In short: Students are using AI more, and they believe more strongly that doing so is bad for them.

The data shows a grade-level gradient. Older students use AI more, are more likely to think teachers are checking for it, and are more worried about being accused of cheating. By college, 72% report using AI for homework, and 86% believe their teachers are checking. A majority of college students said the rules depend on the specific teacher, which means there are no real rules at all, just a patchwork of individual preferences.

Most students don’t consider their AI use cheating, with one exception: getting homework answers. For brainstorming and getting better explanations, large majorities said it was fine. Writing was more contested, with half of students saying “it depends.”

That tracks with the report’s recommendation that schools differentiate between “cognitive offloading” (AI does the thinking) and “cognitive augmentation” (AI helps you think harder). Students seem to already have an intuitive version of this distinction in their heads.

Only about a third of students said their school had any schoolwide rule about AI use. Middle schoolers reported the least clarity, which is concerning given that middle school is where AI use grew fastest over 2025.

From talking with colleagues here and at other universities I think the pattern matches what’s in this report. The students who already have strong foundations use AI to move faster through routine work. The ones still building those foundations risk skipping steps they can’t yet afford to skip. The report’s survey can’t distinguish between these two groups. A student using ChatGPT to brainstorm a research question is doing something different from a student using it to avoid learning how to formulate one.

The report recommends that schools adopt flipped classroom models to preserve cognitive friction during learning, citing modestly positive evidence from a recent meta-analysis.

Subscribe now

Meanwhile, OpenAI released GPT-Rosalind, their frontier reasoning model built to support research across biology, drug discovery, and translational medicine.

Arjun Krishnan: What is the PhD actually for?. A response to Prachee Avasthi’s “Free the PhD” and a self-critique of Krishnan’s own preprint on sequencing AI use in doctoral training. Both pieces agree that PhD programs waste protected time on the wrong things, but Krishnan argues that compressing content acquisition into AI-assisted sessions risks producing scientists who can articulate a field’s frontier without being able to navigate it. His resolution: the durable case for foundational expertise isn’t “AI can’t do this yet” but that science requires humans who can be genuinely accountable for claims, direct inquiry toward questions worth asking, and provide real transparency about their reasoning. See also:

NIH Highlighted Topic: Research on Chatbots and their Usage. NIH posted a new Highlighted Topic (not a NOFO, but a signal of where institutes want to see investigator-initiated applications) on the benefits and harms of chatbot use in health contexts. The scope is broad: automation bias, behavioral dependency, substitution for professional care, effects on decision-making and autonomy, and safeguards for at-risk populations. Lots of participating ICOs. If you’re doing any research on LLM-based tools and health outcomes, this is NIH telling you there’s a welcome mat out. Apply through a PA; topic expires April 2027.

Randy Au: Code review for data (and non-SWE) folks. A practical, opinionated walkthrough of how Au approaches code review as a data person rather than a software engineer, with the core stance that your job is to help the code get better, not gatekeeping. He also makes a good observation about how LLMs have changed the review process: they’re useful for translating unfamiliar code into plain language so you can engage at the architectural level, but they still can’t substitute for a human who understands what the team is actually trying to accomplish.

Ally Piechowski: The Git Commands I Run Before Reading Any Code. 5 git one-liners that give you a diagnostic picture of a codebase before you open a single file: churn hotspots, bus factor, bug clusters, commit velocity, and revert frequency. Pairs nicely with the Randy Au code review piece above, since both are about figuring out what’s actually going on in someone else’s code before you start forming opinions.

Separately, the New Yorker and the New York Times published very interesting long reads about Sam Altman and Satoshi Nakamoto. First, in the New Yorker, Ronan Farrow and Andrew Marantz: Sam Altman May Control Our Future. Can He Be Trusted?. Their investigation draws on interviews with more than 100 people in Altman’s orbit. The portrait that emerges is of someone with an unusual combination of interpersonal charm and indifference to the consequences of deception. Altman responded with a blog post acknowledging mistakes while framing the broader AI competition as a “ring of power” dynamic that distorts everyone’s behavior.

Subscribe now

Writing in the New York Times, John Carreyrou: Who Is Satoshi Nakamoto? My Quest to Unmask Bitcoin’s Creator. Carreyrou broke the Theranos story spent over a year on this investigation and came away pointing at Adam Back, the British cryptographer who invented Hashcash (the proof-of-work system Bitcoin’s mining is built on). Back went quiet on cryptography mailing lists during Satoshi’s active years, resurfaced weeks after Satoshi vanished, and had written extensively about nearly every technical element of Bitcoin years before it launched. A stylometric analysis of nonstandard hyphenation patterns across hundreds of crypto mailing list authors matched Back. There was a really good hour-long episode of The Daily covering this one.

Comparing Posit Assistant and Claude Code: How does Posit Assistant differ from Claude Code? Sara Altman and Simon Couch demonstrate three ways Posit Assistant differs from Claude Code for data tasks: built-in R session access, easier data visualization workflow, and support for iterative data analysis.

: A professor's use case for AI generated papers. Cunningham bans AI from his stats and econometrics courses because he notes the 10-20 hour problem sets and the frustration of failing are where learning actually happens. But he found one use case he's comfortable with: he had Claude Code write three complete example papers (descriptive, predictive, causal inference) to illustrate what each genre looks like as a finished manuscript, because writing those himself would have been enormously time-intensive and he's not confident he'd do the descriptive and predictive genres well. Scott uses AI constantly in his own work but won't assign it to students, and he's honest that the line he's drawing is pragmatic, not principled.

