personal science. If you have additional thoughts or topics you’d like to discuss, please let us know.

Leave a comment

I’m reminded of this upon seeing Dr. Anthony Fauci called before Congress, only to invoke the Fifth Amendment amid legal questions about a preemptive presidential pardon. A week later the committee voted 8–5, on party lines, to hold him in contempt.

The odd part, and the thing that raises my personal science eyebrows, is how little difference there is between your opinions about Fauci and how you voted in the last election. Fine—he’s a political figure now, in front of a political commission. But I can also predict more generally what you think about six-foot distancing, masks, school closures, vaccine effectiveness, and many other empirical questions. These are scientific questions, unrelated to politics. Shouldn’t we want to know the answers?

The cure vs. the disease

Much of the news focuses on the origins of the pandemic, but regardless of how that turns out, Dr. Fauci was the public face of nearly all the COVID era responses. Whether or not he directly ordered counter-measures like school closures, mask mandates, or restaurant restrictions, he publicly supported and encouraged them.

Many of the consequences will linger for generations. Face masks that were never required anywhere before the pandemic will be mandatory whenever “there’s something going around”. New digital infrastructure makes remote work and food delivery far more ubiquitous than it would have been otherwise. A precedent has been set for lockdowns and other restrictions that were unthinkable before 2020.

Even if our political system can’t answer society-level questions about the effectiveness of these unprecedented measures, every personal scientist should be curious to know what worked and what didn’t—and what we would do differently next time.

What I actually observed

If you’re a professional scientist in need of research grants or publishable results in a politically-charged environment, I don’t blame you for keeping your experimental proposals to yourself. But personal scientists are free from those restrictions. I don’t have expensive lab equipment or a big research staff, but I do have my own observations and powerful LLMs to help me sort facts from opinion. Here is my own personal experience:

1. I know several people hospitalized with COVID, some seriously. Most or all were unvaccinated — either infected before vaccines existed, or they declined.

2. I don’t personally know anyone who died unexpectedly. Not one. Among the hundreds of people I’ve asked, I’ve never gotten a single example of a genuinely out-of-the-blue death. The death of a grandmother is a very sad, of course, but not really unexpected.

3. All of my 80+ year old family members who were infected recovered just fine, even those who caught it in 2020 before vaccines. Of course I know many other families who weren’t so lucky, but I don’t think my situation is terribly unusual.

4. I personally know dozens of people with long COVID, some severely debilitated. COVID was certainly a very serious issue for them.

Now fine, these are personal anecdotes that are true for me and my direct acquaintances. Can we be a little more rigorous?

Do the math

Many of the numbers that were unknown or disputed at the time are now known with much greater certainty: the total number of deaths, the infection fatality rate (IFR) and its ranges broken down by geographical area and other demographics like age, sex, and race. The median US COVID death age was 78, and three-quarters were over 65.

How typical was my personal experience?

The average American knows about 600 people by name. US COVID deaths ran about 1.1 million, or 0.33% of the population. So a 600-person network should contain roughly two COVID deaths total, across four years. And yeah, that kind of tracks for me—we have a family friend who succumbed early in the pandemic, and maybe one or two others if I think hard enough.

But what about unexpected deaths — healthy, under 55? That’s maybe 2% of COVID deaths, or ~22,000 nationally: 0.007% of the population. The simple math:

22,000 / 330,000,000 = 0.007% then 600 × 0.00007 ≈ 0.04 unexpected COVID deaths among people I know well.

You’d need to know about 15,000 people to expect one unexpected COVID death. So for me, the fact I know nobody is not at all unusual. In fact, it’s true for 96% of Americans, including you.

Another way to think about this: a 600-person network sees between three to eight deaths a year from all causes in a normal year. COVID added two, nearly always an older person; call it a 10% increase in funerals. Too small to be detectable.

These are real deaths, and every one is a tragedy of course. But context is everything.

More context and a control

Now let’s think about another group of people I “know”: There are 100 governors and lieutenant governors, 535 members of Congress, and some number of actors, athletes, and executives whose names everyone knows. We hear about famous deaths all the time. If this was a stop-the-world event, where were they?

First, look at the ordinary background rate of unexpected death in healthy people under 55 — cancer, cardiac, accident, overdose. I count four unexpected non-COVID deaths among the “famous” : Basketball star Kobe Bryant, an actress from Glee, an actor from Black Panther, and a rapper.

For COVID deaths, I first looked in Wikipedia, where to have a page you must be a “notable” person—a pool of 150,000 Americans. I found just 3 COVID deaths that met the criteria (”under 55, unexpected”): Nick Cordero, Luke Letlow, and Adam Schlesinger. But I’ve never heard of them. Have you?

In other words, by the “famous people” standard, we get COVID death numbers that track my own personal experience: an uptick you’d only notice if you track very carefully.

Woodstock

For more context, I looked at another serious pandemic. The 1968–69 Hong Kong flu killed ~100,000 Americans out of 200 million, against COVID’s 1.1 million out of 330 million. Age-standardize — the 1968 US had a median age of 28 and 9.6% over 65, versus 38 and 16.5% in 2020 — and the overall menace looks similar.

I looked up “famous deaths” for that period and, as with COVID, I found just a few unexpected deaths from the pandemic (a CIA director, a broadway star), versus many non flu-related deaths (a Soviet Cosmonaut, a boxer in his prime, a Formula One racer).

Except. There were no nationwide restrictions, no daily infection dashboards, no masking, no vaccine ID cards required to enter a restaurant.

In fact, the famous Woodstock concert was held in August 1969, in the trough between the waves. Nobody remembers 1968 as an event, because nobody was counting.

So I wonder, based on my own personal experience, what would have happened if we hadn’t been watching Dr. Fauci, but instead just relied on our own personal observations that “something is going around”.

But the vaccines

Of course, without the public attention to the real dangers of SARS-COV-2, it’s hard to imagine the political will to do anything like Operation Warp Speed, that gave us a vaccine in record time. It’s impossible to tell how many lives were saved, including perhaps many famous ones. It’s a counter-factual that we’ll never answer to everyone’s satisfaction.

Here I was surprised at the range of estimates: various groups using different methodologies claim anywhere from a few hundred thousand on the low end, to 3 million+ on the high end. But all of them agree that 90%+ of the lives saved are in people over 60.

Another undisputed fact: before vaccines the average age at death was in the high 70s and 80% were over age 65.

So what would I have done differently? Honestly, I think it’s hard to defend Dr. Fauci’s draconian recommendations.

Personal Science Weekly Readings

I realize that my skeptical conclusions are not particularly welcome among people who would like to believe all those masking/hygiene efforts were worthwhile. To see many more details for how I derived the above figures, see my lengthy Claude transcript. And please leave a comment to tell me what I got wrong!

Speaking of errors in science, Paul Litvak did his own informal test of AI Peer review in order to see how well LLMs can catch errors in scientific papers. tldr; current AI models are pretty good, especially when you double-check with multiple LLMs. Although his testing isn’t super-rigorous, it’s pretty good for a first pass. He had Claude help him take the original papers and subtly modify them with substantive errors that undermine the paper’s conclusions. You can read his entire methodology

Finally, I can’t wait to read Gary Wolf’s new book; preorder a copy here: quantifiedself.com/book.

About personal science

Just because government agencies or other professionals can’t agree doesn’t mean the rest of us can’t make up our own minds. Personal science is the idea that anybody can use scientific techniques to find actionable insights for everyday problems, including how to respond responsibly in the next pandemic.

We publish every Thursday. If you have other topics to discuss or comments, let us know.

Leave a comment

This week we’ll look at whether personal science can become too much of an obsession, and what are better alternatives.

We all know “that guy” who is seemingly over-obsessed with tracking all things related to fitness and health. These people are over-represented in the high tech industry, where wealth and success are highly correlated with the invention of new gizmos that promise ever-higher productivity thanks to automated shortcuts. And yes, I’m proud to include myself in those ranks, for better or worse. Like many in the original “Quantified Self” or “Biohacker” groups, the subject resonated with me, not because I had stumbled upon something new and exciting, but because I learned that I’m not the only one who likes to treat my everyday life like a science experiment. In the tech world, you settle arguments by building stuff, not through fancy presentations or appeals to credentials. Shouldn’t we all strive to live that way?

But there’s also “Cyberchondria — escalating health anxiety from searching symptoms online”, Valetudinarian - “A person morbidly preoccupied with their health, usually with the connotation of elaborate regimens and instruments.” (like Jane Austen’s Mr. Woodhouse), and Healthism, Robert Crawford’s term for those who view health as an individual moral duty achieved through personal vigilance.

These words have different manifestations, but they all get at something real: the measurement itself was the harm. By trying too hard to be ‘personal scientists’, we can make ourselves worse off.

Back in the old days, a self-tracking obsession was limited by the amount of data you could collect. I remember years full of pen-and-paper notebooks where I carefully logged personal data about sleep, exercise, food, and much more. Wearables eventually made that much easier, and the limiting factor became analysis. “Dashboard” apps like Zenobase or dataviz platforms like Observable helped, but they still required some effort; you’d need a free evening or weekend to figure out how to connect your data, filter and manipulate it, and then prepare it in a form that made it easier to analyze. (see PSWeek240815 for many examples).

Ironically, this exposes a weakness at the heart of the personal science obsession with tracking and analysis. What if all those weekends spent collecting and analyzing data weren’t a troublesome but necessary friction standing in the way of an important insight? What if that time spent was really a filter that prevented me from asking the questions I should have been asking all along?

Now that a question costs thirty seconds, I ask worse ones, and I ask many more of them. The tooling that finally let me analyze everything also removed the only mechanism that made me decide what was worth analyzing.

Personal science walks the fine line between being open-minded (to even the most whacky-seeming ideas) and skeptical (about even the most respected people and ideas). A professional’s reputation and livelihood are at risk if they push too hard on an idea that the “consensus” rejects. (See PSWeek240104 for how often that happened under COVID). But personal scientists have no such restrictions. Our main constraints are time and imagination.

Professionals have access to expensive libraries of information, knowledgeable colleagues and skilled research assistants. But now, thanks to LLMs, we have all of that too. Hardly a week goes by without news of another unsolved problem being cracked by AI. Many of those solutions are being found by people who have no more access to Claude or ChatGPT than you do.

So once our analytical resources have caught up to the professionals, I wonder if our failure mode isn’t measuring or analysis; it’s letting what’s measurable crowd out the important — the streetlight effect turned inward. Spending so much time obsessing about the data collection and analysis may have come at the expense of trying to understand the decision that the answer would change. If I can’t name the decision, then it’s not science—it’s just a plain ole ordinary obsession.

So what’s next? I’m going to spend more time thinking of questions and less time obsessing about tracking or analysis. Sometimes you really can miss the forest for the trees.

Speaking of how easy LLMs have made the conduct of personal science, I took up last week’s question about Timestamps and used my new MCP to generate a full-size book Timestamps.

Or if you don’t like that one, here’s my Self-Experimenters Peptide Handbook.

Although (or maybe I should say because?) I used AI heavily in the making of these books, I think you’ll find them accurate, comprehensive, and readable—everything you’d want from a book-size treatment of a non-fiction subject.

Where is this headed?! I’m sure that real writers will authors will find plenty to criticize, but c’mon, the same is true of most human-authored books. I’ve been a voracious reader my whole life (see PSWeek260625) and I have to honestly admit that I’m finding these AI-authored books helpful as a quick way to survey a complicated new topic. What will we do in—pick a short number of years, like 5 or 10—when the AI models are many times better than today?

In PSWeek260716, we mentioned the great pricing on blood diagnostics at Goodlabs, and now I had a chance to try out their free service with Bloodworks Northwest, the blood donation center in Seattle. It works just like a regular blood donation (which you should do anyway), but they’ll take part of your sample and send to a lab that can return results for dozens of different blood parameters. I tried it with Cystatin-C (normally $48+) and IGF-1 ($139+), tests that are unusual and expensive enough that I ordinarily wouldn’t have bothered, but you can get these and many more:

Oh, and one thing I learned: you don’t need to be fasted for most blood tests. The old guidelines, that you shouldn’t eat 12 hours before a lipids (cholesterol) or other common tests—the latest studies say it’s not necessary. Check with your doctor (or LLM!) but except for a few insulin-dependent tests (e.g. glucose, insulin), it’s probably not important to skip breakfast before doing a blood test.

About Personal Science

Personal Scientists use science for personal reasons, not as part of a job. We treat science as a verb — something you do — not as a noun, and certainly not as a synonym for “truth” or “cool facts about nature.” Following the 1660 motto of the Royal Society, nullius in verba, we take nobody’s word for it.

