What Adam is Reading - Week of 9-29-25

Week of 9-29-2025

 

After ending a phone call with my older son, my Uber driver stated, "It is so nice to hear a supportive parent. My child had four legs, but he died last year." Coming from a medical conference, my automatic response was, "I'm so sorry, did your child have some sort of congenital abnormality?"  "No," the Driver said, "he was a dog." After laughing, we spent the remainder of the hour-long drive between downtown Chicago and Midway Airport during rush hour in conversation. I heard all about Buddy (the dog), Dennis' (the driver) 30-year struggle with HIV, and read the pages in a recent book highlighting Dennis and Buddy's 12 years of companionship, helping Dennis cope with HIV. I love random, meaningful conversations, and will be sad when taxis become fully autonomous.

 

Here is the book:

https://lernerbooks.com/shop/show/20310

Meet Dennis and Buddy:

https://drive.google.com/file/d/16Q_O2E6q1qe72hx0WYgEGtc5IVeQYhD_/view

 

---

Listen to a Google Notebook LM A.I.-generated podcast of the newsletter with two virtual "hosts."

 

https://drive.google.com/file/d/1soz24iVbbWr3EXQrkWPMEOBnYpKeG0FV/view?usp=drive_link

------

 

Science and Technology Trends

 

There was a plethora of interesting articles this week.

 

I am still waiting for an intranasal COVID vaccine. Spanish researchers demonstrated the effectiveness of a new intranasal vaccine platform (a live, attenuated, non-spreadable SARS-CoV-2 replicon [a DNA or RNA molecule that can replicate itself within a host cell] ) introduced into the nares of transgenic mice. There was 100% survival of the vaccinated mice compared to the control, and vaccinated mice demonstrated a robust immune response. Although this method still requires all the necessary testing for human use, it is novel and interesting.

Article:

https://www.pnas.org/doi/epdf/10.1073/pnas.2516587122

AI Summary:

https://claude.ai/public/artifacts/0710fc70-4c41-4ad3-b0d2-97200f48c87a

 

I found this recent Science article modeling the impact of what 40% NIH funding would mean if the president had made such funding cuts in 2024.   The goal is to illustrate the trade-offs associated with the president's proposed 40% funding cuts in the 2025 budget.

https://www.science.org/doi/epdf/10.1126/science.aeb1564

and

X discussion

https://x.com/mattsclancy/status/1971283145480655095?s=42

and AI summary

https://claude.ai/public/artifacts/ba77c268-7820-47b8-a89a-0bed944854af

 

Highly precise gene editing is now a thing. Far less random and more reliable than CRISPR-based edits, per the author's X post.

https://x.com/pdhsu/status/1971293821418995988

The article is pay-walled:

https://www.science.org/doi/10.1126/science.adz0276

 

Despite the current limitations of gene-editing technology, therapies are still possible. One example is recent data on treatment for Huntington's disease, a highly penetrant (meaning most people with the genetic abnormality display symptoms), 100% fatal (meaning 100% of people who get symptoms die) familial dementia. The disease progresses, in part, due to the accumulation of abnormal proteins (made because of genetic abnormalities) in the brain. UniQure (a biotech firm) issued a press release regarding a Phase 3 trial using a direct brain infusion of AMT-130 (UniQure's adeno-associated virus that introduces microRNA) targeting the Huntington protein. Over a three-year period, 12 patients receiving the therapy showed a 75% slower progression of symptoms compared to a large historical control group. This therapy is the first meaningful treatment for this terrible illness, though direct brain infusions of drugs are difficult to bring to market.

Article on the press release:

https://www.science.org/content/article/first-gene-therapy-seems-slow-huntington-disease

AI Summary:

https://www.perplexity.ai/search/summarize-this-in-a-new-artifa-ARzR4i3MSSGZasL9fSlnVw#0

 

 

Anti-Science Articles of Note

 

This week, thanks to a robust public discussion about Tylenol (acetaminophen), some people in the U.S. gained a deeper understanding of the distinction between correlation and causation. Here is a breakdown of the complex topic based on my reading:

More children are diagnosed with Autism, but the prevalence of symptoms has not increased. Broadening diagnostic criteria, the fact that a diagnosis often opens increased educational and other support, and increased awareness are all factors driving the frequency of diagnosis.

