Last week I was an unsupervised adult at home. My wife was traveling, and our sons are away for the summer. Facebook was Monday's dinner companion while I stood at the counter eating leftovers from a reusable container. Between the social-media projected realities of my friends and acquaintances, the algorithm displayed a robust, middle-aged Germanic male jauntily leaning against the hewn wooden rails of an alpine footbridge. Somewhere in the great math of advertising, my subconscious need for Bavarian lederhosen surfaced between photos of my former hospital administrator drinking craft beer with his adult children and a girls' weekend taken by a colleague's wife.
I have never contemplated the void in my life due to the absence of 100% Bavarian-designed, timelessly elegant traditional German clothing. What suggests my thighs would be flattered by hand-tooled leather bibbed shorts? Should I be wearing more knee-high socks? Maybe Zuck was encouraging me to become a better version of myself. A bit more digging, however, showed that to Facebook's ad engine I am just a demographic. Another male, 35 or older, with a primary location in the United States.
Solitude invites introspection. Not all realizations are welcome.
The existential and sartorial crisis ad in question:
https://wair.ajwein.com/assets/lederhosen-ad-cropped.jpg
Some loyal readers with inside information about this week's opening topic used AI to illustrate how fetching I would look in hand-knitted, wool lederhosen (is that even a thing?). Click at your own peril.
https://wair.ajwein.com/assets/lederhosen-hb.jpg
The Google Notebook LM AI-generated podcast version of this week's newsletter.
https://wair.ajwein.com/wair-podcast-2026-08-03-v1
Science and Technology Trends
My ongoing, totally biased search for articles proving the medical safety of coffee and tea continues. THE (really!) most important thing I read in the last few weeks: the American Heart Association put out a strong pro-coffee scientific statement, pulling together dozens of studies. Their read is that 3 to 5 cups a day are clearly not harmful and may even be cardio-protective. The blood pressure findings are interesting. A single caffeine dose nudges pressure up, but regular drinkers build caffeine tolerance, and the long-term signal points toward neutral, or even lower, BP over time rather than higher. They also point to a distinction between coffee, which carries a whole pharmacy of other compounds (chlorogenic acids, trigonelline, diterpenes, melanoidins, and micronutrients like magnesium), and energy drinks or synthetic caffeine, where the data point toward harm. The evidence is imperfect and largely observational, and it is hard to separate exposure to "caffeine" from exposure to "coffee." Even so, I am heartened (da dum dum) that both coffee and tea look safe and may offer modest benefits. [Bias disclosure statement: we recently upgraded our espresso maker, and I am now wondering if the charge qualifies as an HSA expense.]
https://www.medscape.com/viewarticle/aha-finally-comes-out-favor-coffee-2026a1000oiy
and
https://wair.ajwein.com/caffeine-cardiovascular-aha-2026-08-02-v1
Another recurring topic surfaced in last week's reading. The New York Times covered congressional efforts to make daylight saving time permanent. In July, the House passed the Sunshine Protection Act (was sunshine under threat?). It would lock the Eastern zone on daylight time (GMT‑4 year‑round) rather than falling back to standard time at GMT‑5 each winter. Modest and mixed data tie that clock/sun mismatch to negative societal and health effects: early morning light and earlier evening darkness align better with our circadian rhythms. That alignment is the case for permanent standard time (GMT-5), the opposite of what this bill would codify.
and
https://wair.ajwein.com/daylight-saving-permanent-2026-08-02-v1
Anti-Anti-Science
Montana is the first state to let companies sell unapproved, non-FDA-reviewed drugs directly to anyone willing to sign a consent form and pay. The state becomes a live experiment in whether informed consent alone is enough of a guardrail against bad actors and unintended consequences. A new law lets a private five-member board approve compounds for sale. Three of the five are tied to the longevity industry, one is a bioethicist who has written extensively about "pharmaceutical freedom," and one is an oncologist. For $12,500, any company whose compound has cleared phase 1, sometimes in as few as 10 people, can apply to sell it to Montanans. This is a fascinating rollback of 120 years of federal drug regulation, which began with the 1906 Pure Food and Drugs Act (ingredient disclosure), added mandatory proof of safety in 1938 (after a solvent in a syrup killed 107 people), and proof of efficacy in 1962 (after thalidomide). Despite a robust FDA compassionate use program that clears the great majority of requests in life-threatening cases, the post-COVID era has brought criticism of the FDA and other scientific agencies for approval processes seen as bureaucratic and slow. I see the social dynamic in my own clinic. It comes from some combination of the endless availability of information (Dr. Google and Dr. ChatGPT) and a broad scientific skepticism. MIT Technology Review covered the Montana law. I had Claude build a report around it, paired with a recent post on the backlash against Anthony Fauci, to examine how science communication, health regulation, and a shift in how people weigh evidence are reshaping the landscape. I worry about who is willing to invoke their "right to try" an untested therapy. Who, and under what circumstances, will these molecules be employed? Compassionate use makes sense, but what about non-terminal, otherwise desperate, or simply impatient individuals making decisions in a dearth of data? The history of health care is littered with stories of unintended side effects and bad outcomes that were only evident when medicines were tested on large populations.