But taking nobody’s word for it isn’t the same as measuring everything. The Royal Society’s founders weren’t drowning in data; they had to decide what was worth observing before they could observe it, a constraint that involved real work. Now that for us that constraint is disappearing, the discipline has to come from somewhere else — from us, deciding what actually.

We publish this short newsletter each Thursday, free to anyone who likes to use science in their daily lives. If there are topics you’d like us to cover, let us know.

Leave a comment

This week: what AI-generated text means for reading, writing and Personal Science Week.

What’s the point of writing?

When Quantified Self co-founder Gary Wolf recently asked on Substack whether he should write an essay about timestamps, my first thought was, well, sure. But why should he write the article and not Claude? If I really want to know about timestamps, does it matter how I learn? I trust Gary’s judgement, so if he signs his name to a post, do I really mind if he used an LLM to generate the text?

You don’t care who wrote your car manual, as long as it’s accurate and well-written. If you need help fixing your dishwasher, you don’t care who or how the YouTube got made—you just want easy-to-follow and concise instructions. Why would a timestamp tutorial be any different?

Detecting AI-generated writing

Professional writers are, understandably, terrified—in the way blacksmiths and buggy whip artisans were once terrified.

No doubt that was the motivation behind Substack’s recent introduction, with algorithm company Pangram, of the feature “Scan for AI Text” that estimates how much of a post was written by a human vs an AI. Just click the three little dots at the top of the post and select it from the menu item

Pangram’s algorithm is actually quite sophisticated. Instead of simply looking for examples of what most of us have learned to recognize as AI-generated text, Pangram tries to classify writing by the author’s “voice”.

It works for the same reason that you and I can reliably tell the difference between a passage from the Iliad and an article in Sports Illustrated, even if we know nothing about either: the styles are so obviously different in each. Similarly, with enough careful training, a computer classifier can learn to tell differences between any author, whether human or otherwise. These classifiers can be good enough that they can distinguish between different LLMs as well—even when you tell the LLM to radically change its style.

You might think this is a ‘cat and mouse’ game where the LLMs and the detectors go back and forth trying to outdo each other, but it won’t be that simple. As LLMs get upgraded, they tend to inherit the styles of their predecessors, and any text with an identifiable style will be pegged by a Pangram-type classifier.

In fact, Arvind Narayanan used Claude Code to access the Pangram API and then iteratively modify his posts until they came back 100% human-authored. But he says now Pangram seems to have strengthened their detector and that tactic no longer works.

What about me?

Personal scientists don’t care about general statistics or claims that something works in general. We care about what works for us. So here’s what Pangram says about my recent post PSWeek260723

Which I think is probably about right. Most of everything I do these days is run through an LLM one way or another, and my writing is no different (See PSWeek260402 for details). A series of Claude “routers” and skills let me quickly pull up a summary of whatever I’ve encountered in the past. For example, when I look for “timestamps” I’m reminded of something I wrote back in PSWeek250206 along with a bunch of other related rules I’ve accumulated for various tracking purposes over the years.

So yes, after I’ve settled on an idea for a post, I’ll ask my Claude project to pull together anything related while I work interactively to understand the personal science consequences. Once I’m happy with the final version, I ask another Claude to do a critical review and fact-check.

Interestingly, I don’t think Claude saves me all that much time in the writing; the real time saver is in the analysis, which pre-LLM would have taken weeks—if I’d even bothered to do it at all. When I ran the Pangram algorithm against several other Substack writers I like, I found several that were mostly or even 100% AI-written. At first I thought it would be nice to show you some named examples, but you know what? I don’t care. Some of these “AI-assisted” posts are from people who are not native English speakers. Some are from busy doctors who otherwise wouldn’t have time to write a thoughtful well-researched post. But all of them—all the people I subscribe to—are real people who take responsibility for what they write. In each case, I learned something that would have been more difficult to learn otherwise. Isn’t that what matters?

Sure, there are some people who will tell their Chatbot to “write 100 different Substack posts about timestamps”, maybe assign pseudonymous author names to each of them and “flood the zone” in search of clicks and eyeballs. Maybe they’ll all be well-researched with plenty of accurate citations and workable examples.

But then what’s the point? If I really want to know about timestamps, I could have my own Claude write one.

I’m not a professional writer, nor do I aspire to be one. I write because (1) the act of expressing a thought coherently is the best way to internalize it for myself, i.e. to “decide what I really think” and (2) the discipline of forcing myself to publish something publicly every week is a good way to keep me focused. As Steve Jobs used to say “Real artists ship”.

The hard part is thinking that timestamps are important enough to write about in the first place. Claude will never replace that.

Personal Science Weekly Readings

Speaking of AI writing, this How I use AI (and How I Don’t) is a well-written explainer for how one professional writer, Jeremy Caplan writes his excellent Wonder Tools Newsletter. tldr; “I’m responsible for everything I write”, he says.

If you feel like you’re good at spotting AI faces, try this BBC Quiz that will present you with pairs like these and ask you to click on the “real” one:

And speaking of great personal science, Jane Cook How to survive boiling water is a well-written account of an experiment at an MIT dorm involving a carton of milk that has been sitting unrefrigerated since 1994. It’s a long story of fermentation and a new experiment where the author did PCR-testing on a commercial probiotic tea to see whether the microbes actually survive boiling (spoiler: they do, but it’s unclear whether that matters).

Finally, another plug encouraging you to subscribe to Wondertools, which always turns up things like this from a recent video discussion with writer A.J. Jacobs: a wine aroma training kit with dozens of individual scents (including, um, horse sweat). A.J. spends a few minutes each morning trying to identify them — smell being, as he puts it, our least-trained sense. $400 on Amazon, or roughly $30 to build your own.

About Personal Science

This newsletter is a weekly summary of ideas and techniques that we hope will help others who want to get better at personal science. If you have additional thoughts or topics you’d like to discuss, please let us know.

Leave a comment

I spent a few weeks trying to measure my own Zone 2 as precisely as I could, instead of reading it off a formula. This week I’ll explain how — and why the most interesting thing I found wasn’t a number at all, but the fact that my Zone 2 moves.

The backstory: winded on the stairs, fine on the bike

About five years ago I noticed I was getting seriously winded doing simple things. Walking up to the second floor of our house left me panting. That, plus a year-over-year drop in my Apple Watch–measured VO₂ Max, sent me straight to a cardiologist and a battery of clinical-grade tests. The tests concluded two things: (1) the Apple Watch VO₂ Max was badly wrong — the lab put me more than 20 mL/kg/min higher than the watch claimed — and (2) my heart is fine.

What I have instead is exertional dyspnea: an apparently benign condition where my body just takes a little extra time to “warm up.” Climb the stairs cold and I’m out of breath. Go for a 30-mile bike ride and I’m fine. What gives?

Why I don’t trust the formula

The famous podcasters insist we all chase “Zone 2.” The idea is that effort divides into five intensity levels (that you measure based on heart rate) from 1 (barely distinguishable from rest) to 5 (you can’t hold it more than a few minutes). Zone 2 is supposed to be the sweet spot: enough to build aerobic fitness, not so much that it wears you out.

You can compute your zones from formulas based on age, sex, and resting and max heart rate. Google it or ask Claude and you’ll find various online calculators that, for me, give numbers from 117 to 140. But these formulas are averages, computed over zillions of people. But to a personal scientist, I’m not an average — I’m me. So what’s Zone 2 for me?

How I went about finding it

The right tool is a chest strap — a Polar H10 (see PSWeek260409) — not an Oura, not an Apple Watch. Wrist optical sensors are fine for a rough heart rate, but they smear the beat-to-beat timing, and the beat-to-beat timing is the whole game here.

After lots of back-and-forth with Claude, the method I settled on is DFA α1: a number computed from the tiny variations between consecutive heartbeats that falls in a predictable way as you work harder. When it drops through about 0.75, you’ve crossed your aerobic threshold — the top of Zone 2. It’s the closest thing to a lab lactate test, and unlike a formula, it’s measuring me.

At first I tried a ramp test at a professional fitness center: warm up, then step the treadmill up a notch every three minutes until you can’t hold a conversation, watching α1 fall as you go. Clean in theory. In practice it was a mess. My tests conducted over a couple of weeks gave me numbers all over the map.

So I ditched the treadmill for something that fits me better: a 30-minute neighborhood walk I already do most mornings. I live in a hilly area with a set route that has a couple of steep ups and downs. Instead of trusting the math alone, I added a second, independent instrument — my own breathing. I pressed a marker button the instant my breathing broke from nose-only into open-mouthed huffing, the oldest and most honest threshold signal there is. It landed at heart rate 140. Better yet, the α1 math agreed to within a couple of beats. Two completely different instruments, one answer.

The real finding: my Zone 2 moves

But the real surprise was learning that my threshold isn’t a fixed number — it depends on whether I’m warmed up.

Two climbs on the same hill, fourteen minutes apart, at the same heart rate of 133: on the first, cold out of the gate, I crossed threshold and my breathing broke; on the second, warmed up, I cruised comfortably below it. So I don’t really have one Zone 2 ceiling. I have a cold one around 136 and a warm one closer to 148 — call it a working aerobic ceiling of about 140 once I’m moving.

My zones — and why the standard formula mis-sorts me

Here are the numbers for me, built on a measured max of 182 and a resting rate of 57 (so a heart-rate reserve of 125):

One interesting result if you want to use a zone calculator. The two most common formulas disagree wildly at the low end; the plain %-of-max-HR model would cap my Zone 2 at about 127 — thirteen beats too low. But the heart-rate-reserve (Karvonen) model puts it at 145 — within five beats of my measured 140. For me, at least, reserve-based zones track the physiology and percent-of-max doesn’t. If you’re going to use a formula, use Karvonen.

So what about my dyspnea?

Finding my zones ended up explaining it.

My dyspnea is a transition problem, not a capacity problem. The mechanism is slow VO₂ on-kinetics: at a rest-to-hard transition, oxygen delivery lags demand, the gap gets covered anaerobically, ventilatory drive spikes, and I feel winded — at a low heart rate, because my heart rate hasn’t caught up yet. Once I’m warmed up, the same load is comfortably aerobic. That’s why the stairs get me but I barely noticed the 2-hour-39-minute bike ride I did last weekend.

So the prescription is easy: deliberate sub-threshold surges. From easy up to about 140 for two or three minutes, back down, repeat — staying under my conversational ceiling (call it ~145) the whole time. It’s the same intensity I already ride at, just no longer smooth about getting there.

Best of all, I finally have the dyspnea as a number. My cold-to-warm gap was about 8 beats. If my on-kinetics improve with training, that gap should shrink. Now, that said, 8 beats is close to the run-to-run noise of a chest-strap threshold estimate, so before I can claim the gap is closing I need to log it across several mornings and watch the average move. But that’s the whole point of measuring anything — turning a vague “I get winded on the stairs” into a quantity I can actually track.

Personal Science Weekly Readings

Quantified Self co-founder and OG, Gary Wolf has a new substack. Drop everything you’re doing and subscribe! While you’re at it, please pre-order his new book: The Quantified Self: Learning to Observe.

Speaking of learning to observe, The Cosimo Research Journal of Non-Institutional Science publishes numerous interesting, if eclectic, crowd-sourced studies, worth reading, on subjects like Does Mouth Taping Improve Sleep (tldr; no)

Personal scientists keep saying that everyone is unique, but now Uri Bram has an online tool that lets you be more quantitative about it. For example, if you have a bachelor’s degree or higher, have children, and both your parents were born in the US: only 16% of Americans fit that description. Calculators like this are becoming easier to create thanks to Claude, so expect a zillion more in the future.

About Personal Science

Be open-minded, be skeptical: those are some of the tenets to remember if you want to use science in your personal life. We publish ideas and thoughts each Thursday, but if you have more topics you’d like to discuss, let us know.

Leave a comment

This week I’ll show what I learned from Goodlabs, a new company that promises free blood tests when you donate blood.

For the past couple of years, I suspended my regular blood donations when I found out that my blood iron levels had collapsed. (See PSWeek230427). I had been persuaded by the idea that male humans, in particular, had evolved in an environment that expected regular cuts and bruises, and that an occasional blood-letting and the re-charge that it forces, would be good for me. I still believe that, but apparently I overdid it (Pro tip: the official guidelines “no more than every eight weeks” is way too aggressive—I’d shoot for every six months or so).