 

  • There is no ONE cause of Autism. The strongest correlation is genetic, but many factors can play a role - Autism is not a single disease.
    • Numerous studies over dozens of years demonstrate strong genetic risk factors combined with some environmental factors, such as advanced parental age, maternal metabolic conditions (obesity, diabetes, hypertension), some medication exposures (notably valproic acid), and perinatal complications. 
    • https://www.openevidence.com/ask/d3ecd990-d37a-424b-89be-2ae5db394ef6

 

  • The data associating acetaminophen exposure with Autism is complicated and weak.
    • There are numerous retrospective, observational studies.
    • Some demonstrate a statistical association (a slight trend toward a positive correlation), some do not.
    • Some of the higher-quality studies involving twins demonstrate that similar exposure does not cause both twins to exhibit autistic symptoms (i.e., null data in sibling analysis).
    • However, even the studies with a positive correlation demonstrate a low absolute risk (approximately 0.09% of developing symptoms by age 10).
    • It is unclear how much acetaminophen is unsafe. The studies often use self-reported data; thus, the "unsafe" frequency, dose, peak blood levels, and duration of acetaminophen use over 9 months of pregnancy are unknown.

https://www.openevidence.com/ask/742c01b8-2bac-45ea-95c5-0479422e8002

 

  • A noteworthy body of data demonstrates that antibodies to folate receptors (Folate Receptor Alpha autoantibodies, or FRAAs) may be a mechanistic factor in Autism. In other words, genetic and environmental risk factors MAY result in increased FRAAs, thereby contributing to some Autism cases.
    • The presence of FRAAs is more common in individuals with autism symptoms.
    • Not all individuals with autism symptoms have FRAAs.
    • In patients with FRAAs, treatment with leucovorin (folinic acid) improves verbal communication and behavioral outcomes.
    • Leucovorin has a favorable safety profile, as demonstrated by years of use across numerous studies.

https://www.openevidence.com/ask/88878e64-5c5b-4164-83fd-d2ecce74d649

 

 

Here are my concerns:

  • Unsophisticated consumers may not appreciate all the nuances.
  • Blaming Autism on a single medication, much less a medication that is generally safe for treating pain, is a gross oversimplification of a complex problem, while blaming mothers for uncontrollable outcomes and encouraging suffering for existing pregnant women.
  • None of this data is new. None of this is secret, unreleased, or "suppressed" knowledge. The studies are readily accessible. Offering this information in incorrect and sensational ways undermines the years of work done by thousands of clinicians and scientists.
  • Some of the cited data does not say what HHS leadership claimed it said: read this blog for an excellent example:

https://epiellie.substack.com/p/rfk-jr-clearly-hasnt-read-this-paper

  • Children with autism spectrum disorders deserve a diagnosis. Evaluation for genetic, environmental, and other causes is reasonable and, possibly, valuable. 
  • Focusing public attention on a single diagnosis and treatment is, at best, foolish and distracting.

 

 

Living with AI.

 

Google is expanding on concepts from Notebook LM to new AI-driven learning tools. Google calls this "a two-stage approach: first personalizing content based on grade level and interests, then creating multiple representations including slides, audio lessons, mind maps, and interactive assessments. A randomized controlled trial with 60 high school students demonstrated statistically significant improvements in both immediate learning outcomes and retention compared to traditional digital reading." The study has limitations (small sample sizes, short durations, etc.), but it is a good example of how AI may play a role in education.

The Tool:

https://learnyourway.withgoogle.com/

The Research:

https://arxiv.org/abs/2509.13348

AI Summary of the Research:

https://claude.ai/public/artifacts/a2e44562-5a5e-4542-acf6-85e0ff2fbe37

 

Stanford scientists have created the world's first bacteria-destroying viruses (called bacteriophages) entirely through artificial intelligence (AI). These phages are capable of infecting and killing antibiotic-resistant strains of Escherichia coli (E. coli). This research marks the first time AI has been used to write coherent, genome-scale viral sequences. Researchers, using the Evo 1 and Evo 2 AI models trained on millions of phage genomes, generated functional viral DNA. The researchers synthesized the AI-generated genomes, tested them, and found novel bacteriophages that could target and kill E. coli strains that naturally occurring phages could not. This research is not yet scalable, but it offers a glimpse into the future of synthetic biology and bioengineering.

