Reminder: giving an AI the prompt "You are a world expert in X" doesn't make it a world expert. It doesn't give the model additional knowledge or improve its reasoning. All it does is change the *style* of the output, so it sounds like an expert, even when it's wrong.

Chicago riverwalk, memorial day weekend

Wabash Avenue bridge, Chicago, today

Five Things (May 23, 2026): AI in life sciences. doi.org/10.59350/810...

What you see in the Sun, is the Chicago skyline from the Indiana Dunes beach, across Lake Michigan. You can see it from 50 miles of distance due to a form of superior mirage, because the skyline is seen above where it's actually located. [📷 wesskywalker]

"This is why LLM AI will not solve cancer or climate change. Those problems require long time horizons, willingness to fail publicly, cooperations across competitors, and subordination of short-term returns to outcomes that may take decades to materialize." www.tomorrowsmess.com/p/the-fox-an...

me checking eRA Commons:

Some thoughts on where AI in genomics stands right now, building on a symposium we recently hosted at UChicago. Seven points, ranging from why scaling DNA models hasn't delivered, to why metadata is the real bottleneck, to the weird backlash against AI in academia open.substack.com/pub/blekhman...

good morning

reading about people taking ivermectin for hantavirus and the stories are really heartworming

Happening today!!

🔬How do clinical exposures shape the gut microbiome, and how do longitudinal clinical–microbiome signatures predict patient outcomes? Come find out at my poster #194 today at 7:30 pm at @cshlmeetings.bsky.social Biology of Genomes 2026! #bog26 @blekhman.bsky.social

Fabulous day in Chicago. Feels like summer 🌞

🧬 AI in Genomics Symposium next week! We're bringing together leading scientists from across academia and industry pushing the boundaries of AI in genomics research Join us May 8 at the University of Chicago - KCBD Auditorium, 9:30–5:00, reception to follow. No registration needed!

My take is that "Competitive but not discussed" is essentially not competitive. Meaning, these will not be funded. Program Officers are struggling to fund even highest scoring proposals right now. I don't see a world where C-NDs are even considered.

This is totally reasonable and normal and not a bubble www.cnbc.com/2026/04/15/a...

Cool new study -- they used historical water and well water fluoride levels across Wisconsin counties, and linked these to IQ and cognition from age 16 all the way to age 80. Guess what they found? No association between fluoride levels and cognition - at any age www.pnas.org/doi/full/10....

I'm moving to Purdue University as a tenured professor this summer! Excited to hire postdocs there and get a lab up and running with @runxishen.bsky.social as co-lead. More details: www.linkedin.com/posts/anneca...

Visited Zion National Park for the first time - impossible to convey how beautiful it is

Sean Davis & I are hiring a postdoc to work on turning massive public biological data collections into reusable engines for discovery. #ML / #AI + large-scale omics + open software Details + apply: cu.taleo.net/careersectio... seandavi.github.io | thekrishnanlab.org @cubiomedinfo.bsky.social

I'm generally a fan of AI, but I just don't see how this is going to improve polling. Will just make an already messy data even more biased and problematic

“first drafts of Results, Discussion, and Introduction should be the trainee’s own work. AI can help with grammar…tightening prose…critiquing the logic of an argument you’ve already made. But the hard part of forming the argument needs to be yours.’ blekhman.substack.com/p/you-need-t...

A lab using my recent articles on AI use during PhD training doi.org/10.5281/zeno... & doi.org/10.5281/zeno... to kickstart a discussion and draft guidelines for their group is exactly the kind of use I hoped these articles would inspire! Highly recommend reading Ran's post.

I wrote about why every lab should have AI use guidelines, and how to do it. open.substack.com/pub/blekhman...

After 13 years, the Blekhman Lab has a new website: blekhmanlab.org The old one was hand-coded HTML from my first week as a faculty member. It... showed its age. This one was built with Claude Code. Progress on multiple fronts.