Well now I have a reason to donate again: free blood tests! Goodlabs is a new diagnostic company with an idea that seems obvious: if you’re already donating blood, why not run some other tests at the same time? Meanwhile blood donation centers are always strapped for donors, so why not use the promise of a diagnostic as an incentive to donate? For personal scientists, a free blood test is way better than the cookies and orange juice they usually give you.

I signed up for my next blood donation, thanks to Goodlabs, but the company also offers a la carte blood testing at unbelievable prices, so I tried that too. Long-time readers know that I have essentially two go-to places for regular blood testing: SiPhox (for at-home testing, see PSWeek221222) and OwnYourLabs for discounted professional blood draws through LabCorp (see PSWeek220622). I’ve also tried Function Health and others (see PSWeek240912) Well, now I think I’ll standardize on Goodlabs.

After ordering on their website, the rest of the experience is just like you get with the other providers:

1. Confirmation email that lets you schedule a time and place for the professional blood draw. Both Quest and LabCorp have dozens of locations near me, so I just pick the one that’s earliest in the morning (I do fasted tests). Usually a 10-minute drive.

2. A PDF requisition order with a barcode. I print this and show it at the reception desk at the lab.

3. Results arrive in my account, usually within 24 hours.

Goodlabs has an export feature that lets you export as either CSV or the raw PDFs from the laboratory (Quest, in my case). I found the raw PDF to be more complete, with specifics such as fasting status and time of collection (which matters for some parameters like iron). Also some of the Goodlabs data is “derived” (i.e. a ratio computed from the raw numbers), and so redundant if you’re going to use Claude for the analysis anyway. So my advice: just download the lab PDFs and skip the CSVs, which you can always generate yourself later in Claude.

Goodlabs, like Function Health and SiPhox and most every diagnostic company these days, has a pretty UX to display your results along with an AI “chatbot” that can answer basic questions. The Goodlabs chatbot is pretty good, but honestly I think personal scientists will just want to export everything and study it ourselves. My Claude project contains way more information about myself and my history, including a connection to Apple Health. It’s just easier to get more comprehensive answers with my own LLM.

For example, here’s something actionable I learned thanks to Claude and the Goodlabs result:

I’ve been taking 1 softgel of a Sports Research Triple Strength Omega 3-Fish Oil (Costco $26 for a bottle of 150), and it seems to have actually improved my Omega-3 levels noticeably compared to the Nature’s Bounty brand I’d taken previously. I was able to show that Sports Research brought me into the normal range. (6.0, up from 4.8 in Jan 2025).

Claude also added that there’s a genetic variant (FADS1/FADS2 cluster, e.g. rs174537) that can affect EPA/DHA intake, and when I gave it my 23andMe SNP data it found specific variants that affect how my body handles fish oil. It’s always nice when genetics and blood results agree—a concordance that before LLMs would have required hours or days of additional research.

More on Goodlabs

I have no affiliation with Goodlabs other than as a (so far) happy customer. They gave me this code 9VWGYY (or this link) which I think might give you a discount. But with or without the discount, these prices are unbelievably cheap, and I can attest that everything about the experience is legit. Highly recommended.

Incidentally, the Goodlabs people told me they’ll be in Seattle in early August, so I’m trying to organize a biohacker/longevity event for people who are interested in meeting to discuss blood testing and more. The tentative date is Monday August 3rd, as I look for a venue (If you want to attend, or can help us find a place, let me know).

Personal Science Weekly Readings

We all know about the record-setting heat wave in Europe over the past month. Since air-conditioning is relatively uncommon there, news headlines emphasized the heavy death toll among the elderly and other vulnerable people. So then, asks the ornery team at Trust the Evidence , why can’t you see a rise in the average death rates in the UK compared to previous years? Their answer: because the warnings are based on modeled projections, which are always dependent on the assumptions that go into the model. In fact, people die for all kinds of reasons in good weather and bad. A reminder to (1) be wary of in-the-moment calculations made before all the data is in, and (2) never assume the “first draft of history” is correct. The world is a complicated place, and we should be humble when making statements like “this caused that”.

Speaking of academic research that makes me wonder, this new paper from team of microbiome and nutrition researchers reviews Applying Artificial Intelligence and machine learning to precision nutrition. It defines all the latest AI buzzwords (CNN, RL, RAG) and describes how each applies to precision nutrition. It’s fine and I’m sure given all the co-authors it’s been carefully edited and accurate for when it was made. But how actionable is it? Will it affect any future research program or clinical decision? what people really want to know is “Is such-and-such food good for me?”, or better yet “What diet will let me fix such-and-such problem”. I’m not sure that papers like this help lead to those sorts of answers, especially in a world where LLMs are soon able to do more comprehensive literature reviews than any human team can hope for.

Have we reached Peak Protein? Here’s a new, apparently serious product, “Man Cereal“ for people who really really want more protein.

Remember Colon Blow?

About Personal Science

“Following the science” used to mean obeying highly-credentialed people from prestigious universities and research labs, but these days there are so many different, often contradictory voices telling you what’s true, that it’s hard to know what to believe.

Personal scientists know that real science is never settled, and that what works in a laboratory study may not work for what matters to you or your loved ones. But science itself—careful observations, willingness to remain open-minded and skeptical—is the best way to uncover the truth, and it’s something we can practice every day in our own personal lives.

We publish this newsletter each Thursday. If there are other topics you’d like to explore, or if you have other questions, please let us know.

Leave a comment

This week I’m looking at germs: what should I know about how to keep food safe?

Personal scientists get skeptical when a nameless bureaucrat tells us what we can or cannot do, so I look suspiciously at any “expired by” label.

That’s how I felt when I opened my refrigerator and inspected the various uneaten containers inside. Some were easy to chuck: the cottage cheese was covered in green “fur”, the sour cream smelled obviously wrong. But I kept most of the remaining items: the heavy cream had “expired” weeks ago but it smelled and looked fine. Some of the cheese had a little mold on it, but nothing that I couldn’t cut around. The carrots and the onion in the bottom drawer looked (and tasted) okay. It’s not like I’m a hippie who refuses to throw anything away, but I don’t see a need to be extra careful either.

Back in PSWeek250109 we discussed food safety in the context of cooking, where I wondered why we are so paranoid about raw meat. After all, that’s what animals eat.

Most of us wash fruits and vegetables before eating (why else would you need a salad spinner?), but I wonder how necessary that really is. These days nearly all store-bought produce has already been treated with peroxyacetic acid (PAA), a vinegar-like liquid chosen because it degrades safely into harmless byproducts like water and oxygen. You can buy PAA for home use (SaniDate™) but why bother when your produce has already been washed with it?

Washing raw meat — especially chicken — can actually increase risk rather than reduce it. When you rinse a chicken in the sink, water droplets can aerosolize and spread bacteria (particularly Campylobacter and Salmonella, which are very common on raw poultry) onto the surrounding counter, sink, and nearby foods. This is a classic example of cross-contamination. Because you’re going to cook the chicken thoroughly anyway, the rinsing step usually removes very little of the risk while potentially spreading pathogens to things you won’t cook, like salad ingredients or utensils.

For the same reason, I generally don’t wash meat or vegetables I intend to cook thoroughly, unless there’s visible dirt or residue. The heat of cooking takes care of surface bacteria far more reliably than rinsing does.

Cooking, while effective, is also over-rated. Pretty much all pathogenic microbes die at temperatures above 75ºC (165ºF), which is the origin of the better-safe-than-sorry advice you’ll find in cookbooks. But that’s only if you want to be really, really safe. In fact, pathogens stop multiplying above ~120°F and are progressively killed over time at temperatures well below 165°F — it's the combination of temperature and time that matters, not a single magic number. When the scientists at Nathan Myhrvold’s Modernist Cuisine Lab tested USDA’s claims about microbe survival, they produced this chart:

In general, these microbes reproduce best at just above body temperature, but get even a little warmer (120º or so) and they basically stop growing. You don’t need to get anywhere near the 165º it takes to kill them.

That said, my personal science experiments have not been without risks. My lackadaisical attitude caught up with me recently—not from my past-the-due-date food from the refrigerator, but when I ate some reheated leftover Costco Southwest linguini with sausage. It looked and smelled fine. And I nuked it well beyond 165º. So why did I get sick?

At first I blamed “fried rice syndrome,” the classic Bacillus cereus illness. Its spores shrug off both cooking and freezing, hibernating until the food sits out long enough for them to wake up and multiply. As they grow they pump out cereulide, an emetic toxin so heat-stable that reheating destroys the bacteria and leaves the poison untouched — which is why leftover rice and pasta can sicken you even when they look and smell perfect. Staphylococcus aureus runs the same play with its own pre-formed toxin. No amount of microwaving saves you.

But my dish was sausage, and meats and gravies are the calling card of a different bug: Clostridium perfringens, which you and I already carry in our guts at low levels. It doesn’t poison the food; it poisons you. Its spores survive cooking, multiply while the dish cools slowly or sits warm, and then — once you swallow a big enough live dose — sporulate in your intestine and release their enterotoxin there. Here reheating should have protected me, except a microwave heats unevenly: “well beyond 165º” is an average, not a guarantee, and the cold spots are exactly where the survivors ride through.

So the personal science lesson is: use your senses, but be careful. In general, traditional foods that were available before refrigeration (e.g. butter, yogurt) can last a while; but things that have traditionally always been cooked (e.g. pasta, rice) had better have some other means of preservation, like salt or fermentation, before they’re worth eating unrefrigerated.

Personal Science Weekly Readings

When Euan Ashley published one of the first clinical analyses of a human genome back in 2010, it was a big deal that required a team of 30 people working for 9 months. Over the July 4th weekend, he tried it again with Claude. He was able to get the same results in 30 minutes using an 800-word prompt and about $5 worth of tokens. As he says, “There is no world in which this is not utterly remarkable.”

Eric Topol, who is as optimistic about AI in medicine as anyone, cautions that “there are no studies in real world medicine; all we have now are simulations, case vignettes, patient actors, etc.” One problem is that medical publishing is extremely slow by computer industry standards, with many of today’s headlines serving up results from models that are a year or more old. Yishan Wong built a Github repo that lets anyone re-try those results with any model, including new and open-source ones.

And check out Open Science, an open source competitor to Claude Science.

You give it a goal, and it works through the research loop the way a capable collaborator would. It reads the papers that matter, forms a hypothesis, writes and runs code, runs experiments on real compute, queries the major scientific databases, and writes up the result. It runs as a workspace in your browser and works with any frontier or open-weight model from Anthropic, OpenAI, Google, and dozens of other providers, using your own API keys. No account is required.

About Personal Science

Personal scientists don’t take the label’s word for it — but we don’t take the contrarian’s word for it either. A printed date is a quality hint, not a safety command; In each case the interesting truth was one layer below the slogan, reachable by anyone willing to check. Nullius in verba — take no one’s word for it — applies as much to the “expired by” stamp as to the influencer telling you it’s fine.

We publish every Thursday. If you’ve run your own experiment on food safety, raw milk, or germophobia, let us know.

Leave a comment

This week I’ll summarize some of what I saw.

Biohackers, like personal scientists, are open-minded about new techniques for optimizing health and wellness. That’s why I enjoy attending biohacker conferences like the one held last weekend in New York City. More than 1,000 people paid $300 for a two-day pass that includes talks by various health-optimizing speakers, training sessions, and access to an exhibit hall featuring dozens of biotech-related products.

There is often a fine line between genius and crazy, between under-rated medical miracle and scam-driven snake oil. If you’re somebody who relies on “experts” to decide the difference for you, then you won’t enjoy biohacking events. But we personal scientists are open-minded and skeptical about everything—whether the claim comes from credentialed mainstream researchers or from sincere amateurs. So that was the attitude I took while passing through the booths at the show.

Peptides!

The biggest trend I’ve noticed at these shows is the surging interest in peptides. Several booths were selling the usual suspects—BPC-157, TB-500, CJC-1295/ipamorelin, and various thymosins—most of them labeled “research chemical, not for human consumption,” the legal fiction that lets sellers sidestep the awkward fact that none of these is an FDA-approved drug.

Knowledge of peptides is an occupational hazard for personal scientists, and I’m doing my best to come up to speed. But like I keep saying, personal scientists are open-minded and skeptical, so I don’t try anything—especially something that goes directly into the bloodstream—unless I see very clear evidence that it’s safe. For BPC-157 (“tissue repair”), that evidence still doesn’t exist: the encouraging results are almost entirely from cell and animal studies, and nobody has established a safe human dose. “Popular at the expo” is not a clinical endpoint.

