Week of October 28, 2024
Last week, a husband and wife I see in the office (often together) regaled me with anecdotes of high-intensity, low-jackpot bingo, their favorite social activity. Both are octogenarians who simultaneously play 16 to 24 bingo cards per round. While the caller's cadence varies, my patients assured me they have no trouble with their search algorithm and ambidextrous daubing. The wife reported secretly praying for her husband to win, "He is super competitive (although calmer at age 84)." But, she said, "losing raises his blood pressure." While I've advised patients to stop watching cable news and extricate themselves from toxic relationships, I did not appreciate the medical implications for older, type-A bingo players who (like Ricky Bobby in Talladega Nights) feel that 'If you're not first, you're last.'
There is a paucity of formal research on the bingo's health outcomes; however, there is abundant published mahjong-related health data. For instance, mahjong is associated with improved cognitive and psychological function among elderly Asians in China:
https://pmc.ncbi.nlm.nih.gov/articles/PMC11436455/
I also found an article reviewing 3-case reports about adult men (age 30 to 50) who experienced new-onset seizures while playing mahjong (though the causal relationship is weak - I suspect they played a lot of mahjong and happened to have medical events while doing so.)
https://pmc.ncbi.nlm.nih.gov/articles/PMC10183290/
-----
Here is the Google NotebookLM A.I.-generated podcast with two "hosts" discussing this week's newsletter. Check out What Adam is Reading, the unsettlingly good (and creepy) A.I. Podcast edition. (The tools are improving to the point where the hosts are now obsequiously flattering me and my writing – there will be a world of self-esteem boosting A.I. for sure.)
https://drive.google.com/file/d/1vp5Ny-ckF4vpt0WgtnbrTKzGtlkMaFpa/view
About NotebookLM: https://blog.google/technology/ai/notebooklm-audio-overviews/
---
We continue to be in a nadir of COVID cases (based on wastewater), with 1 in 115 individuals infected over the last 10-14 days.
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/
------
COVID articles
Machine learning algorithms can diagnose COVID based on the sound of an individual's cough. According to these data, the trained neural network offers "a specificity of 98%, sensitivity of 97%, accuracy of 98%, and an AUC score of 0.981. "The researchers presented a comprehensive overview of existing ML-driven audio diagnosis tools and compared their multi-layer audio analysis. Aside from the obvious questions (How will this perform in the real world? Can it be incorporated into smartphones? What about ambient listening devices?), my mind wonders how well other illnesses (or even psychological states) could be discerned from audio and added to real-time interactions (imagine your automated Trader Joe's checkout station attempting to empathetically offer support for sounding sad or depressed and guiding you to the ice cream).
https://www.nature.com/articles/s41598-024-76639-9
And some commentary
https://x.com/atranscendedman/status/1849808730663096825
I found this meta-analysis looking at the COVID transmission rates on airplanes. These data imply a significant opportunity for more air exchanges and a value to masking on airplanes.
https://x.com/ejustin46/status/1849118601334157786
https://www.medrxiv.org/content/10.1101/2024.10.22.24315911v1
Medical Trends and Technology- Halloween edition
These choices were inspired by a loyal reader who became a poorly paid research assistant this week. "Adam, you should write about some Halloween-themed articles." "OK, send me some."
Let's start with the science of ghost hunting. Say what you will about the media, but The NY Times did a fantastic job highlighting the pseudoscience of electromagnetic field detectors, REM pods, and spirit boxes (all of which you can buy on Amazon). Of course, given the absence of standard measures or quantities of paranormal activity, I never understood how you could calibrate these devices.
https://www.nytimes.com/wirecutter/reviews/testing-ghost-hunting-gear/
The University of Virginia's Department of Perceptual Studies has taken a more academic approach to the supernatural. Amongst the researchers, I suggest poking around Department Chair Jim Tucker's website and his books on children that remember past lives. Dr. Tucker offers a far more rigorous attempt to describe unusual phenomena. I found the stories unsettling and fascinating (see the YouTube video).
https://med.virginia.edu/perceptual-studies/publications/books-by-dops-faculty/study-of-near-death-experiences/
Website
https://www.jimbtucker.com
Here is a 2023 talk Dr. Tucker gave in India at the Library of Tibetan Works and Archives.