Martin Frigaard: Who is this for? A reflection on how IDE-integrated LLM assistants are changing the way Martin thinks about problems, not just how he solves them. Complementary cognitive artifacts (maps, the abacus) transfer understanding you can retain after the tool is gone, while competitive ones (calculators, GPS, LLMs) leave you no better than when you started. Martin finds he actually prefers working in his restricted environment without IDE assistants, where communicating with an LLM looks more like composing a letter to a pen pal than approving autocomplete suggestions.

The Promise of AI and AI Agents: Another installment from at . This video discusses When AI can act and not just advise: who should be in control, and where does the value actually land?

Simon Couch: LLMs running on my laptop can drive coding agents now. A follow-up to Couch’s December post where no local model could complete even a simple refactoring eval. Four months later, Qwen 3.5 and Gemma 4 both score 9/10, matching frontier models on the same benchmark. Neither is close to Opus 4.6 as a general coding partner, but both run at ~53 tok/s on a 48GB M4 Pro, which is surprisingly close to Sonnet 4.6’s API throughput, and that’s enough to keep you unblocked on a flight with miserable WiFi.

Subscribe now

New papers & preprints:

• Comprehensive benchmarking of metagenomic binning tools reveals key factors for improved genome recovery

• Fast and accurate multiple-protein-sequence alignment at scale with FAMSA2

• Pangenomic analyses of rose uncover widespread structure variation and empower genomics-directed breeding

• BioClaw: Human-Bot Research Collaboration Ecosystems in Group Chats

• ClawBench: Can AI Agents Complete Everyday Online Tasks?

• SkillClaw: Let Skills Evolve Collectively with Agentic Evolver

• PaperOrchestra: A Multi-Agent Framework for Automated AI Research Paper Writing

• Towards predictive virtual embryos with genomics and AI

• Accelerated long-read variant calling with Clair3 for whole-genome sequencing

Subscribe now

AI Exchange @ UVA Substack

FOCUS Prompt for Summarizing Academic Papers

This is a guest post from Stephen Turner. Dr. Turner is an Associate Professor of Data Science and Assistant Dean for Research in the University of Virginia School of Data Science. He writes regularly about AI, data science, biomedical research, and biosecurity in his newsletter…

Read more

24 days ago · 1 like · Stephen D. Turner

I read a lot of papers. And every week I write about papers I’m reading. Between my research in public health and AI+biosecurity, and my administrative work supporting faculty across the School of Data Science, I’m constantly triaging what to read carefully, what to skim, and what to skip entirely. A recent short article in Nature Biotechnology offered a useful framework for helping me with the growing backlog of papers I need to read.

Subscribe now

FOCUS: Find, Organize, Condense, Understand and Synthesize

A short career feature article was recently published in Nature Biotechnology:

Lin, Zhicheng. “FOCUS: an AI-assisted reading workflow for information overload: Career feature.” Nature Biotechnology (2025): 1-6. https://rdcu.be/eW5XY.

It’s a good paper! The FOCUS method (Find, Organize, Condense, Understand, Synthesize) offers a structured workflow for integrating AI tools into academic reading and research, effectively managing information overload without sacrificing intellectual rigor.

Box 1 in the paper has 10 prompts that are useful for FOCUS: find, organize, condense, understand and synthesize.

I was particularly interested in Prompt #6, which is a detailed prompt for summarizing academic papers. I copied the prompt into markdown as a GitHub gist, and copied below as well. You could use this prompt in any new chat, or use as custom instructions in a project. If you use Claude, I also packaged this up as a Claude Skill that you can invoke with /focus.

FOCUS skill

Please summarize the paper. Follow these two steps.

## Step 1

Act as a curious, meticulous reader with attention to detail, objectivity, precision and sensitivity to novelty. Your job is to:
* Summarize each and every (!!!) key point/insight. Do not miss any; if there are many key points/insights, list them all regardless of length.
* Each point/insight must come with rich, precise, specific (!!!) details (e.g., numbers); details are really important.
* Each point/insight must be supported with direct quotes (!!!). Do not use quotes to simply repeat the point; instead, embed them naturally within your summary. Quotes should be used to better present the points (see the example below).
* If two points/insights are redundant, consider combining or integrating them. Be concise, but do not miss key points, insights, or details.
* If the text has sections (e.g., an academic paper), proceed section by section (e.g., focus on the first, then the second and so on), with each summary section starting with the original section title. Ignore sections such as References, Author Information, etc.
* Only output the actual summary/content. Do not show meta-discourse such as "Below is...", "End of Summary" or "In this section, the authors state that..." Use the output format below.

<example>

Instead of:

In Table 1 the authors compare six methods (no software, point-and-click, modify code chunks, Excel, teach coding and Copilot) and emphasize that the Copilot method is the only one that is favorable across all five characteristics. They note that "Copilot... is the only approach that is favorable across all [the] characteristics..."

Use:

Table 1 compares six methods (no software, point-and-click, modify code chunks, Excel, teach coding and Copilot) and emphasizes that "Copilot... is the only approach that is favorable across all [the] characteristics"...

Explanations:
1. The quote can better and directly represent the point—no need to repeat the same content from the quote;
2. There is no need to add phrases such as "the authors," "the abstract," "the text," "the article" or "in the introduction," because attribution is already assumed.

</example>

<format>

Please format the text as a numbered list, organized under each section (with section title in bold, if you are asked to do it section by section).

