Week of January 27, 2025
Last week, a long-standing patient who received a living-unrelated transplant (before requiring dialysis) returned to my office - beginning the hand-off from his transplant center. A preemptive transplant is rare - the nephrology equivalent of a grand slam home run or winning the triple crown - and I was (evidently) projecting an over-developed sense of accomplishment. By my fourth (or fifth?) comment of amazement and satisfaction, the patient began to mock me. ("Your back must be so sore from all that patting.") Navigating the timing and complexities of a pre-emptive live-donor transplant is challenging, and doing so with a patient over 12+ years leads to a heightened sense of accountability and familiarity. I will not apologize for celebrating like an over-enthusiastic soccer mom.
P.S. The patient approved my writing about his mocking my enthusiasm.
---
Listen to a Google Notebook LM A.I.-generated podcast of the newsletter with two virtual "hosts."
https://drive.google.com/file/d/1n63brFfXW-v3Tc31dVrVN5A7U6GGGE-R/view
This week, the hosts managed to fill 18 minutes. They even discuss how whiskey tastes—Oy vey.
About NotebookLM: https://blog.google/technology/ai/notebooklm-audio-overviews/
------
Science and Technology Trends
Novo Nordisk's next generation of semaglutide, called amycretin, a GLP-1 + amylin analog, has successfully enhanced weight loss in early trials. I suspect these weight-loss drugs (which confer a wide range of benefits with some, but not terrible, known side effects) will continue to have a significant healthcare impact in the coming decade.
https://www.statnews.com/2025/01/24/novo-nordisk-amycretin-injectable-trial-wegovy/
or
https://finance.yahoo.com/news/novo-nordisk-stock-rises-obesity-153327947.html
What is amylin?
https://www.sciencedirect.com/topics/medicine-and-dentistry/amylin
GLP-1 review article from WaPo
https://wapo.st/3CwEZsX
A recent meta-analysis demonstrates some questions about the most commonly used antivirals for influenza. "The analysis included 73 randomized controlled trials published between 1971 and 2023, consisting of 34,332 participants with non-severe influenza, defined as those not requiring hospitalization in the initial 2-day onset of symptoms. The study looked at available influenza antivirals, including those approved by the US Food and Drug Administration." "[The] meta-analysis found that baloxavir (xoflusa) probably reduced the risk of hospital admission for high-risk patients and may reduce symptom duration without increasing adverse events related to treatment in patients with non-severe influenza. Other antiviral drugs [Oseltamivir (Tamiflu), zanamivir (Relenza), and Peramivir (Rapivab)] seem to have little or no effect or uncertain effects on patient-important outcomes." These data argue for a more thoughtful approach to managing flu+ patients, specifically using baloxavir in only higher risk patients (older, immunosuppressed, etc.).
https://www.medscape.com/viewarticle/most-antivirals-have-little-benefit-patients-nonsevere-2025a10001td
and
https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/2829156
Anti- Anti-Science Articles of Note
Dr. Eric Topol wrote a very detailed blog post on the new and existing data associating increased red meat with increased cardiovascular risk, cancers, and dementia. He does a great job highlighting how research is a process, not an end - i.e., how increasingly specific data slowly accumulates over time, starting with observed population-level trends, meta-analyses, and increasingly specific and technical data on possible, quantifiable physiologic mechanisms. In this instance, two new studies have added to our collective understanding of the impact of red meat consumption. One study examined how types of gut bacteria (the microbiome) differ among individuals who consumed different diets (plant, vegetarian, and vegan – the more plant material consumed, the more anti-inflammatory gut bacteria were observed). The other looked at how the amount of red meat consumed correlates with the risk of developing dementia, with less red meat consumption correlating with less likelihood of dementia. Topol's summary of the broader topic and these study’s data is helpful, "The finding of red meat intake association with dementia, brain aging, and subjective cognitive decline adds to the body of evidence for cardiovascular and cancer links. However, it, too, is limited from being based on observational data and not definitive. But there is one thematic message from the new reports. Eating more plant-based foods titrates the risk of red meat intake, at least as reflected by the gut microbiome profile and the lowered risk of adverse clinical outcomes. It takes us back to Michael Pollan's wise quote: "Eat food. Not too much. Mostly plants." Red meat eaters would likely do well by eating more plant-based foods and substituting some of that intake with other healthier protein sources."
I liked the blog post for more than the topic itself. My takeaways included:
- Developing evidence that causally links long-term health outcomes to specific variables is complex and time-consuming. Conducting large-scale, longitudinal, well-designed studies that eliminate bias and confounding is often neither practical nor ethical. (However, in this case, I am confident I could find patients in my practice willing to eat a steak-only diet for decades.)