Sleep in a bubble: AirTulip

AirTulip (airtulip.co) seemed like a genuinely clever idea for people who suffer from allergies. It’s a headboard that runs roughly $2,000–$3,500 depending on size—founded, appropriately enough, by a Dutch aerospace engineer—that uses laminar airflow (the unidirectional, non-mixing flow used in industrial cleanrooms) to push HEPA/ULPA-filtered air across your pillow. Instead of filtering the whole room, it creates a “100% particle-free” zone right where your face is. The company cites testing at Eindhoven University of Technology and has appeared on Shark Tank.

Light and sound: Braintap

I’ve seen Braintap at these shows before. They make a headset that uses carefully timed sequences of colored light flashes to help your brain enter a state of better concentration. This time I was able to talk to the founder, Patrick Porter, who explained how this is more effective than binaural beats (your brain quickly adjusts to sounds).

Photic entrainment is real: rhythmic light does produce measurable EEG responses (steady-state visual evoked potentials), and 40 Hz “gamma” light-and-sound has real, if early, clinical evidence—mostly in Alzheimer’s patients, though, not healthy people chasing focus. The weak link is that consumer devices like Braintap—and roXiva, which I’ve tried—are open-loop: they deliver a precise stimulus but never measure whether your brain actually synchronizes to it. When I tried a roXiva session I came away feeling more relaxed, but the effect quickly faded—exactly what you’d expect from a pleasant experience that isn’t necessarily changing anything durable. Maybe if you do it regularly you’ll have different results.

Go to personalscience.com/docs to see a complete list, with photos, of all the products I saw, including some for beauty and hair tools and the Oxefit AI strength trainer.

Personal Science Weekly Readings

I used my new Claude skill to help me look through the show booths. I took photos at each booth, and after I got home I pointed Claude to the photos. It automatically read the text from the photos, looked up the company website for more information, and generated a handy summary of the exhibits I’d visited, making it much easier to remember the products I saw. You can add this to your Claude and give it a folder of photos:

claude plugin marketplace add Proofbound/conference-booth-report
claude plugin install conference-booth-report@proofbound-plugins

If you miss the days when people were recruiting subjects for COVID studies, SeaPrep is now looking for volunteers for their study of respiratory infection outbreaks. You must live in the Seattle area and be willing to provide nasal swab samples if you or someone in your household comes down with something.

People walking in crowds tend to move counter-clockwise, a phenomenon that seems to be true in all cultures and ages studied. I wonder how hard it would be to make an iPhone/Watch app that would keep track of my movements throughout the day and see if I more often turn to the left (counter-clockwise) or right? (Spoiler alert: not hard.)

In the age of AI, analysis is easy and data collection is hard, but the hardest part of all is thinking of interesting new questions to ask about the data. One cool example: After studying data on how many people are displaced after dams are built in rural areas, water resource management experts in Finland think we may be greatly miscalculating the number of humans . Headlines promptly inflated this into "1–3 billion more people on Earth," which the lead author has explicitly disowned—the paper never estimates a revised global total. Still, once you see how census-dependent the UN's ~8.2 billion figure is, it gets easier to distrust the precision. Countries have a built-in incentive to overestimate their population numbers (e.g. Africa’s largest country, Nigeria, hasn’t had a census since 2006 and allocates government money based on reported population).

About Personal Science

Personal Science is the process of using the scientific method to solve problems and get better results on an individual, personal level. Following the motto of the Royal Society, established in 1660, nullius in verba, we take nobody’s word for it.

This newsletter is a weekly summary of a few observations we think will be interesting to anyone who wants to be a personal scientist. If you have other topics you’d like to cover, let us know.

Leave a comment

But for the past two years I’ve barely cracked any books at all, and so far this year my tracker shows I’ve read exactly zero books.

What happened?

Tracking your books

Personal scientists like to keep score. In the old days, some of us would keep a notebook or a spreadsheet whenever we finished a book. Now it’s easier to just use Goodreads. Once you’ve entered your books, the site allows some rudimentary graphic views to help you visualize how much you’ve been reading across time. And—critically—you can export all your info to a CSV file for deeper analysis.

I’ve tried other book-tracking sites, notably LibraryThing, but ultimately came back to Goodreads because too many of my friends are there. (Friend me on Goodreads or LibraryThing).

Here’s my Goodreads scorecard for the past few years. Note there’s nothing in 2026, not because I forgot to enter anything. Because so far I haven’t completed any books. Last year was a poor reading year as well, and even the books that I did read were mostly concentrated in the first few months.

It’s Not Just Me

A CBS News/YouGov poll from earlier this month finds a third of Americans say they read fewer books for pleasure than a decade ago, across every age and education group. A quarter don’t generally read books at all; only 12% clear twenty a year.

The poll also asked where the hours went: of those reading less, about half blame lack of time, about half blame too many distractions, and more than a third say their attention spans have gotten worse. In place of books: TV and movies, social media (named by nearly two-thirds of under-30s), and gaming (more than half of men under 50). My own biggest new distraction isn’t any of those.

The Real Reason I Don’t Read

The Wall Street Journal says sales of serious non-fiction (aka “Dad books”) have been in steep decline. They blame podcasts, and I agree that’s a factor for me, but it’s only part of the story.

The bigger story for me is LLMs, and fellow personal scientist Tim Ferriss nails it when he asks Has AI Already Killed How-To Nonfiction? with a detailed stats from the supply side—the big drop in sales of his own books.

His five #1 bestsellers had been a dependable annuity. Then the BookScan print numbers year-over-year fell off a ledge of their own: −5% in 2023, −13% in 2024, −46% in 2025, and a −57% run-rate so far in 2026 — roughly 80% fewer print copies than 2022, almost all of it after LLMs went mainstream. This is consistent across the industry: Publishers Weekly had adult nonfiction down 9% in Q1, with self-help the worst subcategory at −26%.

Fundamentally, the Tim Ferris books I like to read are a lookup table: “how do I lose fat / fix my sleep / automate my income.” In 2019 the best interface to those answers was 600 pages. In 2026 it’s a free chatbot that has read his books and ten thousand others and will hand you a personalized protocol in fifteen seconds. Prescriptive nonfiction, he argues, is just the canary in the coal mine; how-to YouTube, advice podcasts, and online courses are next, because the LLM is becoming the interface to everything.

It’s already true for me. I have a Claude Project that contains the manually-entered summaries and takeaways from decades of book reading. But now, when I hear about an interesting new title, I ask Claude to summarize and help me contextualize it relative to what I’ve already read. My Claude rules also pull reviews and counter-arguments to the author’s thesis, presenting me with a pretty good description of the book before I even think to crack it open. The best part is that the summaries are not only highly-personalized, but also interactive. Before long, I’m interacting with the ideas behind the book—which I often find more fulfilling than reading the book itself.

My personal experience shows that the carnage will not be limited to how-to books. All non-fiction (and probably all but the best fiction) is becoming a poor, inefficient substitute to a meaningful LLM interaction.

Is this good?

Arnold Kling thinks reading fewer books is just fine because overall we’re reading as much as ever, only of shorter and more personally relevant online pieces (like this Substack!) Many/most books didn’t deserve more than that anyway. (like PSWeek! 😬)

Ultimately I think this will all reverse itself, though in directions we may not be able to predict. Remember the “digital divide”, that long-ago fear that the socioeconomically-disadvantaged would miss out on the power of technology? Media theorist and Marshall McLuhan scholar Andrey Mir points out the digital divide has already reversed itself: now it’s the affluent and well-educated people who read physical books.

The downside of less book-reading is that I don’t know what I don’t know. Sometimes it’s better to spend 20 hours with a smart author who leads you down paths you didn’t think were important. As we noted in PSWeek251002, sometimes the best ideas are the ones you weren’t looking for.

Personal Science Weekly Readings

Speaking of reading, I liked Naomi Kanakia‘s tongue-in-cheek thought experiment of how probably there’s some MIT psychologist studying ways to “prove” that book-reading is bad for you:

That the stream of words in a book trained the human brain into a habit of self-consciousness, that reading books forced human beings to think of themselves as a stream of text, processed through time, making a coherent argument of some sort. And that this overall flattening effect forced readers to ignore aspects of their personality or their situation that were not otherwise in line with the overarching story they’d created about themselves

Meanwhile, If you are going to read, so much of the best content is free: Project Gutenberg has 70,000+ public-domain titles; Planet eBook formats many of the classics nicely. Anna’s Archive is the largest catalog anywhere (caveat emptor). Calibre manages it all locally and converts between every format. To read with others, hypothes.is and Fermat’s Library let several people annotate the same passage.

Jonathan Bi recommends these as the best four philosophy books for understanding how AI will change everything: 1. Max Weber’s Protestant Ethic, 2. Aristotle’s Nicomachean Ethics, 3. Rene Girard’s Deceit, Desire and the Novel, 4. Jeff Kripal’s Secret Body. I’ve read the first three, but Claude gave me a good enough summary of the fourth that maybe now I don’t need to bother. 🤣

ARPA-H launched a program to redefine sleep as a measurable, controllable driver of health. If you have ideas for projects to fund, check out their application page. (due August 12, 2026)

SOND just got $7M in funding for new Airpods-like sensor that tracks respiration, HRV, snoring and a bunch more metrics that it feeds to a real-time sleep coach.

Meanwhile, Kimba is a new $600/year bedside gadget that releases carefully-timed scents that supposedly will help you sleep better. The company notes that smell is the only sense that reliably influences sleep without waking you.

Speaking of wearables, Rock Health released their latest survey of 8000 people, which among other results shows that Apple continues to dominate with almost 2/3rds of the market.

About Personal Science

Personal Science is about using the methods of science for personal rather than professional reasons — delivered each Thursday as precisely the kind of short, online substitute for a book that this whole issue is about. The irony isn't lost on me. So read fewer books if you want; just don't take my word for what that's doing to you — not mine, not Ferriss's, not the chatbot's. Pull your own numbers. Nullius in verba.

Do you read fewer books now? Have LLMs reduced your interest in tackling new titles? Let us know:

Leave a comment

This week I’ll summarize some of the tips and tricks I’ve learned while using Claude Code and other LLMs.

By now pretty much everyone has used ChatGPT and its cousins, including my favorite LLM Claude. But most people still use it as a “better Google”, a convenient way to get a quick answer to a specific question without the need to click through to another site. If this is how you use it, you may not even see the need to upgrade to the paid version, since most of the answers are adequate—at least as good as what you get from Google. And even an old-fashioned Google search now gives you an “AI Overview”. If the free version is good enough, why bother with more?

Oh boy, to answer that question where do I start?! I subscribe to the Claude Max plan (either $100 or $200 / month), so my actual cost is capped, but here’s what it would cost me just last week if I had to pay by the token. (A ‘token’ is the unit for how LLMs process information. It’s a little complicated but you can think of it as slightly smaller than a word or about 4 English characters; a typical issue of PSWeek runs about 2000 tokens.

How did I use up that many tokens in a single week? Here’s a brief overview

How I use Claude

The first rule of effectively using an LLM is give it as much context as possible. If you only do Q&A-style prompting from the chatbot interface, the amount of help it can offer is limited to what you can type. I first started using Claude Code last summer (see PSWeek250710) because it makes it easy to add hundreds and thousands of additional documents to every prompt. Your computer is probably organized around folders full of related documents (e.g. taxes, photos, budgets, health records, etc.). Claude code lets you use the entire folder as the context for every prompt. When I ask Claude a health-related question, for example, it has access to every single medical record, including images, PDFs, spreadsheets; it takes all of that into account when it gives the answer.

Tips and Tricks

A few basic suggestions that will greatly multiply the usability of Claude, ChatGPT, Grok, or any other chatbot:

Customize! Every chatbot allows you to personalize your experience. Do it! Go to settings and introduce yourself and your priorities for how to get the most out of the system. I tell each of my chatbots as much about myself as I can (age, bio, interests). My Claude Code repos include very detailed claude.md files that make explicit every aspect of the project.

Use the terminal. You’re getting only a fraction of an LLMs power if you continue to exclusively use the desktop or mobile app. Claude CoWork is a reasonable substitute for a while, but trust me: nothing beats the terminal. Don’t know how to start? then…

Ask for help! One of the simplest tricks that people forget is that Claude knows way more than you do about how to use Claude. Chatbots reward people who just ask, simple, “How can I work with you more efficiently”. “Help me get started with a terminal”.