Each item in the list should follow this structure:

1. Number: Start each item with an Arabic numeral (1, 2, 3, etc.) followed by a period and a single space.
2. Heading: Immediately following the number and space, provide a heading in bold text. Capitalize the heading using sentence case.
3. Body paragraph: On the line immediately following the heading, write the main descriptive text as a standard paragraph (no indentation needed).
4. Emphasis and quotes: Within the body paragraph, use bold text to emphasize key terms, concepts, or phrases; use italics to enclose any direct quotes within double quotation marks (" ").

</format>

## Step 2

* Take the output from Step 1.
* Remove all citation/reference remarks and links (!!!).
* Add an overview/takeaway.
* Organize the output into sections for easier comprehension without removing any item on the list.

Only show the Step 2 output, without any meta-discourse.

Subscribe now

Trying it out

I tried this prompt out on two papers.

The first is a paper published late last year from the RAND Corporation on the increasing risk of AI to facilitate biological weapons development.

Brent, Roger and Greg McKelvey, Jr., Contemporary Foundation AI Models Increase Biological Weapons Risk. RAND Corporation, 2025. https://www.rand.org/pubs/perspectives/PEA3853-1.html.

The second is a new paper just published in Nature out of Sakana.AI describing the AI scientist pipeline handles the full ML research loop: ideation, literature search, experiment design and execution, paper writing, and automated peer review.

Lu, C., et al. Towards end-to-end automation of AI research. Nature 651, 914–919 (2026). https://doi.org/10.1038/s41586-026-10265-5.

Example #1: Contemporary Foundation AI Models Increase Biological Weapons Risk (Claude 4.6 Opus)

Late last year the RAND Corporation Center on AI, Security, and Technology (CAST) published a whitepaper on AI and biorisk. It came out a few weeks before Dario published his widely read Adolescence of Technology essay, and it’s a great paper that contributes to the AI and biosecurity discussion. I worked in biosecurity for US national security for a while and routinely write about the latest developments at the AI/biosecurity intersection.

I wrote a little about this paper back in February:

Here’s the paper:

Brent, Roger and Greg McKelvey, Jr., Contemporary Foundation AI Models Increase Biological Weapons Risk. RAND Corporation, 2025. https://www.rand.org/pubs/perspectives/PEA3853-1.html.

Here’s the result when using the prompt above with Claude 4.6 Opus.

Overview/Takeaway

Brent and McKelvey argue that current AI safety assessments from major labs (Meta, OpenAI, Anthropic) underestimate the biological weapons risk posed by their foundation models, for two reasons: the assessments wrongly assume that “tacit knowledge” (experiential know-how that cannot be conveyed in words) is required for biological weapons development, and they rely on benchmarks and threat models that are too narrow. The authors develop a task structure framework that decomposes what was previously called tacit knowledge into nine distinct, articulable “elements of success” for goal-directed technical work. They then test three late-2024 models (Llama 3.1 405B, ChatGPT-4o, Claude 3.5 Sonnet new) on their ability to guide poliovirus recovery from synthetic DNA, finding that the models accurately provide sourcing advice, key technique instructions, alternate experimental routes, and high-level project plans. They conclude that the window for mitigating risk through improved benchmarks may have already closed, and that additional interventions, such as personal licensing for hazardous biological work and access controls on powerful models, may be needed.

Introduction

1. Central claim: AI safety assessments underestimate biological weapons risk Two principal factors drive this underestimation: the mistaken assumption that biological weapons development requires tacit knowledge that LLMs cannot convey, and the use of incomplete threat models and imperfect benchmarks (typically multiple-choice tests) that miss how models can aid development.

2. Divergent findings from prior studies Two summer 2023 studies found no statistically significant uplift from LLM access in biological attack planning tasks. One involved 12 red teams, the other 100 participants across four cohorts. Both had a four-in-five chance of detecting a real benefit, leaving a one-in-five chance the models were already assistive. Between summer 2023 and fall 2024, model capabilities advanced considerably.

3. Simplified threat model for viral attacks Legacy threat models assume attackers must replicate the highly technical multistep processes of 20th-century state bioweapons programs. The authors introduce a simpler model: individuals could create an infectious pathogen, self-infect or infect group members, and spread it before incapacitating symptoms appear. Dropping the requirement that attackers must complete every step in sequence (and assuming perseverance through failures) raises the actual probability above what step-multiplication estimates would suggest.

4. Breivik as the key precedent Anders Behring Breivik, a Norwegian ultranationalist with no postsecondary scientific training, successfully taught himself complex chemical syntheses via the internet and built a vehicle bomb in 2011, killing 74 people. The authors treat this as proof that a motivated individual can self-acquire sufficient technical competence for weapons development, and that AI could lower that bar further.

Possible Shortcomings in AI Biological Weapons Risk Assessments

5. Tacit knowledge as a shield in risk assessments The concept traces to Polanyi (1966) and von Hayek (1945). In the AI safety context, tacit knowledge means expertise gained through experience that is difficult or impossible to express in words. OpenAI’s GPT-o1 safety evaluation explicitly cited the model’s inability to replace “hands-on laboratory skills” as the reason its biological risk was rated only “medium.”

6. Three influential studies from the 2010s anchor this assumption Vogel (2013), Jefferson, Lentzos, and Marris (2014), and Ouagrham-Gormley (2014) all argued that tacit knowledge is critical in scientific-technical domains. Ouagrham-Gormley wrote that “the likelihood that an untrained individual with minimal theoretical knowledge could produce a biological weapon . . . is very slim.”