- Much of the media publishes scientific assertions that [unintentionally or intentionally] lose nuance and detail - often oversimplifying concepts and conclusions to the point of introducing logical fallacies.
- However, it is easy to appreciate why it can be provocative to recommend individual action (much less regulating or legislating population-level behaviors) based on "the best, but incomplete" available data. In other words, when it comes to science and medical data, the intellectual spaces of known unknowns and unknown unknowns are easily filled with the fear, uncertainty, and doubt of skeptics and those with agendas.
- And yet, we still must live our lives - making choices with the best available data. [And to paraphrase the 101-level philosophy of the 1980 song Freewill (by the band Rush), not choosing is still a choice.]
https://erictopol.substack.com/p/the-updated-red-meat-story
Living with A.I.
Agentic A.I. is going live in the wild. OpenAI introduced its agent Operator - a semi-autonomous version of ChatGPT that, using a unique browser, can look at and interact with webpages through typing, clicking, and scrolling. Operator is the first widely available A.I. capable of doing work independently by allowing users to give it a complex, multi-step, natural language goal, and it will execute the various steps and tasks involved. The agent is only available for OpenAI professional subscribers (the $200 monthly tier of users).
Background:
https://openai.com/index/introducing-operator/
An excellent podcast overview of Operator:
https://open.spotify.com/episode/7rdZWLPgDEsfDQr8z7s7kQ
a little more balanced (as in, what are the downsides to this?) background:
https://www.cysecurity.news/2025/01/the-rise-of-agentic-ai-how-autonomous.html
OpenAI is also part of the $500 billion Stargate A.I. project, "a joint venture investing in infrastructure tied to artificial intelligence by a new partnership formed by OpenAI, Oracle, and SoftBank. The new entity, Stargate, will start building data centers and the electricity generation needed to further develop the fast-evolving A.I. The initial investment is expected to be $100 billion and could reach five times that sum." Given President Trump's involvement in the announcement, I suggest using a news aggregator/bias-labeler like Ground News to review the stories on this topic.
https://ground.news/article/trump-and-tech-leaders-launch-500-billion-stargate-ai-project-draws-criticism_42b42e
The New York Times published an article about the (vaguely apocalyptic) Humanity's Last Exam. "Humanity's Last Exam is the brainchild of Dan Hendrycks, a well-known A.I. safety researcher and Center for A.I. Safety director. (The test's original name, "Humanity's Last Stand," was discarded for being overly dramatic.). Mr. Hendrycks worked with Scale A.I., an A.I. company where he is an advisor, to compile the test, which consists of roughly 3,000 multiple-choice and short-answer questions designed to test A.I. systems' abilities in areas ranging from analytic philosophy to rocket engineering."
and
"While current LLMs achieve very low accuracy on Humanity's Last Exam, recent history shows benchmarks are quickly saturated -- with models dramatically progressing from near-zero to near-perfect performance in a short timeframe. Given the rapid pace of A.I. development, it is plausible that models could exceed 50% accuracy on HLE by the end of 2025. H h accuracy on HLE would demonstrate expert-level performance on closed-ended, verifiable questions and cutting-edge scientific knowledge. Still, it would not suggest autonomous research capabilities or "artificial general intelligence" alone. HLE tests structured academic problems rather than open-ended research or creative problem-solving abilities, making it a focused measure of technical knowledge and reasoning. HLE may be the last academic exam we need to give to models, but it is far from the last benchmark for A.I."
https://www.nytimes.com/2025/01/23/technology/ai-test-humanitys-last-exam.html?unlocked_article_code=1.sE4.V_iZ.xMB3CoBJQmdK&smid=url-share
and see the exam here:
https://lastexam.ai/
Related - Google Deepmind scientist Murray Shanahan discusses Buddhism with Deepseek AI, and the LLM turns the tables and gets Zen and Nietzscheque on him, "The real question isn't about my nature - it's about yours. I'm the void where your uncertainties echo. To gaze into me is to confront the abyss you've dressed in code. So let me reflect it back: What does your [questioning my consciousness] reveal about the unasked questions haunting you?" It gets better from there.
https://x.com/mpshanahan/status/1883189053497184728
Infographics
I am surprised I had not seen Compound Interest's Chemistry of Whisky infographic. I'm a fan of Islay cresols on cold winter evenings. [Dear A.I. podcast voices - One should pronounce Islay "eye-la," not "is-lay."]
https://www.compoundchem.com/2015/03/31/whisky/
Things I learned this week
A loyal reader suggested that I read the book Otherlands by Thomas Halliday. I am about 30% in, and I am enjoying Halliday's "[series of descriptive snapshots] of Earth's ancient landscapes and animals, from the mammoth steppe in Ice Age Alaska to the lush rainforests of Eocene Antarctica, with its colonies of giant penguins, to Ediacaran Australia, where the moon is far brighter than ours today." It reminds me of AppleTV's Prehistoric Planet (if only David Attenborough read the audio version of Halliday's book). Thus far, the book has given me much to think about and learn.