Compact! LLMs at heart are nothing more than a fancy auto-complete. When you hit ‘submit’ you are sending the entire conversation—not just the latest question—over to the LLM. If you start with a 100-token question and it answers with a 1000-token response, asking a 200-token followup question will actually send 100+1000+200=1300 tokens. That will compound with every interaction. So to save tokens, the best thing you can do is regularly clear the memory and start over on a new thread as often as possible.

Ask to generate html. LLMs normally give results in plain or markdown text, but if you want to share your results, ask Claude to turn it into a pretty-formatted html page that you can copy/paste into email.

For example, I recently attended a networking and brainstorming session at the excellent Idea Factory at Compound VC in New York. Each participant was asked to come with a few ‘breakthrough’ ideas to kick things off, so I generated a proposal for ‘exotic’ probiotics.

Use Skills. When you find yourself repeating the same task over and over, ask Claude to turn it into a skill. These are self-contained packages of instructions (including code) that make it easier to do something that would otherwise require lots of repeated explaining.

I’ve published several of my most useful skills here: Proofbound.com/workshop. One of my favorites is Conference Booth Report: Point to a folder of photos taken at conference booths and get a summary like this one

Personal Science Weekly Readings

Speaking of LLMs, back in PSWeek260423 we discussed OpenEvidence, a doctor-exclusive search engine for heavy-duty medical questions. Now a large study published in Nature Medicine claims that off-the-shelf LLMs like Gemini and Claude out-perform OpenEvidence and a similar one from UpToDate. No so fast, says @adamrodmanmd: it’s really hard to measure the ‘correct’ answer on some of these tricky medical questions, and ultimately what really matters is how effective you are at prompting. My opinion: stick with Claude. The frontier LLMs are getting better, faster, and any temporary advantage of the specialized LLMs will be overtaken by the better UX and tools available to the general-purpose ones.

For years, cancer researchers have studied p16-INK4a, a protein involved in a tumor suppression, also considered a key biomarker of aging. But now, Sholto David found, hundreds of papers have been published using the wrong antibody, apparently because it’s easy for labs to mistakenly order from suppliers another, incorrect p16 antibody instead.

Calcium and vitamin D probably aren’t saving your bones. A new BMJ systematic review (Massé et al., 2026) pooled 69 randomized trials and 153,902 participants and found little to no effect of calcium, vitamin D, or the combination on fractures or falls.

About Personal Science

Professional scientists have fancy degrees and spent their time writing grant proposals. Personal scientists skip all that overhead and focus on what really matters to our everyday lives: practical, experimentally-based answers to questions and solutions to problems.

If you have questions or comments about personal science, let us know.

Leave a comment

This week we’ll look at what’s changed — and how a personal scientist can decide for themselves when the experts won’t.

The Downside of Alcohol

The downsides of alcohol at a population level are undeniable. Nearly all sexual abuse and most crime is accompanied by drinking. A massive 2016 global study commissioned by the Gates Foundation concluded that “the safest level of drinking is none” — and in early 2025 the outgoing U.S. Surgeon General went further, calling for cancer-warning labels on every bottle. That may be why many notable leaders, from Presidents Biden and Trump to Warren Buffett, never touch the stuff.

Personal scientists prefer to make up our own minds, so we don’t delegate our decisions to experts and their fancy studies, so we’re not surprised that the “no safe level” consensus has been publicly splintered. To inform the 2025–2030 Dietary Guidelines, the government commissioned two evidence reviews that reached opposite conclusions — the National Academies found moderate drinking associated with lower all-cause mortality, while a second federal panel found risk climbing from the first drink. The tie was never broken. In January 2026 the new guidelines simply dropped the numbers: no more “two drinks for men, one for women,” just “drink less.”

The story didn’t end there. Just this week the stricter analysis — the Alcohol Intake and Health Study, federally commissioned to inform those very guidelines — was published in peer review anyway, reporting no protective effect at low intake and an alcohol-attributable death risk of about 1 in 25 at 14 drinks a week, with a recommendation to tighten the limit to a single drink a day for everyone. So within five months the government both removed its numeric limit and watched its own commissioned scientists publish the case for a stricter one. Did the government cave to Big Alcohol pressure, or did the Prohibitionists just have an ax to grind? Like we keep saying, most of us decide who to believe before we decide what to believe.

A key premise of personal science is that we’re all different and your experience could differ. But when the experts fracture this badly, “follow the guidelines” stops being an option — so how do you decide?

Alcohol Isn’t That Popular

First, a reality check on the social pressure. Most drinkers assume “everybody does it,” and are surprised how many people never touch the stuff. As of 2025, only 54% of U.S. adults say they drink at all — the lowest Gallup has measured since it began asking in 1939, down from 62% just two years earlier. Forty-four percent now call themselves total abstainers.

The drop is steepest exactly where you’d least expect it. Young adults (18–34), who two decades ago drank more than anyone, have fallen from 59% to 50% in two years — now drinking less than their parents and grandparents for the first time on record. And among those who still drink, it’s light: about 2.8 drinks a week.

That said, most of the decline has happened among those who didn’t drink all that much in the first place. Among the heavy drinkers, not much has changed.

It isn’t only an American story, nor a new one. More than a quarter of Italians don’t drink at all — a similar share as Russia — and today’s French drink only a quarter as much as they did in 1960. You’re not weird if you abstain. Increasingly, you’re the majority.

How Alcohol Affects My Body

The known effects are real but modest: alcohol shortens the time to fall asleep while degrading sleep quality (especially REM), temporarily raises heart rate, and shifts the gut microbiome. Although I’ll still occasionally have a drink in special social situations, I concluded through my own testing that alcohol just “isn’t worth the squeeze.”.

Case in point from a wedding I attended last month: half a glass of champagne and a Stella — about 20 g of ethanol — finished roughly 3.5 hours before bed, with the Polar H10 recording all night. The first five hours were wrecked. My heart rate held at 75 bpm (a clean night runs ~64), and rMSSD — the beat-to-beat metric that tracks parasympathetic recovery — collapsed to 27 ms, about half my clean-night baseline of ~54. Then, as the ethanol cleared, the second half rebounded: rMSSD climbed back toward 45 ms and heart rate fell. Whole-night LnRMSSD came in at 3.53 against a morning baseline of 4.05 — roughly two and a half standard deviations down. (I carry the ADH1B slow-metabolizer variant, so I clear it slowly; your mileage will differ.)

One clarifying detail: my heart rate’s floor for the night was identical to a clean night. What alcohol crushed wasn’t my minimum heart rate — it was the variability, the flexibility of the system. That “first half ruined, second half recovers” shape is exactly what Whoop and Oura have published at population scale.

Now this is just one night—an anecdote confounded by all the rest of the wedding-related travel, eating, socializing—but I don’t need more data to confirm what I already know: drinking and restful sleep are not compatible for me.

Red Wine Headaches

Do you get a headache after red wine? UCSF researchers think it might be related to how your body reacts to the flavanol quercetin, which is found in higher quantities in expensive, aged wines and is also sold as a cheap supplement. Here’s how to test it on yourself:

1. Drink a form of alcohol that doesn’t contain quercetin (vodka, gin, whisky).

2. Take a quercetin supplement (under $10 for a month).

If you get the headache, you’ve found your culprit — and ironically, because aged wine has more quercetin, you may find that cheaper wine treats you better.

Quercetin is wine-specific, but it’s one case of a broader rule: the darker and more “crafted” the drink, the more congeners — fermentation byproducts like acetaldehyde and methanol that track with hangover severity. It’s why Tim Ferriss’s standing advice is clear, low-congener alcohol with club soda, popularized lately as “just drink a White Claw” — hard seltzer being low-congener and low-ABV. The skeptic’s footnote: congeners modulate the severity; ethanol is still the toxin doing the real work. A clear drink is a smaller tax, not no tax.

Or you could skip ethanol entirely. I tried ZBiotics (a B. subtilis strain that breaks down acetaldehyde in the gut) with no effect I could detect — but I never tested it rigorously. Now I’m testing Hard Ketones, which swaps ethanol for R-1,3-butanediol (”ketohol”) — a non-intoxicating alcohol the liver converts to ketones rather than toxic acetaldehyde. The pitch: a buzz with no hangover, appetite suppression, and raised blood ketones. I’m withholding judgment. When a food chemist tested it on herself under continuous blood monitoring, the result was “not a slam dunk” — inconsistent ketosis, no reliable buzz, dose hard to dial in — and the molecule was originally shelved because the lab mice started stumbling. Loud claims, thin independent data, trivial to measure on myself: a perfect personal-science candidate. Results in a future issue.

Personal Science Weekly Readings

Speaking of measuring sleep, we’ve previously recommended the Polar H10 for serious HRV measurement, but people who don’t like the chest strap often go for the Oura Ring. But one of my favorite weekly email newsletters Recomendo, suggests the RingConn Gen 3 as a better alternative because it doesn’t have the monthly subscription.

A brilliant new Lesswrong piece takes this 2013 observation from Scott Alexander:

Byte for byte, an average Game of Thrones reader/viewer probably has as much Game of Thrones information as a neuroscience Ph.D has molecular biology information, but getting the neuroscience info is still a thousand times harder.

and proposes a way to use LLMs to translate intricate details of cell biology into easy-to-remember stories.

And speaking of stories that try to explain complicated things, please read The World Is Bigger Than You Can Imagine by economist Scott Sumner, who offers powerful reminders of how national economies seem to plod along despite all attempts to predict or change them. Almost any grand summary of economies, history, cultural products like movies or music—it’ll barely scratch the surface of what’s really happening. We humans know far less than we think.

About Personal Science

Personal scientists prefer to figure things out for ourselves. We listen to experts — not because they’re always right, but because they’ve usually thought about it longer. This year the experts handed us a rare, clarifying gift: two federal panels looked at the same evidence and disagreed so completely that the government gave up on a number. When that happens, Nullius in verba — take no one’s word for it — stops being a slogan and becomes the only practical option.

We publish every Thursday for anyone who uses science for personal rather than professional reasons. If you have topics you’d like us to explore, let us know.

Leave a comment

This week we’ll dig deeper into the actual privileges afforded by the “doctor” credential, and what it means for personal scientists.

Nobody asks a programmer for a diploma

I grew up in the personal computing industry, where nobody cared about your credentials. They cared whether your code ran. That really hasn’t changed. The compiler is the most ruthless examiner ever built — it doesn’t read your transcript, it just tells you, instantly and without sympathy, whether the thing works. It is nullius in verba enforced in silicon: take no one’s word for it, least of all your own.

Just one example: Matt Mullenweg was a jazz saxophonist studying political science when he co-created WordPress; he dropped out, and the thing now runs something like 42% of the web. No CS degree, no permission slip, no board exam. The work was directly inspectable. Who cares about his credentials?

But software engineering is different from “doctor” because it has two properties at once — the output is self-verifying, and there’s no legal gate in front of doing it. Other professionals must rely on credentials and we tell ourselves the credential certifies competence, when mostly it certifies the right to do gated things.

The anatomy of a “doctor”

I worked with Claude to build the following table of doctoral credentials, ordered not by prestige or years of training but by the breadth of legal monopoly each one grants. Watch what happens to the other columns as you read down.

Read it down the “monopoly” column and the gated powers widen from nothing to everything. But read down “what it takes to get” and the years barely move across the middle five rows — yet the legal powers swing wildly. Effort, legal power, and certified competence are three separate things fused into one word.

Dentistry and podiatry are interesting for the fence placed around their practices: only one part of the body. The credential isn’t a rank. It’s a territory.

And notice the rows marked with the state-level caveat (dagger symbol). These professions — naturopathy’s very existence, optometry’s fight to hold a laser, chiropractic’s periphery — are exactly the ones whose boundaries are still being litigated. The core of medicine and dentistry don’t vary, because their territory was won a century ago and nobody’s contesting it. Jurisdictional variation is a turf-war tell: it marks the monopolies that aren’t settled yet.

So what does the title actually buy?

An old joke that is more true than we’d like:

“What do you call the dumbest, least competent graduate of medical school?”

Answer: “Doctor”

But strip away the title, and a clinical doctorate is a bundle of specific legal monopolies:

• The prescription pad — the crown jewel. Controlled-substance authority, gated by a DEA registration that requires the license first. This single privilege is what puts the chiropractor below the naturopath.

• Signature powers. Death certificates, involuntary psychiatric commitments, disability and medical leave certifications, DOT and FAA exams, medical-marijuana and disabled-parking sign-offs — documents only a licensed clinician can legally sign.

• Billing and ordering. Diagnose as a billable act, order labs and imaging, admit to a hospital.