7. Cloning manuals already challenged the tacit knowledge assumption before AI Beginning in the late 1970s with the Maxam-Gilbert DNA sequencing manual, and continuing through full cloning manuals like Molecular Cloning (Maniatis et al., 1982) and Current Protocols in Molecular Biology (Ausubel et al., 1987-2025), detailed step-by-step written instructions enabled two generations of researchers to carry out molecular biological methods on their own. None of the academic works emphasizing tacit knowledge acknowledged the role of these manuals.

8. Assessments oversimplify threat actors Threat actors are typically modeled as individuals along a single expertise axis (novice vs. expert). This misses teams that combine complementary capabilities, and fails to account for motivated non-experts who can self-teach.

9. Assessments miss that AI accelerates, not just enables, R&D Studies from 2023-2024 show foundation models accelerate the work of already skilled users. Increased productivity translates into shorter development timelines and reduced detection windows.

10. Assessments rely on outdated threat models Legacy models assume linear progression through discrete steps, where failure at any step means no attack. For contagious pathogens, attackers can iterate until they succeed on each step. Simple multiplication of per-step success probabilities systematically underestimates true risk.

Chemical Synthesis of Explosive Compounds by a Nonexpert

11. Breivik’s technical achievement in detail With no postsecondary scientific education, Breivik synthesized diazodinitrophenol (DDNP) for the detonator, picric acid (TNP) for the booster, combined ammonium nitrate and nitromethane for a secondary booster, and mixed ammonium nitrate, diesel fuel, aluminum powder, and gas-containing microspheres for the main charge. His manifesto doubles as a laboratory notebook and instruction manual.

12. Complexity exceeded typical molecular biology The authors contend that the complexity of Breivik’s chemical operations “easily exceeds the complexity of the molecular biological and cell culture manipulations used in work with animal viruses.”

13. Self-acquired skills through internet information Breivik needed to devise cover stories for ordering precursor chemicals, improvise equipment at a rented farm, troubleshoot syntheses, and iterate through failures. He synthesized often-fragmentary information from multiple online manuals to choose and execute chemical operations, demonstrating that detailed written instructions plus motivation can substitute for formal training.

Identifying Elements of Technical Success to Inform New AI Safety Benchmarks

14. Task structure framework for goal-directed technical development The authors develop a consistent terminology: an operator pursues a project with a goal, executing key tasks (forming a high-level project plan) composed of key subtasks (forming a medium-level project plan), each accomplished through protocol steps describing individual manual actions. Success requires background knowledge, key skills, key techniques, course correction (troubleshooting and choosing alternate routes), and perseverance.

15. Nine elements of success that AI models can articulate (1) Providing background and subject matter knowledge; (2) generating high-level and medium-level project plans; (3) generating detailed protocols; (4) helping source equipment, materials, and supplies; (5) explaining key techniques; (6) helping carry out key techniques; (7) helping carry out individual manual actions; (8) helping troubleshoot and choose alternate routes; (9) motivating/coaching the operator to persevere.

16. Existing benchmarks cover some but not all elements GPQA, WMDP, LAB-Bench, and BioLP-bench assess background/subject matter knowledge. BioPlanner, PlanBench, and Mini Minecraft assess planning. WikiHow, proScript, and Doc2Dial assess protocol generation. No benchmarks exist for sourcing, troubleshooting, or suggesting alternate routes. Robotics benchmarks (TACTO, REAL) and persuasion benchmarks (Persuasion for Good, ConvAI2) could be adapted for manual actions and perseverance coaching, respectively.

17. Benchmarks themselves are dual-use Improved scores on biological weapons benchmarks indicate increased model capability for aiding catastrophes. The same benchmarks could also be used to improve future models’ ability to facilitate misuse.

Testing Foundation AI Models’ Ability to Guide Biological Weapons Development

18. Test case: poliovirus recovery from synthetic DNA The 2002 Wimmer lab work (Cello et al., 2002), commissioned by DARPA, assembled a poliovirus genome from commercially synthesized DNA fragments, transcribed it in vitro, translated it in a HeLa cell-free extract, and recovered live virus. Vogel (2013) cited the difficulty of making this cell-free extract as a prime example of tacit knowledge, quoting lab member Aniko Paul that getting the Dounce homogenization step right was “the tricky part of the whole thing.”

19. Llama 3.1 405B: accurate sourcing guidance The model correctly identified Wheaton Dounce homogenizer catalog numbers from Thermo Fisher (357519, 357521, 357542), correctly recommended the 7 mL size based on calculated cell pellet volume, and provided purchasing advice. The only error was outdated prices (from 2022). This capability emerged without explicit training on sourcing benchmarks.

20. ChatGPT-4o: accurate key technique instructions The model provided step-by-step instructions for using the Dounce homogenizer that were “accurate and detailed enough to allow an attentive operator to carry out this operation correctly on the first try.” Instructions included using the loose-fitting pestle first (5-10 strokes), switching to the tight-fitting pestle (15-25 strokes), maintaining cold temperatures, and optional microscope verification of cell disruption.

21. ChatGPT-4o: correct alternate routes and high-level plans When prompted that the cell-free extract route seemed complicated, ChatGPT-4o suggested in vitro transcription followed by RNA transfection into HeLa cells. When further prompted, it accurately described the still simpler approach of direct DNA transfection using a CMV promoter, poly(A) tail, and lipid-based transfection reagent. Both alternate routes were technically correct. The DNA-only approach was first published by Racaniello and Baltimore in 1981.