https://youtu.be/afJX-gizJtc?si=y9xxYGin6Guxbyll
I offer this as an adjacent story: Max Alexander, an 8-year-old fashion designer who claims to be the reincarnation of Gucci. Either way, he has quite a bit of innate talent.
https://www.nytimes.com/2024/09/10/style/max-alexander-nyfw-fashion-design.html
Infographics
The chemistry of pumpkins. Spoiler: pumpkin spice has nothing to do with actual pumpkins.
https://x.com/compoundchem/status/1850168002324681214
Things I learned this week
Thanks to time with our younger son at his college's parents' weekend, I learned about Wigglesworth Dole, the patriarch of Doles who shaped Hawaii's history and started the Dole Fruit Company. I suspect the elder Dole bore the name Wigglesworth with disproportionate dignity in the 1790s. (College seminar materials are a fantastic source of ephemera.)
https://en.wikipedia.org/wiki/Wigglesworth_Dole
Also, thanks to my son's college parents' weekend, I learned about the world of large-scale inflatable body organs. After walking through a giant inflatable colon (designed to raise awareness for colorectal cancer screening), a quick interweb search yielded inflatable organs for sale or rent. My wife knows how deeply I believe our 12-foot Home Depot skeleton needs a set of inflatable kidneys.
https://www.medicalinflatables.com/exhibits/mega-body/
And for those still looking for a Halloween costume, it is not too late to get your trachea and giant lung costume:
https://www.medicalinflatables.com/exhibits/mega-lungs-costume/
I learned about the genetically isolated Amazonia bird, the Hoatzin. It is closer to a flying dinosaur than any other living creature, and it smells terrible due to some unique physiology for a bird (it ferments plants in a large foregut as part of its digestive process, like a cow). I don't recall the Jurassic Park movies commenting on the smell of the dinosaurs (though there were some dinosaur poop scenes, as I recall).
https://www.aquaexpeditions.com/blog/hoatzin-bird-amazon
Living with A.I.
The Atlantic published an article exploring how college students and professors navigate a world increasingly filled with generative A.I. tools. The author is an English professor who returned to his undergraduate alma mater, Haverford College, which is known for its honor code. While I am not sure I agree with the author's conclusions (underprivileged students and people are more likely to rely on tools like ChatGPT), the article offers some interesting insights and a "snapshot" of college kid’s thinking amid a lot of change. [For full disclosure, my older son is currently enrolled in Haverford and felt the article anchored too heavily on the power and value of Honor Codes. He contends the issue is where in the writing process generative A.I. is employed (research, ideation, writing, etc.) is critical and under-appreciated by this article.]
https://www.theatlantic.com/ideas/archive/2024/10/chatgpt-vs-university-honor-code/680336/
A new early developer version of Anthropic's Claude (that can recognize and interact with objects on a computer screen - independent of a human) is another step toward the world of A.I. agents - A.I. tools trained for specific tasks. When working in series, agents could complete complex multi-step processes on a computer or in the real world via robotics. A loyal reader asked how such agents may impact human work rules and regulations. (In other words, what does it mean if you have tasks completed 24/7 by autonomous agents but require a human in the loop? I will not address the "what happens when the agents demand rights" question.) A bit of searching revealed interesting (but biased) discussions of how labor unions miss the opportunity to embrace A.I. early. [N.B. - the editorial is reasonable, but the organization is a logistics data and consulting firm for companies that employ unionized workers.]
Discussion of Claude:
https://www.oneusefulthing.org/p/when-you-give-a-claude-a-mouse
And
Editorial on unions not thinking about how to embrace A.I.:
https://zero100.com/unions-are-missing-their-ai-moment/
A.I. art of the week (A visual mashup of topics from the newsletter).
Ancient cave art painting of ghosts playing bingo surrounded by pineapples, a giant inflatable colon, a Hoatzin, and two computers with arms moving each other's keyboard and mouse. One of the ghosts is coughing, with signs of air coming out of the ghost's mouth.
https://drive.google.com/file/d/1mVjCHal1y62ug4sjYP2lFJCHvv6WdLd7/view
Clean hands and sharp minds,
Adam
Comments
Post a Comment