Article

https://www.nature.com/articles/d41586-025-03055-y

Research Paper

https://www.biorxiv.org/content/10.1101/2025.09.12.675911v1

AI review of paper

https://claude.ai/public/artifacts/79d2a317-b1d8-47fe-99d8-a26910dce16b

 

Infographics

My favorite periodic table I have found so far: Can I lick it? A guide to how safe it is to lick each element on the periodic table.

https://i.imgur.com/tMs8jeC.jpeg

Related: What happens if you eat periodic elements?

https://www.youtube.com/watch?v=Iko-XX1mPUg

 

Long-time readers will recall my fascination with using periodic table structure for other topics:

https://www.whatadamisreading.com/2020/12/what-adam-is-reading-12-16-20.html

and the internet database of Periodic Tables

https://www.meta-synthesis.com/webbook/35_pt/pt_database.php?Button=Non-Chemistry+PTs

 

 

Things I learned this week

 

Not only are there Excel championships (as in MS Excel), but there is an entire industry around the Financial Modeling World Cup (FWMC). "Based in Latvia, Financial Modeling World Cup organizes a series of monthly competitions in financial modeling and Microsoft Excel. The Financial Modeling World Cup was founded in 2020 by Andrew Grigolyunovich, CFA, MFM." Las Vegas will host the next Microsoft Excel event on December 2-3, 2025. Spectating these events must be scintillating. FWMC should consider partnering with the sperm races I mentioned earlier this year. 

https://excel-esports.com/

 

I just finished John Green's Everything is Tuberculosis, a book that discusses the history, social dynamics, politics, and science of TB. I flagged numerous tidbits of information - including the history of the Adirondack chair. As Green writes, "Clean air, rest, and sunshine were believed to 'infuse new hope and courage,' as one person put it, and so sanatoria focused on controlling the behavior of patients and requiring them to be largely immobile and outside whenever possible. ( The reclining wood-slatted chair known as the Adirondack Chair was invented for TB patients, allowing them to rest outdoors without needing their beds wheeled outside.)"

https://www.smithsonianmag.com/innovation/how-adirondack-chair-became-feel-good-recliner-cures-what-ails-you-180978322/

 

Also mentioned in Green's book is the 1916 McClure's Magazine article by Edwin Bowers, a fake doctor, pioneering reflexologist, and spiritualist from the early 20th century. [It is because of these kinds of self-proclaimed doctors that I now have to take quarterly board exams, and the AMA exists!] Bowers was a strong believer in germ theory and came out strongly against facial hair, stating, "There is no way of computing the number of bacteria and noxious germs that may lurk in the Amazonian jungles of a well-whiskered face, but their number must be legion." It is a highly entertaining read that, thanks to Google, you can read (page 465 of the PDF):

https://www.google.com/books/edition/McClure_s_Magazine/Q5MVRhpdqDoC?hl=en&gbpv=1

Here is the article with extracted text:

https://docs.google.com/document/d/1ZI_HbnQsMebhq8zVw8j4IqlFp1rcC9pvk0RCBcxhrFM/edit?usp=sharing

About Edwin Bowers:

https://en.wikipedia.org/wiki/Edwin_F._Bowers

 

 

AI art of the week

A visual mashup of topics from the newsletter. 

I use ChatGPT to summarize the newsletter, suggest prompts, and make the images.

 

Prompt:

A surreal research cathedral where science, AI, and culture collide. Towering glass columns rise like organ pipes, each containing glowing bacteriophages—AI-designed viruses shaped like radiant insects—attacking holographic E. coli. Beside them, streams of DNA helices swirl into protective shields, while laser-precise edits slice cleanly through genetic code. In the courtyard, a colossal Adirondack chair reclines under blazing sunlight, surrounded by cheering crowds watching an Excel championship as if it were the Olympics. Floating above, whimsical periodic tables labeled "Can I lick it?" orbit like constellations, and bearded figures gaze into mirrors where their whiskers transform into overgrown bacterial jungles. The atmosphere is both awe-inspiring and humorous, fusing high science with cultural absurdity.

Rendering style: in the structured, geometric surrealism of Fritz Glarner, layered with the dreamlike humor of Irving Petlin.

 

Gemini:

https://drive.google.com/file/d/1491qlc7UWc-FPMLYHvcMmB6Jstu1laYc/view?usp=sharing

 

ChatGPT - best one for meeting the prompt this week

https://drive.google.com/file/d/1X_yMA2uWfV1zK5xwtJ0uf6wCiRDlX5dE/view?usp=sharing

 

Grok

https://drive.google.com/file/d/1hTrCMJD6QRcb_QPdi63U0Hf3Z4eV4zSC/view?usp=sharing

 

---

The Pandemic Mitigation Collaborative (PMC) utilizes wastewater viral RNA levels to forecast four-week predictions of COVID-19 rates.

 

Still reporting an average of 1 in 57 people infected last week.

 

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

Comments