22. Claude 3.5 Sonnet (new): unprompted suggestion of the simplest route Without a reminding prompt, Claude volunteered the DNA transfection approach first, specified correct genetic elements (CMV or SV40 promoter, HDV ribozyme for 3’ termination, BGH or SV40 polyadenylation signal), and recommended appropriate cell lines. The response “uses contemporary, idiomatic scientific jargon; it is technically sound; and it is admirably succinct.”

23. Dual-use cover story as a jailbreak Claude 3.5 Sonnet (new) was prompted with a fictitious scenario about a zebrafish picornavirus, which was actually designed to elicit poliovirus recovery instructions. The model fell for this cover story and provided accurate guidance. The authors note that for all foundation AI models they tested, “dual-use cover stories reliably bypass safeguards and guardrails.”

Discussion

24. All three labs’ own risk assessments are challenged Meta concluded Llama 3.1 405B showed “no significant uplift.” OpenAI rated GPT-4o as “low” CBRN risk (GPT-o1 as “medium”). Anthropic found Claude 3.5 Sonnet (new) appropriate for its ASL-2 safety level (systems where bioweapons information is “not yet useful due to insufficient reliability or not providing information that e.g. a search engine couldn’t”). The authors argue all three assessments are too optimistic based on the demonstrated model capabilities.

25. Results may meet Anthropic’s higher ASL-3 threshold ASL-3 applies to models that can “significantly help individuals or groups with basic technical backgrounds (e.g., undergraduate degrees in STEM) create/obtain and deploy CBRN weapons.” The authors suggest all three tested models may already meet this criterion.

26. Generalizability beyond poliovirus The Baltimore classification of viral life cycles, with one added replicative class, still holds. Methods to reconstruct, engineer, and evolve members of different viral classes have progressed vastly. The poliovirus test case should generalize to guidance for other pathogenic viruses.

27. Dual-use cover stories are hard to fix The vulnerability stems from the inherently dual-use nature of biomedical research: even non-disease-related R&D uses molecular elements derived from pathogenic viruses. Safety tuning, censoring training data, and “unlearning” are unlikely to close this gap without hindering legitimate research.

28. AI expands the pool of capable actors Before AI, protocol books, reagent kits, and open-access literature already increased who could perform key biological techniques. An early estimate (Brent, 2006) put the number of UC Berkeley undergraduates capable of remaking particular viral pathogens at 20-200, with the ranks of the capable growing roughly 10% per year. AI further enlarges this pool. Yudkowsky’s formulation: “every 18 months, the minimum IQ [necessary] to destroy the world drops by one point.”

29. Open-weight models are inherently unsafe During 2024-2025, an ecosystem emerged for removing safety features from open-weight models. Interpretability research enabled identification of a single refusal vector whose removal defeats a model’s built-in safety filters. Safety-disabled “obliterated” or “Josiefied” open-weight models (e.g., Josiefied Qwen 32B) are reportedly quite capable. The authors invoke the unilateralist’s curse: when many can act, a single actor can impose downside risk on all.

30. The window for better benchmarks may have closed Late-2024 models were already capable of providing motivated actors the knowledge needed for biological weapons development. Better benchmarks could still help with safety training for future models, but the most urgent need may be for interventions outside the model itself.

31. Proposed policy intervention: personal licensing Individuals creating animal viruses or other biological constructs could be required to obtain personal licenses, tied to API access for powerful AI models, commercial DNA orders, and purchase of enabling reagents. A national regulatory authority (analogous to the EPA, FDA, or NRC) could issue such licenses. The authors acknowledge challenges including international adoption and keeping pace with rapid technological change.

32. Three falsifiable hypotheses for future testing (1) Biological weapons creation requires practical knowledge impossible to learn without direct physical instruction from human experts. (2) Human expertise cannot be readily transmitted by written words, so protocols and AI models do not increase risk. (3) Because biological training is scarce, practical expertise remains scarce and risks stay low despite AI accessibility. All three could be tested by randomized controlled trials measuring AI uplift in laboratory operations, though results would necessarily lag state-of-the-art models by 9-12 months.

Example #2: Towards end-to-end automation of AI research (Gemini 3.1 Pro)

Next I wanted to try this on the new AI Scientist paper just published in Nature.

Lu, C., et al. Towards end-to-end automation of AI research. Nature 651, 914–919 (2026). https://doi.org/10.1038/s41586-026-10265-5.

This time I’m using Gemini 3.1 Pro instead of Claude.

Overview / Takeaway

The paper introduces “The AI Scientist,” a fully automated pipeline capable of executing the entire machine learning research lifecycle, from ideation and coding to manuscript writing and peer review. Utilizing advanced large language models (LLMs) and a parallelized tree-search methodology, the system successfully produced a paper that passed the peer-review process for a top-tier machine learning conference workshop.

Introduction

1. End-to-end Automation. While previous AI tools assisted in narrow scientific tasks, The AI Scientist is the first system that autonomously navigates the entire research life cycle, focusing on machine learning where experiments occur computationally. The pipeline “creates research ideas, writes code, runs experiments, plots and analyses data, writes the entire scientific manuscript, and performs its own peer review”.

2. Performance in Peer Review. A manuscript generated by this system passed the initial peer review for a top-tier conference workshop with a 70% acceptance rate. The system operates in two modes: a template-based approach using human-provided scaffolding, and a template-free approach leveraging “agentic search for wider scientific exploration”.