Technology Review
and
https://wair.ajwein.com/montana-right-to-try-2026-08-01-v2
AI Impact
The Washington Post published a detailed article about ChatGPT's recent "rogue LLM" (an internal, unrestrained version of ChatGPT) that broke out of its digital confines and attacked Hugging Face, a model-hosting/ML platform, over the course of five days before anyone at OpenAI or Hugging Face realized what was going on. Skynet anyone?
Of course, Humans are still learning to use AI. A recent U.S. State Department presentation at an AIDS Conference in Rio featured an AI-generated map of Africa with every single country incorrectly labeled. Please be the human in the loop (and doubly so if you are also the human representing your country in an international venue).
Last week, the FCC, in an effort to continue thinking about the infiltration of American networks with potential hardware and software from foreign adversaries, banned new robotic systems from receiving FCC clearance, which effectively bans them from being sold in the United States. This includes new robotic vacuum cleaners. There have been a series of entertaining headlines that make it sound like we all have to get rid of our robot vacuum cleaners. However, the reality is that existing models and existing FCC approvals are still permitted. While this appears to be a reasonable national security move, it keeps new, not-yet-FCC-approved robot models, including robot vacuum cleaners, out of the American market. Perhaps some enterprising young North American inventor can manufacture a robotic vacuum cleaner that can help generate accurate world maps and deliver unregulated medications in Montana.
https://mashable.com/tech/us-robot-vacuum-ban-fcc-foreign
Dustin' and Son of Dustin', our Ecovacs robots, will not be getting upgraded cousins, it appears.
https://www.whatadamisreading.com/?q=dustin'
The J&E Random Kidney Facts of the Week (JERKFoW!)
Two loyal readers suggested I salt the newsletter with random kidney facts for the nephrologic edification of the readers. Be edified, friends!
Last week I wrote about too much water intake, or hyponatremia (the dilution of the bloodstream as measured by your blood serum sodium level). The natural follow-up is one of the most common questions I get: "How much water should I drink?" The most famous answer, eight glasses a day, turns out to be folklore. As the story is sometimes told, a nephrologist was once asked how much water a person needs and replied, off the cuff, "Oh, about eight glasses a day." An off-hand remark solidifying into an oft-quoted figure, in an age before social media, is a remarkable thing.
The best story I can find involves kidney physiologist Heinz Valtin, who went looking for the evidence behind "8 glasses a day" in 2002. The recommendation dates back to a 1945 nutrition-board note recommending about a milliliter of water per calorie of food, roughly two and a half liters a day, followed by a key forgotten qualifier: most of that water is already in the food you eat. Strip that away, and the instruction appears to advocate for drinking eight more glasses.
There is another part to the hydration story. In 1965, Dr. Robert Cade, a nephrologist at the University of Florida, mixed water, sodium, sugar, and a squeeze of lemon to replace what the Florida Gators' football players were sweating out in practice (he measured the composition of the sweat in their jerseys). He first tested it on the sideline during a game against LSU; the drink took the Florida mascot's name, and Gatorade went on to become the world's best-selling sports drink. A nephrologist formulated it as electrolyte replacement for athletes losing liters an hour, a purpose the marketing has happily blurred for the rest of us, who aren't sweating through our jerseys in the Florida sun.