Southern Utah countryside

"AI scientists are building ever-larger models, and then claiming that each model can solve even more than the previous one. Then, model in hand, they go around looking around for problems to solve" stevensalzberg.substack.com/p/ai-is-star...

Layers. Today in Chicago

Predicting host sex from metagenomic data Cool approach that uses the "host contamination" reads in shotgun metagenomic data, and (using as few as 450 reads) can accurately predict host sex. Works well across different host species and body sites www.biorxiv.org/content/10.6...

Project Hail Mary is one of my favorite books of the last few years, but it was hard to suspend disbelief with some of the story lines. Like that the main character is a PhD microbiologist who is also an expert in physics. Impossible

We're searching for the next Director of the Duchossois Family Institute at UChicago The most exciting leadership role in microbiome science right now: lead faculty, cutting-edge cores, and an FDA-approved facility for microbiome therapeutics Email/DM me if interested apply.interfolio.com/182203

Something that keeps happening recently: I open an email (or essay, cover letter, whatever), within 0.5 seconds recognize it's AI-written, then I delete it immediately. Anyone else's brain developed this new pattern recognition skill?

Leipzig U and the MPI for Evolutionary Anthropology (MPI-EVA) have an open faculty position (W2) in evolutionary population genetics! This position is tenured and comes with generous core funding. We are eager to welcome a new colleague! Deadline March 11. www.uni-leipzig.de/en/newsdetai...

Got 5 NIH study section invitations in one week. For 5 different panels in May. Never happened before. Not sure what's going on but feels like there is a serious strain on the system.

AI has huge promise for genomics -- but it has consistently failed at microbiome-based prediction. My new post on why simple models keep winning, where deep learning actually earns its place, and where the field is headed blekhman.substack.com/p/ai-keeps-f...

Last week I took one of my favorite Chicago photos of all time. This city is LONG

Our latest is out today in @plosbiology.org: "Metabolic modeling reveals determinants of prebiotic and probiotic treatment efficacy across multiple human intervention trials" @isbscience.org @uwbioe.bsky.social @uwgenome.bsky.social doi.org/10.1371/jour...

Yes, they can hallucinate papers that don't exist, discuss results that seem to be imaginary, and can be confusing and inconsistent. But talking to tenured professors may still be helpful

Terrible. Still need to see a bit more of the timeline, but based on this trend I don't see how the NIH will fund even 10,000 new awards this fiscal year. Maybe closer to 8000

In 1800 (not that long ago!) the average life expectancy was ~30 years.

I've been hearing from more and more people who received peer reviews that were clearly written by AI. It's frustrating, and no one knows what to do. I have some thoughts. blekhman.substack.com/p/so-your-pa...

Good morning from Chicago

I am seeking a postdoc to join my group at UCLA -- ideally the candidate would have some experience in either population genetics or microbes/microbiome (computational background needed). We have a range of projects and are happy to tailer to your interests. Please dm/email me if interested.

We should talk more about the graduate admissions crisis that's taking shape in the US. Talking to colleagues across universities, many graduate programs are slashing admission numbers and reducing class sizes substantially for the upcoming academic year. This is shortsighted

if you're in line to sack drake maye, STAY IN LINE

Our new paper, where we use metabolic modeling to show Fusobacterium grows faster in colorectal tumor vs normal tissue microenvironments, and use computational + experimental approach to find specific metabolic pathways driving host-microbiome interactions in cancer www.biorxiv.org/content/10.6...

This is not heartbreaking. Keep in mind that the authors perspective: they probably disagree that their paper "needs a lot of work". Instead they find another journal (and another set of reviewers) that agree with them, and save months of expensive, time consuming, unnecessary work.

Breast milk isn't just nutrition – it delivers live bacterial strains that colonize the infant gut and persist for months. Happy to share our new paper, where we used metagenomics to track bacterial strains between 195 mother-infant pairs over the first 6 months of life: doi.org/10.1038/s414...

I just signed up and read the first of Ran's posts. Excellent stuff, highly recommended.