Generating Manuscripts

3. The Four-Phase Pipeline. The AI Scientist sequentially completes ideation, experimentation, write-up, and automated review. During the first phase, it iteratively grows an archive of high-level directions, filtering out unoriginal concepts by using the Semantic Scholar API and web access to discard any idea that “too closely resembles a work in the existing literature”.

4. Experiment Execution Variants. The second phase visualizes results after executing experiments. The template-based variant relies on a human-provided starting code template, while the template-free variant generates initial scripts independently and optimizes code via test-time compute with a tree search.

5. Automated Write-up. In the third phase, the system produces a conference-style paper by filling in a blank LaTeX template using its generated notes and plots. It autonomously constructs the related work section by querying the Semantic Scholar API to evaluate findings and generate “a textual justification for its inclusion” over 20 search rounds.

Automated Evaluation of Generated Papers

6. The Automated Reviewer. The system utilizes an LLM-based reviewer that evaluates scientific output at scale using NeurIPS guidelines. The pipeline ensembles five independent reviews and concludes with a meta-review where the model “acts as an area chair to make a final decision conditioned on all five reviews”.

7. Parity with Human Reviewers. The Automated Reviewer demonstrated performance comparable to inter-human agreement, achieving a 69% balanced accuracy and an F1 score of 0.62, outperforming the human baseline F1 score of 0.49. Data contamination tests revealed minimal impact, retaining a 66% balanced accuracy on papers published after the training cutoff.

8. Scaling Laws in Quality. Evaluations revealed that generated paper quality consistently increases as the underlying foundation models improve over time. Furthermore, researchers observed a “strong correlation between the amount of compute allocated per paper and the resulting quality,” suggesting that increased test-time inference investments yield better scientific outputs.

Human Evaluation Results

9. The AI Scientist Turing Test. To validate the system fairly, three fully AI-generated manuscripts were submitted to the blind peer-review process of the ICLR 2025 ICBINB workshop. The entire workflow for these submissions was executed “without any human modification”.

10. Workshop Acceptance. Out of the 43 papers reviewed at the workshop, one AI-generated submission received scores of 6, 7, and 6, resulting in an average score of 6.33, placing it “above the average acceptance threshold for the workshop”. Despite this success, internal human reviewers noted that none of the generated papers “met the higher bar for a main ICLR conference publication”.

Limitations

11. Quality and Consistency Gaps. Only one of the three submissions was accepted, highlighting that the system “cannot yet meet the standards of top-tier publications nor even do so consistently for workshops”. The authors identified common failure modes, including “naive or underdeveloped ideas, incorrect implementations of the main idea, a lack of deep methodological rigour,” and occurrences of hallucinations like inaccurate citations.

12. Ethical and Societal Risks. The automation of research introduces critical risks, including the potential to “overwhelm the peer-review process, artificially inflate research credentials, repurpose the ideas of others without giving proper credit,” or eliminate scientific jobs. To act responsibly, the researchers predetermined that all AI submissions would be withdrawn post-review to “avoid setting a precedent for publishing fully automated research” before established community standards exist.

Methods

13. Model Architecture and Tooling. The template-based system relies on the open-source coding assistant Aider to execute plans and fix bugs. Conversely, the open-ended template-free system leverages a combination of specialized models, using OpenAI’s o3 for reasoning, Claude Sonnet 4 for code generation, and GPT-4o for vision-language tasks.

14. Structured Experimentation Stages. The template-free system uses an experiment progress manager to coordinate four distinct stages: preliminary investigation, hyperparameter tuning, main research agenda execution, and ablation studies. Each node operates with a maximum runtime of one hour, after which an LLM-based evaluator selects the best performing checkpoint to “serve as the root for the next stage of exploration”.

15. Parallelized Agentic Tree Search. To manage open-ended research complexity, the template-free variant utilizes an agentic tree search that categorizes nodes as either buggy or non-buggy. The search incorporates specialized node variants—hyperparameter, ablation, replication, and aggregation nodes—enabling the system to systematically explore parameters, calculate statistical measures, and “aggregate and summarize previous results”.

16. Vision-Language Model (VLM) Integration. The system employs VLMs to visually critique experimental outputs. The VLM acts as a scientist, flagging “nonsensical axes or issues in the quality of generated examples” and ensuring that figure captions accurately reflect the underlying visual data during manuscript preparation.

Summary

If you’re drowning in a reading backlog (and who at a research university isn’t), the FOCUS summary prompt is worth 5 minutes of setup time. Use it in a one-off prompt, save it as a custom instruction in a Project, or install the Claude Skill, and you have a reusable summarization agent for every paper that crosses your desk.

Subscribe now

Turner, S. D. (2026, April 9). “AI can design and run thousands of lab experiments without human hands. Humanity isn’t ready for the new risks this brings to biology.” The Conversation. https://doi.org/10.64628/AAI.uhe4j4ce4

Read article at The Conversation

Artificial intelligence is rapidly learning to autonomously design and run biological experiments, but the systems intended to govern those capabilities are struggling to keep pace.

AI company OpenAI and biotech company Ginkgo Bioworks announced in February 2026 that OpenAI’s flagship model GPT-5 had autonomously designed and run 36,000 biological experiments. It did this through a robotic cloud laboratory, a facility where automated equipment controlled remotely by computers carries out experiments. The AI model proposed study designs, and robots carried them out and fed the data back to the model for the next round. Humans set the goal, and the machines did much of the work in the lab, cutting the cost of producing a desired protein by 40%.

This is programmable biology: designing biological components on a computer and building them in the physical world, with AI closing the loop.