Underneath all of it sits the National Provider Identifier — and it’s the perfect anti-credential. There are over 9 million of them, and CMS states plainly that issuance “does not ensure or validate that the provider is licensed or credentialed.” It’s a free, ten-minute identity number that authorizes nothing. Registration is not authority; the NPI is just the floor the real monopolies are built on. (We talked about NPI in PSWeek260423 )

Incidentally, in some states it’s illegal to call yourself a “doctor” in a clinical setting, to stop patients inferring privileges the license doesn’t grant. Much as we can admire the First Lady for her academic achievements, I don’t recommend you call on “Dr. Jill” in a medical emergency.

The caveat

Before declaring it all a racket, it’s important to remember why that gate is there. Personal scientists love to experiment, but some things are best left to people who actually know what they’re doing. There are some very dangerous drugs out there; surgery is not for amateurs. Credentials don’t guarantee competence, but in this case it would be dangerously mistaken to think that gatekeeping is worthless. It’s that we’ve bundled a real safety function together with a legal monopoly and a status honorific, and then called the whole package “competence.” A personal scientist’s job is to keep the three separated.

Personal Science Weekly Readings

Speaking of people who fake credentials, remember that con-man character, Frank Abagnale, played by Leonardo DiCaprio in Catch Me If You Can? the guy who faked his way into all kinds of things, including a stint as a medical doctor? Turns out the whole story may itself be a con. A new, well-researched book claims he made up all his exploits in order to get work as a “reformed bad guy”. Which is itself a pretty good con, come to think of it.

And if you’re convinced by that recent study in Science that concluded LLMs are better diagnosticians than most doctors, check out the counter-argument from GP and FDA Advisor Dr. Shantanu Nundy MD, MBA: Large Language Models don’t reason better than physicians - yet. He argues that the study evaluated LLM answers after the facts of each case had been determined. Real doctors earn their keep by asking the appropriate questions in the first place—before the LLMs are involved.

I’m in New York City this month where I’ve been meeting tons of personal science-minded people, including the CEO of Goodlabs, another low-cost blood testing company. They partner with blood donation centers to offer free blood testing when you donate. You can also order low cost tests (fulfilled by your local Quest Lab) at outrageously-cheap prices. Vitamin D: $14 , CMP: $5, OmegaCheck: $42. These are the lowest prices I’ve seen anywhere, ever. I’ll try them when I’m back in Seattle and let you know. (Note: be sure to select “Quest” in the drop-down on the price list or it will default to New York State prices, which are ridiculously high compared to the rest of the country).

About Personal Science

A credential is a permission slip — something the professional needs before the world will let them prescribe, operate, or sign. Personal science needs no such slip. Science is a set of tools anyone can pick up, not the private property of the credentialed, and the only examiner that matters in your own life is whether the thing actually works for you.

So we respect genuine expertise without mistaking the title for the competence beneath it. We listen to experts — then remember they routinely disagree, which leaves us, not them, on the hook for deciding which one to believe. Above all we run our own experiments and trust what we can measure firsthand, staying open-minded yet skeptical, and never forgetting that you are the easiest person to fool. Nullius in verba — take no one’s word for it.

Personal Science is delivered each Thursday to anyone who wants to use science for personal rather than professional reasons. If you have topics you’d like to see covered, let us know or leave a comment.

Leave a comment

This week I’ve been looking at examples of Zipf-like patterns in my own data.

Predictable Relative Frequencies

Anyone who’s worked on a large enough team is intuitively familiar with Price’s Law — the claim that the square root of contributors in a field produces half the output. Watch any episode of The Office: it’s pretty obvious which one or two people at Dunder Mifflin are doing the real work, and not always obvious what value the rest add.

There are good reasons to be skeptical when this is applied to your own work group — the original law (proposed by bibliometrician Derek de Solla Price in 1963 about scientific citations) has been empirically contradicted by later researchers, who found that real-world distributions are usually more skewed than √n predicts. But the underlying observation is real: outcomes are rarely distributed evenly. They cluster.

A more general version of the same pattern is Zipf’s Law. George Kingsley Zipf observed in the 1930s that in any natural-language text, the most common word appears about twice as often as the second-most-common, three times as often as the third, and so on. In English, the word “the” makes up about 7% of all words you’ll ever read; “of” is half that; “and” is roughly a third. Plot rank against frequency on log-log axes and you get a straight line.

The pattern shows up everywhere humans do things at scale: city sizes, citations, website visits, GitHub stars, app store downloads, book sales. It’s not a universal law — many systems are more skewed (log-normal, stretched exponential) — but as a first approximation for human-generated distributions, it holds remarkably well.

The interesting personal-science question isn’t whether this pattern exists in the world. It’s whether you can see it in your own data, and what it tells you when you do.

The Setup

I’ve had ActivityWatch running on my Mac since December, an open-source, local-first time tracker that I prefer over RescueTime (which we discussed back in PSWeek230720) because nothing leaves the machine. It records the foreground app and window title every few seconds, and a companion watcher flags when the keyboard and mouse go idle.

This is one of those tasks AI agents now do trivially. I just asked Claude Code to look at my ActivityWatch data and plot it. The whole pipeline took about ten minutes.

The Top 10

Six months (170 days), 786.6 active hours, 124 distinct apps:

The top four — Brave, VS Code, Claude, Obsidian — account for 67% of everything. Five and six (Mail, iTerm2) push that to 80%. The tail of 118 apps shares the remaining 20%.

Is It Zipfian?

Yes — and more concentrated than Price’s Law would predict.

A canonical Zipf distribution has exponent α = 1: rank 1 is twice rank 2, three times rank 3, and so on. My fitted exponent over the top 80% of usage is α ≈ 0.76 (R² = 0.91). A shallower slope (α < 1) would normally suggest a less top-heavy distribution — so the exponent alone understates how concentrated my usage actually is. The story isn't in the exponent. It's in the cumulative share:

• 50% of active time → 3 apps

• 80% → 6 apps

• 95% → 21 apps

Price’s Law predicts √124 ≈ 11 apps should produce 50% of activity. In my data, 3 apps do. My usage is roughly four times more concentrated than Price’s Law claims, probably because my Brave Browser is really a whole bunch of apps, if you think of web pages as apps.

A Second Dataset: Voice Calls

My computer activity is one thing. Can I find a similar pattern in other data?

I pulled six weeks of my carrier voice log — 146 calls, 1,163 minutes, 32 distinct numbers. Both rankings below are Zipfian. Interestingly, the two different methods don’t rank the same people first.

Calls ranked by total minutes

• 50% of call-minutes → 2 numbers (Mom and my son)

• 80% → 5 numbers

• 95% → 13 numbers

• Price’s Law would predict √32 ≈ 6 numbers for 50%. The actual count is 2.

Calls ranked by number of calls

• 50% of all calls → 2 numbers (Wife and Mom — together 87 of 146 calls, 59.6%)

• 80% → 9 numbers

• 95% → ~20 numbers

The cross-metric observation

Look closely at rank 1 in each chart. My wife is my most-called number by frequency (45 calls), but ranks only 4th by duration (60 minutes). That’s about 1.3 minutes per call on average — quick “what time will you be home?” check-ins. Mom is the opposite: 42 calls but 428 minutes — about 10 minutes per call, real conversations.

Same dataset, two valid rankings, two different stories. Frequency captures transactional contacts. Duration captures relational ones. Neither one is “the” right ranking; they’re answering different questions.

A bigger caveat: the voice log is only one communication channel. My son shows up at rank 2 by duration because he’s in grad school building his own LLMs — a nerdy topic we both enjoy— so the calls run long. My daughters communicate by text and through their mother, so they barely register in my voice log even though they’re prominent in our family’s broader communication graph. A Zipf curve on one channel isn’t a Zipf curve on a relationship — it’s a Zipf curve on that channel for that person.

Why Time Isn’t Value

Before drawing conclusions, an honest caveat: time-on-app isn’t value-per-app, and call duration isn’t relationship-importance. Mail at 52 hours is mostly low-leverage maintenance. Obsidian at 92 hours produces the notes that become this newsletter. The 8 minutes I spent in Calculator might have been more consequential than an hour in Messages. And a 30-second call from one of my kids carries more weight than an hour-long sales call.

Zipf describes frequency, not importance. The wife-vs-mom contrast above makes the point cleanly: if duration were the only metric, you’d conclude I talk to my mother more than my wife — when really we’re just choosing very different communication patterns with each.

Why This Matters: Agency

In PSWeek250828 I argued that AI agents reward agency — the ability to delegate work without micromanaging it. Looking at the Zipf plots above reminds me of some of the potential consequences.

If three apps consume half my computer time and two people carry half my call minutes, the leverage isn’t in optimizing the long tail. It’s in being deliberate about the head. Am I using Brave for what matters, or am I in the browser by default because it’s where everything ends up? Is Obsidian time generating notes I’ll use, or am I rearranging tags? Could agents handle the Mail block (52 h, all maintenance) and free that for something the tail tells me I rarely do?

I want to think about this more — what the Zipf shape of your life means for how you should spend it, and what AI agents change about the calculus.

Personal Science Weekly Readings

• ActivityWatch — open source, local-first, free. If you self-track, this is the right tool. Works on Mac, Linux, Windows, and Android.

• Anyone with a Mac can do a casual version in two minutes: System Settings → Screen Time → App Usage. Less precise, but enough to see your own Zipf curve. Your iPhone Recents tab is the equivalent for calls.

• Speaking of laws about productivity and expertise, in PSWeek240711 we showed why the famous Dunning-Kruger effect — the claim that novices systematically overestimate their own competence — is probably a statistical illusion rather than a real psychological phenomenon. A useful reminder that “viral” findings about human nature deserve the same skepticism we apply to nutrition studies.

• The Kaguura Gichuru post (The Mathematical Reason Most People Never “Make It”) that got me thinking about this. Worth reading for the underlying intuition even if Price’s Law has been empirically contradicted by bibliometric researchers since the 1980s.

Finally pleaes note that I’m in New York City from now until late June, meeting a number of fellow personal scientists. If you’re in town, let me know!

About Personal Science

Personal scientists prefer to figure things out for ourselves. We listen to experts not because they’re necessarily right, but because they’ve had more experience than the rest of us. Still, experts are often wrong and frequently disagree with one another, so whether you listen to experts or not, you’ll need to make up your own mind.

When somebody publishes a viral claim like “the square root of contributors do half the work”, the personal-science move isn’t to argue with the abstract claim. It’s to check whether the pattern shows up in your data. Mine did, more strongly than the law predicts, across two independent datasets. Yours probably will too.

We publish every Thursday for anyone who uses science for personal rather than professional reasons. If you have other topics you’d like us to explore, let us know.

Leave a comment

This week we update our old post to see how personal scientists can evaluate hypnosis for themselves.

Hypnosis is another way of getting into the Flow

Hypnosis has always seemed a little bizarre, maybe even dangerous. We’ve all seen demonstrations involving seemingly normal people who, under the hypnotists’ spell, appear to willingly do crazy things. Is this just fakery? or is there something more significant happening?

David Spiegel, a Stanford professor of psychology has been studying hypnosis science for decades. He says we shouldn’t think of it as a trance – a spooky loss of control to a professional manipulator. Instead, hypnosis is just an example of a “flow state”. We all know this feeling, sometimes described as “being in the zone”, where time seems to stop and you feel thoroughly immersed in an activity. Hypnosis, like meditation, is just another, normal, desirable human psychological state we enter whenever we focus ourselves.

Spiegel makes an app, Reveri, which attempts to package his hypnosis sessions into a series of recorded steps. For $25/month, the app claims to help you achieve your goals of healthier eating, reduced anxiety, better sleep, and more.

Spiegel says about 2/3rds of people are hypnotizable, and he has a simple test to see if you’re one of them. The “eye roll” test checks whether the whites of your eyes are visible when you look straight up without moving your head. Watch it demonstrated in an Andrew Huberman podcast episode with David Spiegel. (See Podcast Notes and also notes from Juan Pablo Aranovich). Time magazine has a short overview of the subject, including additional details from Spiegel.

Testing the app

In December 2025, npj Digital Medicine published the first large outcome study of Reveri itself: almost 84k users across 300k sessions over four years. Within-session stress dropped on a 10-point Likert scale every time, with Cohen’s d between −0.71 and −0.78 across the first ten sessions. By psychology’s loose standards, that’s a large effect. Standard, interactive sessions worked roughly 1.8× better than brief, non-interactive ones. Higher-hypnotizability users improved more. Only ten of the 84K users reported worsening symptoms — a strong safety signal.