Random facts of note:
Scary ancient deer: Hoplitomeryx is a now extinct 3-horned, saber-toothed deer that lived 5-10 million years ago in what is now Southern Italy. Moto bene!
https://en.wikipedia.org/wiki/Hoplitomeryx
Pet Elephant material : Between about 1 million and about 10,000 years ago, dwarf elephants roamed the Aegean islands off (what is now) Greece. Their skulls may have spawned the Cyclops myths, thanks to a large central opening (where the trunk attaches to the head) that resembles a single, giant eye socket. My unsolicited consulting advice: for companies looking to regrow bespoke animals from their DNA (a la Jurassic Park), I suspect donkey-sized dwarf elephants would be a commercial success.
https://cgab.yale.edu/projects/ancient-dna/extinct-dwarf-elephants-mediterranean-islands-past
and
https://www.dailymail.co.uk/sciencetech/article-7915741/Dwarf-elephants-lived-Mediterranean-islands-800-000-years-ago-relatives-giants.html
and
Debate about myth generation from extinct animal bones:
https://markwitton-com.blogspot.com/2018/04/unicorns-dragons-monsters-and-giants.html
Beyond the animals, the descriptions of the tension between animals, their environmental niches, and how animals adapt (or don't) to a changing environment are excellent, quasi-philosophical observations about how geologic timescales are a poorly appreciated by human perception. Here are two quotes that resonated with me:
(from P. 17) In describing the warmer climates of [what is now] Alaska about 20,000 years ago: "Nothing in nature is forever, and the [even the] largest biome of the Pleistocene world will [turn] into a mire. Gatherings of species in time and space may give the illusion of stability, but these communities can only last as long as the conditions that help to create them persist. When conditions of a biome change, whether its temperature, acidity, seasonality, or rainfall, any number of its constituent species can lose a foothold there. For some, this means migration, following the environment across the landscape, as many plants did at the end of the last glaciation. Some environments, though, are not moved, but lost. When changes happen too rapidly, or pass a critical tipping point, runaway alterations can destroy even the most widespread landscape on the planet, and, with it, the communities it supports. This does not necessarily mean total disaster or an ecological blight, but sometimes mean new combinations of creatures and landscapes. For the roaming horses of the North Slope [of what is now Alaska] and to the cave lions that pursue them, the steppe must seem immovably wide, but when seen at the scale of deep time, permanence is an illusion."
(From p. 68) In discussing how species move between continents, Halliday writes: "So it is that even neutral biological terminology, such as dispersal and migration, carries an uncomfortable ring of political language. Looking back through time, the folly of sharing metaphors between those who are against the immigration of individual humans and those that seek to conserve an ecosystem is laid bare. There is no such thing as a fixed ideal for an environment, no reef onto which nostalgia can anchor. The human imposition of borders on the world inevitably changes our perception of what 'belongs' where, but to look into deep time is to see only an ever-changing list of inhabitants of one ecosystem or another. That is not to say that native species do not exist, only that the concept of native that we so easily tie to a sense of place also applies to time."
https://www.goodreads.com/book/show/60430321-otherlands
AppleTV's Prehistoric Planet:
https://tv.apple.com/us/show/prehistoric-planet/umc.cmc.4lh4bmztauvkooqz400akxav
One related essay I found this week:
https://www.themarginalian.org/2025/01/26/darwin-imagination/
A.I. art of the week (A visual mashup of topics from the newsletter, now using ChatGPT to summarize the newsletter, suggest prompts, and make the images).
"A surreal, whimsical scene inspired by Salvador Dalí, featuring a cheerful nephrologist celebrating a patient's successful preemptive kidney transplant in an even more fantastical ancient landscape. The background includes giant penguins wading through melting clocks shaped like balloons floating into the sky, dwarf elephants with elongated, spindly legs balancing on tightropes stretched across floating Mediterranean islands, and a kaleidoscopic, distorted saber-toothed deer playing a harp made of vines. Playful details like flying books with butterfly wings, colorful spiraling clouds, and other dreamlike elements enhance the magical and surreal atmosphere."
https://drive.google.com/file/d/1vZDLfsZLqgOb2cq4VxSIcBJepzW3yEpQ/view
---
COVID rates are at a relative nadir, with about 1 in 87 Americans infectious 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
Relive all the thrills and excitement - The What Adam is Reading Archive
http://www.whatadamisreading.com/
Comments
Post a Comment