Week of April 28, 2025
Twenty-seven years ago, my patient's camel bolted from a live Christmas nativity scene onto the highway (that is outside my medical office) on Kent Island, Maryland. Sadly, Ernie, the camel, was struck and killed by a Volvo driven by the parents of a local college student. Sometimes, good healthcare is making the correct diagnosis. Sometimes, care is about giving patients time and space to explain life's wins, losses, and absurdities. And while not all of my patients own a menagerie of exotic animals (for tax purposes - read the article below), I delight in earning people's trust - and allowing them to share their diverse experiences. I am also amazed that Volvo did not use this story to advertise their vehicle's safety (sturdy enough to save you during high-speed dromedary impact!)
https://wapo.st/4lMzd8j
and here is a .pdf in case you don't want to log in to WaPo for the free article:
https://drive.google.com/file/d/1QOIwsw5tQWaFszkALgrLefX_bbJ_DY7H/view
Note that my patient approved of me writing this. (How could I not share a story like this? It practically writes itself.)
---
Listen to a Google Notebook LM A.I.-generated podcast of the newsletter with two virtual "hosts."
https://drive.google.com/file/d/1uI1wZm1s2H-P9sOxGsqf0IArdCdhz9aj/view
About NotebookLM: https://blog.google/technology/ai/notebooklm-audio-overviews/
------
Science and Technology Trends
Dr. Eric Topol published an excellent analysis of the wide range of data about alcohol consumption. He highlights this topic's immense difficulty: How do you define intake (measured in precise amounts (often not known) or categories like "moderate," which are subjective)? How do you control for confounding variables? And, what conclusions and recommendations does the best available data suggest? Topol's interview with former Surgeon General Vivek Murthy provides an excellent discussion on causation, bias, and the difficulties of applying population-level data to individual recommendations.
https://open.substack.com/pub/erictopol/p/a-ground-truth-on-alcohol-intake
Anti- Anti-Science Articles of Note
The Kaiser Family Foundation published a new poll on the US public's views on measles outbreaks and misinformation. Despite a partisan divide on vaccine beliefs, increasing exposure to scientific misinformation, and uncertainty about how to interpret the "debate" about vaccines [there is a sizable 'movable middle' of opinion], there continues to be a high degree of confidence in the MMR vaccine's value.
The poll
https://www.kff.org/health-information-and-trust/poll-finding/kff-tracking-poll-on-health-information-and-trust-the-publics-views-on-measles-outbreaks-and-misinformation/
An AI Summary of the polling data:
https://claude.ai/public/artifacts/ca5d1148-daee-40db-97a0-33d005a7616c
Anti-science can be very subtle. One of my patients (who falls in the skeptical but 'movable middle' referenced above) sent me a news article (about research published in a science journal) in response to my recommendation he should begin semaglutide (aka Ozempic): "New study raises concerns over serious side effects of weight loss injections."
https://www.diabetes.co.uk/news/2025/jan/new-study-raises-concerns-over-side-effects-of-weight-loss-injections.html
A few interesting points:
- A few weeks ago, I reviewed the journal article this news article cites. The journal article was generally positive and affirmed that the benefits were greater than the risks for GLP1 use.
- The news article my patient shared was published on a UK diabetes website, but the article, citing the same research, is negative regarding GLP1s.
- It is an excellent example of how a good study with valuable data can be 'weaponized' into scary and negative news through editorial and cognitive biases.
Here is what I wrote back to the patient via email:
-
Thanks for this web article. It is a fantastic example of bad reporting. Here are the major flaws I see.
- Alarming Statistics Without Context: The article mentions a "46% increase in hospitalizations in just the last month" without providing baseline numbers or context for this percentage, exaggerating the risk perception.
- Sensationalist Language: The article uses terms like "potentially life-threatening" and emphasizes severe complications, which may create disproportionate fear relative to the actual risks. Moreover, the article highlights the well-known side effects of GLP1s (nausea, constipation) and well-known problems in patients with weight loss (like not adjusting one's blood pressure medicine and having low blood pressure readings).
- Incomplete Presentation of Benefits vs. Risks: While mentioning that GLP1s "have several beneficial effects, including a reduced risk of heart attack, stroke, and even dementia," the article places much more emphasis on risks and adverse effects - a false equivalency with the known positive data.
- Cherry-Picking Data: The article highlights adverse outcomes while giving less attention to the broader context of these medications' established benefits for diabetes management and cardiovascular risk reduction.
There were some positive points to the UK.diabetes.com article:
- The article quotes the study's lead author, Dr. Ziyad Al-Aly, who acknowledges GLP1s have some side effects.
- The article mentions that GLP1s are associated with a reduced risk of heart attack, stroke, and dementia.
- The article reports on a peer-reviewed study with over two million patients.
The cited research stud:y https://www.nature.com/articles/s41591-024-03412-w
Overall, the article you share is intellectually flawed and more designed to stoke fear (and redistribute it on social media)—it is clickbait.
Thanks for sharing!
-Adam
-
-------
Living with AI.