The paper is the first attempt to measure what self-hypnosis actually does at scale, and it’s by far the most data anyone has on the question.

Press release headline: Stanford-developed app reduces stress in 84,000 users. To the normie science news reader, that sounds pretty good.

The skeptic reads the methods section

It isn’t. The study has no control group. It’s a retrospective look at people who chose to open a hypnosis app, rated their stress on a single quiz item before a 10-minute session, then rated it again after. How could a design like that not show improvement?

1. Regression to the mean. People open the app when they’re stressed. Their next rating will tend to be lower regardless of what happened in between.

2. Demand characteristics. You just spent ten minutes on something marketed as a stress reducer. Reporting “still stressed” feels like grading yourself.

3. The pause itself. Ten minutes sitting still with eyes closed reduces self-reported stress whether the audio is hypnosis, an audiobook, or silence.

The authors concede in their own Limitation #4 that they used a single stress test (Likert) measure rather than a validated scale like PSS or PANAS, chosen for engagement rather than rigor. I don’t necessarily mind these sorts of pragmatic decisions made in order to get as large a sample size as possible. But all of these caveats make the conclusions far more shaky than the headline suggests

And that’s not to mention the obvious conflict of interest, and motivated reasoning bias when the study’s author is the same guy who sells the product. Spiegel has spent 50 years on this work and his clinical record is real, and I appreciate any academic who tries to put his work into mainstream use as a real product, but put it all together and a more honest framing is: the study shows something happens after a Reveri session. It cannot show whether that something is hypnosis.

Large N doesn’t fix lack of control

This is the personal-science lesson worth pinning to the wall. A sample of 84k doesn’t compensate for the absence of an active comparator. Big N makes random noise vanish; it does nothing for systematic bias. If every one of those users had also rated their stress before and after ten minutes of anything else — a podcast, silence, scrolling Instagram — the comparison would be informative. Without it, the number is rhetorical, not epistemic.

This is exactly the trap personal scientists fall into when reading observational health data. We notice a sample size, we notice a p value, we move on. The question that decides whether the result means anything is almost always: compared to what?

The trait may not be a trait

The other genuinely new result, which got less attention: David Spiegel spent decades arguing hypnotizability is a stable, lifelong trait — measurable with the eye-roll test and that ~2/3 of adults somewhat susceptible, ~15% highly so. In January 2024, Spiegel co-authored with fellow Stanford researchers Afik Faerman and colleagues a preregistered, double-blind RCT in Nature Mental Health showing that a minute of transcranial magnetic stimulation (TMS) temporarily raises hypnotizability in low-scoring patients. The effect lasted about an hour. n = 80, sham-controlled, neuroimaging-guided. In other words, hypnotizability can be changed.

Two caveats keep this from being world-shaking: the trial measured the hypnotizability score itself, not clinical outcomes, and the effect was transient. But the implication for personal science is real: hypnotizability looks more like trainable VO2 max than like eye color. Predominantly stable, modifiable at the margins. If you take the eye-roll test and score low, that’s the current reading — not a verdict.

It also opens the testable question of whether hypnotizability, “absorption,” NSDR responsiveness (see PSWeek230314), and placebo response are facets of one underlying trait. They might be.

The n-of-1 anyone can run

Reveri still costs ~$99/year, with a free 7-day trial. The eye-roll test is free on YouTube. So is sitting quietly in your bed.

Here’s a one-week protocol any reader can run:

• Day 1, 3, 5: rate stress 1–10, do a 10-minute Reveri session, rate stress again.

• Day 2, 4, 6: rate stress 1–10, sit quietly for 10 minutes with eyes closed (no audio, no app), rate stress again.

• Day 7: look at your numbers.

If Reveri beats silence, that’s evidence of something specific to the intervention. If they tie, the active ingredient was the ten minutes, not the hypnosis. Either result is useful. Neither requires a subscription.

This is the personal-scientist’s answer to a study that can’t answer its own question. An academic paper tells you the app is safe and that users feel better after using it. Whether you should pay for it depends on a comparison only you can run.

Personal Science Weekly Readings

Speaking of hypnosis, a 2024 systematic review of 679 hypnosis apps found only four had ever been tested in a clinical efficacy trial — Reveri wasn’t among them. So the new 2025 paper mentioned above is also Reveri’s first appearance in the literature, which makes the design choices it made all the more consequential.

At PSWeek, we’ve had a lot to say about plastics over the years (e.g. PSWeek250102), so we should revisit some of our assumptions after a new University of Michigan study points out that researcher gloves can contaminate the lab tests that claim high microplastic levels.

About Personal Science

Personal scientists are open-minded but skeptical — about apps marketed with big numbers, about founder-authored outcome papers, and about our own confident first reactions. Nullius in verba — take no one’s word for it. Especially when the word comes wrapped in n = 84,395.

This newsletter is delivered each Thursday. If you have topics you’d like to see covered — or experiments you’re running that deserve attention — let us know.

Leave a comment

What magnesium actually does

Open any biochemistry textbook and the picture of magnesium has almost nothing to do with sleep. It’s the second most abundant positive ion inside your cells, and it’s a required helper for more than 300 different enzyme reactions. The places it actually earns its keep cluster in three buckets:

• Heart and blood vessels. Magnesium is your body’s natural calcium blocker — it keeps calcium from flooding into muscle cells, which is how it relaxes blood vessels and steadies the heartbeat. Cardiologists have used IV magnesium for decades to treat specific arrhythmias. This is the part of magnesium with real evidence behind it.

• Nerves and muscles. Magnesium sits at the gate of the NMDA receptor (a key brain signaling channel) and helps regulate how nerves talk to muscles. Run too low and you get tremors, twitchy reflexes, and — in serious cases — seizures. This is the pathway the bedtime-supplement story is supposed to work through.

• Energy production. Every ATP molecule — the universal currency your cells spend to do anything — only works when it’s bound to magnesium. No magnesium, no functioning metabolism. Full stop.

Notice what’s not on the list: sleep.

The sleep claim is downstream of the neuromuscular pathway, and on the actual evidence it’s much shakier than the marketing implies. The most-cited meta-analysis (Mah & Pitre 2021) pooled just 3 RCTs and 151 older adults, found a modest improvement in sleep onset latency, and concluded — in the authors’ own words — that “the quality of literature is substandard for physicians to make well-informed recommendations on usage of oral magnesium for older adults with insomnia.” A larger 2025 randomized trial of magnesium bisglycinate (155 adults, 4 weeks, the same form I take) found a real but small effect on insomnia symptoms — Cohen’s d = 0.2, a barely-detectable nudge.

In other words, the textbook biology of magnesium is mostly heart-and-nerves. The sleep angle is a marketing emphasis, not a scientific consensus.

My experiment

To find out which story applied to me, I ran a withdrawal-and-restart test on myself. After 35 nights on Mg bisglycinate 240 mg, I stopped cold for 4 nights, then restarted for 7 more.

Important note on dosage: 240 mg/day is below the doses used in the cited RCTs (typically 300–500 mg of elemental Mg), and below what Huberman and other influencers recommend for sleep. My total daily intake — diet plus supplement — was probably in the normal-to-slightly-above range for an adult male. So what I’m testing is not “does maximal magnesium loading affect sleep” but the more practical question: does the dose I actually take, layered onto an unremarkable diet, do anything I can detect?

The key to this kind of n-of-1 design is holding everything else constant. Across all the test nights — and pre-registering the list before withdrawal began — I committed to: no alcohol, dinner before 7 PM, protein at least 100 g/day, no high-intensity (Zone 3/4) cardio, consistent bedtime. Each morning I took a 3-minute Polar H10 reading (see PSWeek260409) and pulled overnight metrics from Apple Watch. The primary endpoints, fixed in advance, were resting heart rate and N3 deep sleep. Secondary endpoints were WASO (wake after sleep onset) and overnight HRV.

Here’s the summary of what I learned:

The headline result on sleep was straightforward: nothing. Deep sleep didn’t drop during withdrawal. Sleep efficiency didn’t change. WASO actually got better — the two cleanest, lowest-WASO nights of the entire dataset (6 minutes and 5 minutes) happened with zero magnesium in my system, the opposite of the predicted direction. By my own personal-science rule of thumb — if something works, you’ll notice — magnesium failed the sleep test.

What did show up

A different signal appeared on a slower timescale.

By the third clean withdrawal night, my overnight HRV (LnRMSSD) had dropped to 3.78, about 1.35 standard deviations below the Mg-ON baseline of 4.05 — the first overnight HRV depression in the test. When I restarted, recovery wasn’t immediate either. One low night and a recovery curve isn't proof — it's a hypothesis I'll test with a second withdrawal cycle. But the kinetic profile is suggestive.

That kind of result — a clear dip on day 3, recovery taking 4+ days — is wrong for a same-night NMDA/GABA effect, which is what the bedtime-supplement story implies. It is exactly right for a tissue-redistribution effect on vascular smooth muscle and parasympathetic tone, which is what cardiology has known about magnesium for decades.

That a 240 mg dose — below the loading doses in the literature — produced any autonomic signal at all is what makes it worth a second look. Small intervention, detectable response, suggests a sensitive pathway.

So magnesium does seem to be doing something for me. Just not on the metric I was hoping for. The effect is modest, slow, and on the autonomic axis — not the sleep-architecture axis.

The reframe

The cleanest way to describe what 46 nights of careful tracking taught me is this: magnesium is a heart-and-nerves supplement, not a sleep supplement. The mechanism with decades of cardiology evidence behind it — vasodilation, calcium antagonism, autonomic tone — is the one that shows up in my data. The mechanism with three underpowered RCTs behind it — sleep architecture — is not.

This isn’t a takedown of magnesium. As a low-cost cardiovascular adjunct with a defensible mechanism and a small but real autonomic signal in my own data, I’ll keep it up. But I’m taking it for the right reason now, not the influencer reason.

My remaining sleep lever is somewhere else. Last month I wrote about protein turning out to be a bigger deep-sleep variable than I expected. The next experiment, hinted at by an unexpected pattern in the last few restart nights, is whether when I eat protein matters as much as how much. More on that soon.

Personal Science Weekly Readings

• The Mah & Pitre meta-analysis on oral magnesium for insomnia is the most-cited piece of evidence behind the popular claim. Worth reading both for what it found (a 17-minute reduction in sleep onset latency) and for what the authors say about the evidence base they were working with (substandard).

• The 2025 magnesium bisglycinate RCT is the largest trial to date on the form most people actually buy, and the most honest about effect size. Cohen’s d = 0.2 is the kind of effect you can find with a big enough sample but probably can’t feel.

• For contrast, the cardiovascular evidence on magnesium is substantially stronger. For example a 2024 review covers the calcium-antagonist mechanism, vascular tone effects, and arrhythmia data — the parts of the magnesium story that are actually well-supported.

• Previous PSWeek experiments where the popular hypothesis didn’t survive contact with my data: CBD for sleep (no effect), my 3 AM wakeup investigation (still unsolved), and the protein/N3 surprise (a real lever, found by accident).

New York New York

Note: I’ll be in New York City from now until late June, meeting a number of fellow personal scientists. If you’re in town, let me know!

About Personal Science

Personal scientists test the claims that the rest of the world repeats. We track not because the numbers are perfect, but because tracking forces the attention that lets us see when the consensus story doesn’t fit our data. Nullius in verba — take no one’s word for it — is an old motto that applies as well to a supplement label as to anything else.

We publish every Thursday. If you’ve run your own n-of-1 on magnesium and found something different, let us know.

Leave a comment

This week we look at the problem of unspoken knowledge and what to do about it.

A U-Penn student once told me about volunteering with under-privileged mothers in Philadelphia. One mother explained she’d be stopping at the store on her way home to pick up dinner for her kids: a box of Oreos. Just Oreos. Nothing else.

Most of us hear this and assume it’s a joke. What mother would think that’s a normal meal? But the student insisted it was typical, and most of the mothers she worked with thought it was fine.

Here’s the uncomfortable question for a personal scientist: how do you know it isn’t? Where in your upbringing did you learn that Oreos aren’t dinner? Probably not from a 3rd-grade health class. Most likely you absorbed it the way humans have absorbed nutrition knowledge for millennia — from parents, siblings, peers, the smell of your grandmother’s kitchen. Nobody ever taught you “Don’t eat Oreos for dinner,” just like I’m sure that Philadelphia mother was never taught the opposite.