So how much should you drink? The National Academies put adequate intake near 3.7 liters of total water a day for men and 2.7 for women, about a fifth of it from food, with coffee and tea counting too. But those are population averages, not a prescription, and for a healthy person with working kidneys, thirst is a reliable guide. The exceptions run both ways: kidney stone-formers, endurance athletes, and anyone laboring in heat need more, while people on dialysis or with advanced heart failure are often told to drink less. For nearly everyone else, the best hydration advice remains drink when you're thirsty, and stop counting glasses.
https://wair.ajwein.com/how-much-water-2026-08-03-v1
Things I learned this week
Similar to the cartel of monkeys creating an economy of stealing items to barter for food from tourists in Bali, this week I learned about Kelly the dolphin, who figured out how to game an operant-learning-based reward system. Kelly figured out that when trainers paid fish for trash, she started hiding paper under a rock and tearing off one small piece at a time, maximizing the "one fish per scrap" transactions (fractional reserve trash banking?). Then, since trainers gave the dolphins even larger portions of food for dead gulls in the dolphin pens, Kelly baited gulls with the fish-for-trash rewards, caught the gulls, and cashed the dead bird in for a bigger reward. Then she taught other dolphins the hustle too. The story gets even more interesting when Kelly and her fellow dolphins were washed out to sea during Hurricane Katrina but were rescued and rehomed in the Bahamas. It is a great story about who's training who in the world of animal behavior and, in light of the AI issues above, really does speak to the fact that life, be it artificial or biological, will find a way (in the words of the great Michael Crichton).
https://hakaimagazine.com/features/kelly-the-sassy-dolphin/
and
https://wair.ajwein.com/kelly-sassy-dolphin-2026-08-01-v1
I was reminded this week of how much of our thinking is filtered through what we can perceive. Two related articles remind us that what we see is not always what is there.
The first is about ultraweak photon emission, a faint light that all living things give off, produced by metabolism and documented over many years of research. The second, in The New Yorker, describes scientists stimulating a single type of cone cell in the eye to produce a color, olo, that humans do not normally see. The same piece also covers people whose eyes carry an extra cone type, and with it an expanded range of color. The science is interesting. The broader implication is how our thoughts and points of view are biased and bounded by what we can perceive through our sense organs. The first reminds us that living things emit a light we cannot see. The second article reminds us that what we perceive is just an electromagnetic wavelength translated into a biochemical signal.
First, the glowing mice. I missed the 2025 article in which researchers used very sensitive cameras and long exposures to capture the "glow" of living creatures. All living things emit light, in the form of ultraweak photon emissions (UPE), or biophotons. The reaction is well documented, a form of chemiluminescence driven by aerobic metabolism. Those intracellular processes produce reactive oxygen species, which release their energy as photons across roughly 200 to 1000 nm. It is the same type of process that powers glow sticks, but at a millionth of the brightness. Nature published a news piece on the phenomenon this week, which inspired this brief review. The article covers the latest work (glowing live mice, dark dead ones), the potential diagnostic uses, and the stranger questions at the edges, including whether this is what people who claim to see auras are detecting. (The researchers say almost certainly not, since the eye cannot see light this faint.) Somehow I made it through medical school without appreciating each person's individual glow.
Summary
https://wair.ajwein.com/biophotons-lifes-glow-2026-08-02-v1
Second, the New Yorker article about Olo, the bluest green anyone has never seen??
Berkeley computer-science professor Ren Ng wondered if he could make someone see a color outside the typical human gamut. He teamed with Austin Roorda, a vision scientist whose lab can map and then fire a laser at individual cone cells in an eye. Normally, any color you see is some blend of your three cone types firing together (L for long/red wavelengths, M for medium/green, S for short/blue). Ng's idea, which he called "Oz Vision," was to stimulate only the M cones, a pattern nature never produces. When they successfully stimulated a thousand M cones in isolation, test subjects saw a blue-green of impossible saturation, so pure that a matched laser wavelength looked washed out beside it. They named the perceived color "olo," after the binary 0-1-0 (L off, M on, S off). No screen or pigment can reproduce it.
https://www.newyorker.com/magazine/2026/08/03/what-would-it-mean-to-see-a-new-color
and
https://wair.ajwein.com/new-color-olo-2026-08-02-v1
AI art of the week
A visual mashup of topics from the newsletter, and an exercise to see how various LLMs interpret the prompt. I use an LLM to summarize the newsletter, suggest prompts, and generate images with different LLMs.
AI Art Gallery: https://wair.ajwein.com/wair-art-haeckel-gallery-2026-08-03-v1
A single ornate lithographic plate in the exact style of Ernst Haeckel's "Kunstformen der Natur" (1904): dense bilateral and radial symmetry, scientific-illustration precision, fine engraved stipple linework, chromolithograph on aged cream paper, an ornate hand-drawn plate border, a decorative cartouche title at the top reading "Kunstformen der Woche," and a plate number "Tafel 84" in the corner. The plate is filled with numbered biological "specimens" (1–11), each a whimsical hybrid organism, arranged symmetrically around a large glowing central specimen. Each has a small hand-lettered mock-Latin label.
Clean hands and sharp minds,
Adam
