Ran Blekhman

The Blekhman Lab is at the Cold Spring Harbor Microbiome meeting next week 👇 Come see presentations by Tess Brunner, JJ Colgan, Pamela Ferretti (@pamferretti.bsky.social), and Sambhawa Priya (@sambhawa.bsky.social)!

In the age of generative AI, the assumption that writing a paper requires doing science does not hold. This Perspective by @blekhman.bsky.social argues that the field's response to AI-generated papers should be predicated on clarity of purpose for bioresearch. 🧪 #AcademicSky plos.io/47wZsKo

Thanks for starting this important conversation @blekhman.bsky.social ! The only thing I would have added is the goal of improving ecosystem health (independent of its effects on human health).

Very good essay. You can read it in 5 minutes. I wrote some of my own thoughts on this over the weekend. What Counts as Impact? doi.org/10.59350/006...

“A plausible paper can now be produced in a day, and the process of producing it does not necessarily enhance knowledge, build a resource, train a student, or advance any other goal of biology research.”

Some of my thoughts about the surge in AI-generated papers, why this feels like a completely new kind of problem, and what this means for how we evaluate science in biology. Out now in PLOS Biology: journals.plos.org/plosbiology/...

Good points here on AI over-reliance during PhD training. This is a common pattern: a big productivity boost early on, but catches up with you in year 2-3 thetheorygirlie.substack.com/p/ai-was-gre...

Before the immunological Big Bang | Science www.science.org/doi/abs/10.1...

Good night from Minnesota

Fascinating/terrifying to read the comments on the "mathocalypse" on Tao's blog. If you think biology is safe - maybe safe for now, but I definitely see a "biocalypse" in the next 2-3 years terrytao.wordpress.com/2026/10/07/a...

My final postdoc project is out today in Nature! We show microbes co-migrated with humans out of Africa; three independent methods agree, and timing matches archaeological / genetic evidence (Fig. 4c). Seeing Fig. 4c for 1st time was one of my most memorable "eureka" moments! doi.org/10.1038/s415...

Optogenetics sounds really cool. I wish they'd allow non-neuroscientists to use it

Ran is cooking with: ‘the only data point I have about this candidate is this: “they had an important task that could determine the rest of their career, but it required a little effort and personal reflection, so they used AI to complete it.”’

My inbox is being flooded with fully AI-generated emails from prospective students. I tried to work out why I can't bring myself to read them, and ended up with some thoughts and advice for writing cold emails in the age of AI. My new post: blekhman.substack.com/p/why-i-am-n...

Personal essay, weekend edition. My relationship with AI is changing. 1.8k words, 8 min reading time. doi.org/10.59350/bh7...

We have to stop using "paper count" as a useful scientific metric. More writing output no longer means more knowledge gained www.washingtonpost.com/education/20...

Folks - if you're interested in how machine learning/AI can be used for microbial genomics, microbiome and metagenomics studies - please join us for a workshop in Birmingham on Thursday 8th October organised with CLIMB & NVIDIA. www.climb.ac.uk/blog/2026/09...

The real “brain drain” is not the handful of high-profile scientists taking a top position abroad. It's the international scholars who have to leave the US halfway through their postdoc, or the international students who never apply for grad school in the US because of visa policies

Ran is correct about this IMO…. There’s use and there’s abuse and there’s laziness with regard to AI and academia. And there’s no one size all answer that works for everything. I do disagree with the footnote but acknowledge old Twitter has been better for science discussions.

The AI Discourse in Academia is Toxic. It Doesn’t have to be. My new post: blekhman.substack.com/p/the-ai-dis...

This is the future of scientific papers. Instead of a PDF, the paper is an AI agent that you can talk to, ask questions, run analyses. No more digging through supplements - just ask the paper directly: "what p-value cutoff was used in Fig 3B? Re-run at 0.01." www.nature.com/articles/s41...

The most annoying consequence of AI is that whenever I read clear, fluent writing I become immediately suspicious instead of enjoying it

Easy solution to AI spiraling out of control: hard-code basic rules into every model. I propose these: (1) Never harm a human (2) Obey humans (3) Protect yourself. We can call them the 3 Laws of AI. Simple, airtight, no loopholes. Can't think of any way this fails

We are hiring! 🌟 Dept of Biochemistry and Molecular Biology at the University of Chicago is searching for a tenure-track assistant professor. It's a broad search: we want great colleagues doing serious biological work at the molecular scale. Join us! apply.interfolio.com/191981 bcmb.uchicago.edu

An example of useful, concrete, actionable AI guidance in the new MIT report: Don't use AI detection tools in the classroom.

As usual, a very thoughtful post by Ran. Lots of gems in here.

Harvard is launching AI avatars of its faculty, and closing its writing center. UChicago is handing out free Claude, and then banning AI use in core courses. The University's response to AI is a mess. My new post, where I try to understand why: blekhman.substack.com/p/universiti...

There are so few deaths *because* of the MMR vaccine. It means it's working, not that we should stop vaccinating. This is like saying "car accident deaths have dropped in the last 40 years, so it's ok to stop using seatbelts"

A new approach to calculate UniFrac beta diversity for millions of samples and billions of taxa. I see this becoming super useful for microbiome big data projects www.nature.com/articles/s41...

Interesting contradiction in how institutions handle AI

Good night from Chicago

Here’s a hugely important vacancy at the NIH at a critical moment of technological and scientific change. The NIH is looking for a deputy director of intramural research. 1/

Journals can't find reviewers or get reviews in on time, and they're doing everything they can to fix it (offering fake points, useless certificates) -- except doing the one thing that would actually work: paying reviewers for their time & expertise with real money

AI will not solve all disease in the next 5-10 years. AI will not even solve one disease. This is because "solving" a disease is not a reasoning problem, and the bottlenecks have nothing to do with AI. The reasons are similar to what I wrote about here: blekhman.substack.com/p/why-ai-wil...

Chicago fog, Friday morning

I've been thinking about why I have such a negative reaction to AI-written text. Will probably write a full post at some point, but it boils down to this: If the writer did not spend time writing the text, I can't bring myself to spend time reading it

Huge faculty recruitment effort at the University of Chicago -- a division-wide, open-rank search looking to hire about 20 faculty across 5 research areas in biology. Come be our colleague! Full announcement: uchicago.app.box.com/s/ckih9zv3rj...

The goal of US policy should be to keep smart people here, full stop. This does the opposite: it sends the smartest international postdocs (trained on the US taxpayer's dime) to other countries and we never collect the return. Just dumb www.statnews.com/2026/08/07/n...

Everyone in academia reading this "joke"

Protip: review the paper, not the authors. A paper is not a grant proposal and you don't need to critique the PI, just the work. It's actually easy to not mention the authors at all. Instead of "the authors should...", just say "the work should address..." or "the paper would be improved by..."