The AI Daily Brief and Gizmodo recently published the story of Roy Lee, a Columbia University student, who developed an AI tool called "Interview Coder" that could solve coding problems presented during big tech firms' live technical interviews. Several companies hired Lee (he declined the jobs), and he posted the recording of his Amazon interview on YouTube. Amazon expressed concern to Columbia about Lee "cheating" on their technical interview. Columbia University subsequently held a disciplinary hearing, leading to Lee leaving college. Since then, Lee has rolled his Interview Coder into a new startup called Cluely - and has sparked a fantastic debate on how to think about our use of AI - is it cheating, augmenting, or assisting?
https://open.spotify.com/episode/3YOSJ5kRalWZvX2JPfZHh9
(Note: The podcast calls Lee's app Clearly, but it is Cluely.)
and
https://www.linkedin.com/posts/roy-lee-goat_i-got-kicked-out-of-columbia-for-building-activity-7319827343730974720-f4ew/
and
https://fortune.com/article/cluely-ai-cheating-columbia-student-seed-funding/
Infographics
Kremp Florists, like Alan's Factory Outlet [of sheds], has a robust infographic and web design department. In anticipation of Mother's Day, they offer The Best and Worst States to Be a Mom. Please pay attention to how they define the criteria for ranking.
https://cdn.shopify.com/s/files/1/0912/4927/9256/files/best-states-for-moms-5.png
from
https://www.kremp.com/pages/the-best-and-worst-states-to-be-a-mom
Here is the link to Alan's, my other favorite company with a surprisingly robust infographics department:
https://alansfactoryoutlet.com/infographics-alans-factory-outlet/
Things I learned this week
Similar to how scientists have been imaging the carbonized scrolls from Herculaneum, British scientists have recently used CT scanning and 3D reconstruction to "unfold" and reconstruct medieval manuscripts. One such manuscript included a previously unknown, archaic story of Merlin (the wizard of King Arthur's court). The newly rediscovered version of Merlin's story hints at the derivative Celtic origins to the Arthurian legend - and the role of Christianity in shaping earlier pagan beliefs. I recommend the YouTube video below describing how culture, history, and politics shape our stories and legends. I cannot recall thinking about the broader historical and political dimensions of the King Arthur story, but (to paraphrase the philosophy of second-wave feminism) everything is political.
https://www.bbc.com/future/article/20250324-the-magical-medieval-tale-revealed-by-a-100000-dollar-camera
and
https://youtu.be/KmLr76gdXfc
Related: The mission is to decipher Herculaneum scrolls using high-resolution scanning, and artificial intelligence scales up rapidly.
https://www.nature.com/articles/d41586-025-01087-y
What Adam is Reading follow up - whales are (people?) too. After enough time, it is worth revisiting past topics for updates. In June 2021, I wrote about research efforts to translate whale songs into human-understandable language and, perhaps, communicate back with whales. Checking in with Project Ceti (for cetacean, a play on Project SETI, which searches for extraterrestrial radio signals) yields a steady stream of research and publications. The most recent publication is a not-yet-peer-reviewed article from Ecology Law Quarterly (2025) exploring how advancements in artificial intelligence and bioacoustics could transform nonhuman animal law, with a particular focus on cetaceans. Using Project CETI (Cetacean Translation Initiative) as a primary case study, the authors examine how decoding animal communication might reshape legal frameworks in light of a higher degree of animal sentience than we appreciate.
Project CETI
https://www.projectceti.org/news-research-insights
The Paper:
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5165527
The TLDR AI Summary of the article: https://claude.ai/public/artifacts/682a952e-347f-4f5a-9a98-cde95c1c67d9
My original discovery of bioacoustic translation:
http://www.whatadamisreading.com/2021/06/what-adam-is-reading-week-of-6-14-21.html
The SETI listening to space (as opposed to whales): https://www.seti.org/
AI art of the week.
(A visual mashup of topics from the newsletter, using ChatGPT to summarize the newsletter, suggest prompts, and make the images).
https://drive.google.com/file/d/1MmiXDG8Em1wBxpY3lurrjp5eehFeAqj9/view
Prompt:
Create a retro-style 1990s advertisement poster in a clean, colorful, semi-cartoonish 2D illustration style, similar to vintage travel or automotive ads.
Foreground: A gray Volvo 850 station wagon, slightly dented but intact, parked on a simple roadway. There is a significant dent in the car's hood, shaped like the silhouette of a camel. Inside the car, a smiling family (two adults, one child) is seated, wearing casual 90s-style clothing, looking relieved and proud.
On the right side of the image, a friendly, slightly cartoonish camel stands next to the Volvo, looking content. The camel wears a harness decorated with tiny bells, subtly ringing.
Background: A calm blue ocean with gentle waves, a sailboat in the distance, and an enormous whale leaping out of the water. Above the sea and whale, musical notes are floating, gradually transitioning into binary code (zeroes and ones), representing a blend of nature, music, and technology.
Typography: The bold navy blue retro font at the top reads "VOLVO SAFETY."
At the bottom, also in bold navy blue retro font, it says: "BUILT KENT ISLAND TOUGH: SURVIVES DROMEDARY ENCOUNTERS."
Color Palette: Soft, muted tones (dusty blue, gray, camel tan, cream background), clean linework, and minimal shadows emphasize a nostalgic, feel-good vibe.
Composition: The layout is balanced, with the Volvo and camel in the foreground and the ocean and whale background off to the left. The text is legible and centered. The tone is playful yet heroic.
---
COVID rates were down to 1 in 196 individuals last week.
The Pandemic Mitigation Collaborative (PMC) website uses wastewater levels to forecast 4-week predictions of COVID rates.
https://pmc19.com/data/
based upon https://biobot.io/data/
Wastewater Scan offers a multi-organism wastewater dashboard with an excellent visual display of individual treatment plant-level data.
https://data.wastewaterscan.org/
----
Clean hands and sharp minds,
Adam
-DaVita Inc-
Comments
Post a Comment