This is what Hungarian-British philosopher and physical chemist Michael Polanyi called tacit knowledge — the know-how that resists being written down. His slogan: “we can know more than we can tell.” His stronger claim, easy to miss: all knowledge sits on a tacit foundation. The explicit propositional stuff — textbooks, peer-reviewed papers, your favorite Substack — is the visible tip of an iceberg whose bulk is unarticulable.

Now invert the question. How do you know your way of eating is actually better? You could read a stack of nutrition books, become an “expert.” But here Polanyi would push back: book knowledge laid on top of a tacit foundation doesn’t replace the foundation. You picked the books. You picked the authors. Economist Arnold Kling puts it bluntly: we decide what to believe by deciding who to believe. Reading Michael Pollan instead of watching Food Network doesn’t escape that loop — it just dresses it up with better references.

This is the trap personal science is supposed to be a discipline against.

Back in PSWeek221229 we cited philosopher of science Michael Strevens: science is data + methods of data collection + logic that ties them together. That’s the only known procedure for stepping outside your inherited intuitions and your trusted sources and forcing yourself to face what’s actually true about you.

I want to update that older claim with what I’ve come to think since. Strevens isn’t an escape from Polanyi — nothing is. The tacit substrate is always there, in your choice of what to measure, your choice of which expert seems serious, even your sense that a result “feels right.” Personal science is not a way to transcend the tacit layer. It’s a discipline for refusing to be its prisoner.

That means asking, every time you think you “know” something: did I learn this by running the data + methods + logic gauntlet — or did I just pick it up? Most of what most of us “know,” even those of us who read a lot of papers, falls in the second bucket. That’s not shameful. It’s the human condition. But noticing it is the first move.

Including, of course, with this newsletter. Don’t take my word for any of it. Nullius in verba doesn’t go all the way down — Polanyi would say nothing does — but it goes further than anything else we’ve found.

Personal Science Weekly Readings

Speaking of tacit knowledge, nobody disputes that sunburn is bad for you. And common sense says it’s good to spend time outdoors. But how much sun is “good”? Matthew Zirwas is a dermatologist who wants to convince you that avoiding sun exposure is more likely to lead to problems. I double-checked the list of academic papers he cites, and it’s true that the underlying observational studies are real, though as always with this kind of epidemiology, sun exposure correlates with a lot of other healthy behaviors." He also claims that while Vitamin D is important, getting it from supplements may not work.

One way to tell how beholden you are to tacit knowledge is how easy it is to change your mind. Despite very high LDL-C numbers Nick Norwitz MD PhD was skeptical of cholesterol-lowing medication until recently he changed his mind. His thought process is classic personal science: be skeptical but open-minded, and recognize that everyone is different. (We mentioned him previously in PSWeek251127 as the guy who tried an all-sardine diet for a month.)

Speaking of Nick Norwitz, we’ve long been fans of Dave Feldman, a normal guy who has promoted the idea of “Lean Mass Hyper Responders” (LMHR), a specific class of people whose cholesterol shoots up on a ketogenic diet. Feldman’s “Cholesterol Code” concludes that high cholesterol isn’t relevant to LMHR people. So when recently Peter Attia strongly panned the LMHR idea, it’s worth noting both Attia’s case and the response from Feldman and Norwitz. Tldr; Feldman and Norwitz argue that Attia is misrepresenting the LMHR theory and cherry-picking results.

Oh, and speaking of Oreos, Nick is also the guy who published that study we wrote about in PSWeek240125 that showed how for LMHR people, Oreos lower cholesterol more than a statin.

Astral Codex Ten is always worth a read, especially the monthly list of interesting links. The links from April included

An entrepreneur’s dog got cancer, so he worked with ChatGPT to design a personalized mRNA vaccine, and it seems to have kind of helped (the dog still has cancer, but the tumors shrunk and she is feeling better). Here’s a comment with more info by the scientist involved, here’s a reminder that this kind of tumor changes size a lot for no reason; here’s the inevitable prediction market on whether we’ll still believe this is real a year from now (currently at 66%). I am less interested in the fact that one guy says his dog improved than in the comments by seemingly unimpressed scientists saying “Yeah, whatever, big deal, anyone can make a personalized MRNA cancer vaccine that works, the difficulty is studying it and scaling it up” (example).

Finally, I don’t know who is behind the X pseudonomyous account Crémieux, but he often posts interesting content, like a summary of peptides quality that concludes most China-sourced GLP-1RA samples are actually fairly pure. But be careful: there is a lot of variability among vendors. That said, he claims you can get good quality semaglutide for around $15 a month and retatrutide for $30. If you try this yourself let us know.

About Personal Science

Our motto is the 1660 motto of the Royal Society: Nullius in verba — take nobody’s word for it.

That’s harder than it sounds. As this week’s post argues, most of what any of us “know” about how to live well wasn’t acquired through careful experiment — it was absorbed from parents, peers, books, and the experts we happened to trust. Personal science doesn’t pretend to escape that substrate. It’s the discipline of noticing it, and of putting as much as we can through the data + methods + logic gauntlet anyway. Skeptical of experts, skeptical of supplements, skeptical of our own intuitions — and, especially, skeptical of our own conclusions.

Personal Science Week is delivered each Thursday to people who use science for personal, rather than professional reasons. If you have topics you’d like to see covered — or experiments you’re running that deserve attention — let us know.

Leave a comment

Recently I’ve been testing HRV seriously to see what I can learn — and I’ve come away with a narrower, more useful theory than I started with.

On the morning of April 13, my resting heart rate (RHR) was 64—well above my baseline. That’s the kind of number that, on its own, just means something is off. Maybe I’m getting sick. Maybe I overtrained. Maybe I drank too much coffee.

But this time I had more data. I’d put on a Polar H10 chest strap (see PSWeek 260409) and run a 17-minute HRV recording through my HRV app. I got the autonomic signature of psychological stress, not illness, not overtraining.

Why was I stressed? Because I hadn’t started my taxes yet. (Personal scientists are not immune to procrastination.)

That morning is the cleanest example I have of why HRV is worth measuring. Resting heart rate alone could not have told me which kind of “off” I was in—and the right response is different for each.

What HRV is good for

After several weeks of exploring, my conclusion:

RHR is best for overall readiness scoring ; HRV helps figure out why something is off.

RHR is one of the easiest metrics to capture. Any activity wearable will give it to you (and you can even measure it the old fashioned way with a thumb on your wrist and a minute timer). HRV (heart rate variability) is much trickier to capture accurately—and most importantly—in a way that’s actionable. Spoiler alert: your Apple Watch or Oura measure of HRV are not particularly actionable, I’ve found.

The instructive cases in my data are when HRV and RHR disagree. When they agree, RHR is sufficient and HRV is redundant. When they disagree, HRV usually tells you something specific about which physiological state you’re in.

For daily decisions—did I recover, am I overreaching, should I train hard today—RHR captures something like 70–80% of what HRV tells you. It’s cheaper, less noisy, less artifact-prone. The 20–30% where HRV adds something is when it’s worth the extra trouble.

A quick HRV primer

If you’re new to HRV: it’s the tiny variation in milliseconds between consecutive heartbeats, and it’s a window into your autonomic nervous system—the involuntary side that regulates everything from digestion to stress response. Your heart’s pacemaker is designed to fire at a steady rhythm—about 100 bpm without input. The autonomic nervous system rides that pacemaker like a driver: the parasympathetic branch acts as a brake (slowing things down at rest), the sympathetic branch as an accelerator. HRV is a measure of how actively that driver is steering—how quickly your body shifts between the two.

Higher variability generally means a more responsive parasympathetic (”rest-and-digest”) branch; lower variability means the sympathetic (”fight-or-flight”) branch is dominating.

Apple Watch will report an HRV number (SDNN, sampled overnight), but it samples sporadically and the absolute values are noisy. Oura is better, but it’s still only every 5 minutes. For serious and actionable metrics, you’ll want a chest strap.

1. Am I getting sick? (24–48 hours of warning)

This is the case that justifies the overhead for most people, including me.

A controlled influenza H3N2 challenge study (Temple et al., 2023) inoculated 20 healthy adults with flu virus and monitored HRV, heart rate, and activity via a wearable for 7 days before and 10 days after. Their anomaly-detection model correctly flagged 16 of 17 infected participants an average of 23 hours before symptom onset—and 58 hours post-inoculation, well inside the window where rest and hydration actually change the trajectory.

My own data has a blurrier version of this. In March I caught a rhinovirus, and my HRV dropped before I consciously registered anything was off (see PSWeek260402). But I had a similar drop during my Chicago trip earlier this month—cross-country flight, late-night wedding reception (congrats Leah and Keegan!), a glass of wine here and there. Both recovered, but HRV took about four days to return to baseline while RHR was back within two. Nothing viral; just a clean "something is off" signal — visible in both channels.

The practical version: if Watch HRV drops more than 1.5 SD below my baseline for two consecutive nights and RHR is up 5+ bpm, I flag it as “possible incubation” and adjust the day accordingly—hydrate, cancel the hard workout, reschedule travel if I can. Sometimes it’s nothing. Sometimes it’s something.

2. Is this stress, or fatigue, or something else?

Back to the April 13 morning. Elevated RHR has a half-dozen possible causes, and HRV gives the elevation a flavor. Here’s the rough differential that seems to work:

RHR alone can’t make these distinctions. HRV can. What you do about each is different—the stress case calls for a walk and a phone-free hour, the overtraining case calls for a deload week, the illness case calls for bed. Getting the category right matters.

What HRV doesn’t do

Being honest about what you’re not getting is half of personal science.

• It doesn’t predict sleep quality. My best night this month and my worst night had similar morning HRV. What predicts how rested I feel is RHR + N3 duration + what stage I wake up in. Not HRV. (See PSWeek260416 on the protein-sleep relationship.)

• It’s not a fitness benchmark. Resting HRV correlates only weakly with VO2max cross-sectionally. Useful for tracking your own change over months; useless for comparing yourself to anyone else.

• It’s not a mortality oracle. Yes, population-level HRV correlates with cardiovascular outcomes. The effect sizes are small enough that no single personal decision turns on it.

My current HRV toolkit

Here’s what I’m using now:

Free and good enough: Apple Watch overnight RHR and its HRV measure (SDNN). Catches the big perturbations—Chicago-trip-level stuff, illness incubation, holiday weekends. Fine for most people. I wear my Watch to bed anyway, so this is essentially zero marginal effort.

Cheap and much better: Polar H10 chest strap (~$90) plus the free Kubios HRV app. Two minutes of setup each morning. ECG-grade beat detection. Every frequency-domain metric you’d want—and this is what enabled the April 13 stress-vs-illness call.

If you want to go deeper: pair the H10 with the HRV Logger app by Marco Altini ($15). It exports raw RR intervals to SQLite and CSV, so you can run your own DFA α1, Poincaré, and time-domain analyses. I export mine into a longitudinal log and have Claude pattern-match across mornings.

I don’t use an Oura ring. At $400 plus a monthly subscription, it’s expensive for something whose main job—telling you how you slept—you mostly already know by how you feel. The Apple Watch + H10 combo gives me more data for less money.

For more information about personal HRV monitoring, the best source is Marco Altini’s HRV4Training blog. The creator of HRV Logger writes clearly and honestly about what HRV can and can’t do, mostly for high performance-seeking athletes but also for the rest of us.

Personal Science Weekly Readings

Speaking of low cost new ways to quantify yourself:

• Is it really possible to use an off-the-shelf iPhone as an ultrasound camera? Apparently, yes—researchers have built an app that turns the phone’s existing hardware into a fetal heart monitor.

• Seth Howe is a basically normal guy who sequenced his DNA at home using a $3,200 MinION Mk1D sequencer and about $1,100 of reagents. He also needed access to a centrifuge, heat block, and pipettes—items found in any biology lab. (Easier alternatives: a 30x sequence from sequencing.com for ~$400, or The Odin’s $599 deal if you want to keep your raw data private.)

  

• A 2023 study concluded that more than three-quarters of a million Americans are permanently disabled or die each year as the result of a misdiagnosis—about 5% of Americans experience a wrong diagnosis annually. A useful reminder of why personal scientists treat clinical conclusions with the same skepticism we apply to supplements.

About Personal Science

Personal scientists approach wearable data the way we approach everything: with open-minded skepticism and a willingness to experiment. We don’t assume the marketing claims are right, and we don’t assume they’re wrong. We track, we compare, we update our models. Nullius in verba.

We publish every Thursday. If you’ve found uses for HRV data that RHR doesn’t give you, let us know.

Leave a comment