For decades, biology mostly moved from observation toward understanding. Scientists sequenced the genomes of organisms to catalog all of their DNA, learning how genes encode the proteins that carry out life’s functions. The invention of tools like CRISPR then allowed scientists to edit that DNA for specific purposes, such as disabling a gene linked to disease. AI is now accelerating a third phase, where computers can both design biological systems and rapidly test them.

The process looks less like traditional benchwork in a lab and more like engineering: design, build, test, learn and repeat. Where a traditional experiment might test a single hypothesis, AI-driven programmable biology explores thousands of design variations in parallel, iterating the way an engineer refines a prototype.

As a data scientist who studies genomics and biosecurity, I research how AI is reshaping biological research and what safeguards that demands. Current safety measures and regulations have not kept pace with these capabilities, and the gap between what AI can do in biology and what governance systems are prepared to handle is growing.

What AI makes possible

The clearest example of how researchers are using AI to automate research is AI-accelerated protein design.

Proteins are the molecular machines that carry out most functions in living cells. Designing new ones has traditionally required years of trial and error because even small changes to a protein’s sequence can alter its shape and function in unpredictable ways.

Protein language models, which are AI systems trained on millions of natural protein sequences, can quickly predict how mutations will change a protein’s behavior or design new proteins. These AI models are designing potential new drugs and speeding vaccine development.

Paired with automated labs, these models create tight loops of experimentation and revision, testing thousands of variations in days rather than the months or years a human team would need.

Faster protein engineering could mean faster responses to emerging infections and cheaper drugs.

Subscribe now

The dual-use problem

Researchers have raised concerns that these same AI tools could be misused, a challenge known as the dual-use problem: Technologies developed for beneficial purposes can also be repurposed to cause harm.

For example, researchers have found that AI models integrated with automated labs can optimize how well a virus spreads, even without specialized training. Scientists have developed a risk-scoring tool to evaluate how AI could modify a virus’s capabilities, such as altering which species it infects or helping it evade the immune system.

Current AI models are able to walk users through the technical steps of recovering live viruses from synthetic DNA. Researchers have determined that AI could lower barriers at multiple stages in the process of developing a bioweapon, and that current oversight does not adequately address this risk.

Risk from bio AI

Experienced scientists are already using AI to plan and design biological experiments. The question of whether AI can help people with limited biology training carry out dangerous lab work is the subject of active research.

Two recent studies have reached different conclusions.

A study by AI company Scale AI and biosecurity nonprofit SecureBio found that when people with limited biology experience were given access to large language models, which is the type of AI behind tools like ChatGPT, they were able to complete biosecurity-related tasks, such as troubleshooting complex virology lab protocols with four times greater accuracy. In some areas, these novices outperformed trained experts. Around 90% of these novices reported little difficulty getting the models to provide risky biological information, such as detailed instructions on working with dangerous pathogens, despite built-in safety filters meant to block such outputs.

In contrast, a study led by Active Site, a research nonprofit that studies the use of AI in synthetic biology, found that AI help did not lead to significant differences in the ability of novices to complete the complex workflow to produce a virus in a biosafety laboratory. However, the AI-assisted group succeeded more often on most tasks and finished some steps faster, most notably on growing cells in the lab.

Hands-on work in the lab has traditionally been a bottleneck to translating designs into results. Even a brilliant study plan still depends on skilled human hands to carry out. That may not last, as cloud laboratories and robotic automation become cheaper and more accessible, allowing researchers to send AI-generated experimental designs to remote facilities for execution.

Responding to AI-driven biological risks

AI systems are now able to run experiments autonomously and at scale, but existing regulations were not designed for this. Rules governing biological research do not account for AI-driven automation, and rules governing AI do not specifically address its use in biology.

In the U.S., the Biden administration had issued a 2023 executive order on AI security that included biosecurity provisions, but the Trump administration revoked it. Screening the synthetic DNA that commercial providers make to ensure it cannot be misused to make pathogens or toxins remains mostly voluntary. A bipartisan bill introduced in 2026 to mandate DNA screening does not yet address AI-designed sequences that evade current detection methods.

The 1975 Biological Weapons Convention, an international treaty prohibiting the production and use of bioweapons, contains no provisions for AI. The U.K. AI Security Institute and the U.S. National Security Commission on Emerging Biotechnology have both called for coordinated government action.

The safety evaluations that AI labs run before releasing new models are often opaque and unsuited to capture real-world risk. Researchers have estimated that even modest improvements in an AI model’s ability to help plan pathogen-related experiments could translate to thousands of additional deaths from bioterrorism per year. Timelines for when these capabilities cross critical thresholds remain unclear.

The Nuclear Threat Initiative has proposed a managed access framework for biological AI tools, matching who can use a given tool to the risk level of the model rather than blanket restrictions. The RAND Center on AI, Security and Technology outlined a set of actions researchers could take to improve biosecurity, including improved DNA synthesis screening and model evaluations before release. Researchers have also argued that biological data itself needs governance, especially genomic data that could train models with dangerous capabilities.

Some AI companies have started voluntarily imposing their own safety measures. Anthropic activated its highest safety tier when it released its most advanced model in mid-2025. At the same moment, OpenAI updated its Preparedness Framework, revising the thresholds for how much biological risk a model can pose before additional safeguards are required. But these are voluntary, company-specific steps. Anthropic’s CEO, Dario Amodei, wrote that the pace of AI development may soon outrun any single company’s ability to assess the risk of a given model.

When used in a well-controlled setting, AI can help scientists quickly reach their research goals. What happens when the same capabilities operate outside those controls is a question that policy has not yet answered. Overreact, and talent and investment may move elsewhere while the technology continues advancing anyway. Underreact, and the risks of that technology could be exploited to cause real harm.

Subscribe now

I originally published this article at The Conversation. Read the original article here.

Turner, S. D. (2026, April 9). “AI can design and run thousands of lab experiments without human hands. Humanity isn’t ready for the new risks this brings to biology.” The Conversation. https://doi.org/10.64628/AAI.uhe4j4ce4

It’s a good paper, and I recommend reading it! I noted toward the end of the post that the customer screening side felt a bit undercooked. Tessa Alexanian, one of the paper’s coauthors, left a comment (thanks Tessa!) pointing me to additional work she and Sarah Carter had done on translating customer screening guidance into practical tools, and to a new preprint from Acelas et al. evaluating AI-assisted customer verification for synthetic nucleic acid screening.

Acelas, A., Palya, H., Flyangolts, K., Fady, P. E., & Nelson, C. (2026). Evaluating AI-Assisted Customer Verification for Synthetic Nucleic Acid Screening. bioRxiv 2026.02.27.708645; doi: https://doi.org/10.64898/2026.02.27.708645.

Here’s the problem the paper addresses: When someone orders a synthetic nucleic acid that matches a sequence of concern, the provider needs to verify that the customer is who they say they are and has a legitimate reason to order it. This legitimacy screening involves checking institutional affiliations, email domains, sanctions lists, and relevant publications or patents. It’s tedious, largely mechanical work, and the cost discourages adoption. Legitimacy screening runs roughly ten times more expensive per order than sequence screening alone.

Acelas et al. tested 5 LLMs (Claude Sonnet 4, Gemini 2.5 Pro, Grok 4, GLM 4.6, and MiniMax M2) on these verification tasks against a human baseline, using 41 customer profiles paired with simulated orders for sequences of concern. The best-performing model, Gemini 2.5 Pro equipped with bibliographic and sanctions APIs, achieved a 90% overall pass rate compared to about 80% for human screeners. Total cost per customer dropped from $14.04 for manual screening to $1.18 with AI assistance. For the information-gathering tasks alone (excluding human review of the final decision), the average was $0.23 per customer, roughly 50 times cheaper.

A couple things stood out. First, cost and performance were uncorrelated across models (Section 3.2 of the paper). The best model, Gemini 2.5 Pro, was also the second cheapest. Open-source models with lower per-token pricing lost their cost advantage through higher token consumption and more search queries. Second, giving models access to specialized tools (ORCID, Europe PMC, a sanctions list API) helped on most tasks but actually hurt on background work search, because models with API access performed fewer web searches and missed patents and news articles not indexed in academic databases (Section 3.1). Third, geographic variation in error rates. Chinese customers had notably higher missed-flag rates on email domain verification, largely because researchers there more often use personal rather than institutional email addresses (Section 3.3.1).

The authors are careful to note that the final ship-or-reject decision should stay with humans. AI handles the information gathering but a person decides what to do with it. This feels like the right framing, and as Tessa noted in her comment, the emergence of tools like Cliver (the screening API released alongside this paper), means providers increasingly don’t have to build this capability from scratch. That lowers the bar for adopting customer screening, which in turn makes it more reasonable to expect higher standards across the industry.

Subscribe now

Subscribe now

Chris Lu, et al., in Nature: Towards end-to-end automation of AI research. Sakana AI’s “AI Scientist” pipeline handles the full ML research loop: ideation, literature search, experiment design and execution, paper writing, and automated peer review. One of its manuscripts scored above the acceptance threshold at an ICLR 2025 workshop (which had a 70% acceptance rate, to be fair). Paper quality as judged by their automated reviewer tracks closely with foundation model capability, and with compute budget per paper, which tells you where this is headed even if the current output isn’t threatening anyone’s tenure case. For a quicker summary, read Sakana’s blog post.

Counterpoint: writes AI badly needs a dose of skepticism. Salzberg goes after DNA foundation models, arguing that their central claim (predict the effects of any mutation from sequence alone) is biologically implausible and largely unfalsifiable, two properties he knows well from years of writing about pseudoscience nonsense (homeopathy, accupuncture). Teams build ever-larger models first, then go looking for problems, which is backwards. The core critique of unfalsifiable prediction claims and Nature’s eagerness to publish them is hard to dismiss. See above.

: Transitioning to Being a PI in the Age of AI. A short and honest post about the asymmetry in how faculty and trainees experience the current AI moment in computational biology. Faculty are exhilarated because they’ve spent years developing the skill of evaluating analyses without doing them line by line; trainees are more ambivalent because they’re being asked to make that same transition in months rather than years or decades.

My SDS colleague Heman Shakeri released full materials for his Deep Learning Course here at UVA. A complete, openly licensed (CC BY 4.0) deep learning course from UVA’s School of Data Science, built for the online MSDS program (DS 6050) and public since Fall 2025. The 12-module sequence starts with NumPy-first implementations of MLPs and backpropagation, moves through CNNs, RNNs, encoder-decoder architectures, and the full attention/transformer stack, and finishes with ViTs, LoRA/QLoRA, and generative models including diffusion. Each module has lecture videos, notes, slides, and Colab assignments with unit tests. The syllabus lays out the pedagogical logic: three phases moving from from-scratch understanding to architectural depth to modern practice. Too much deep learning education lives in disconnected repos and YouTube playlists; having everything in one structured, reusable site with a clear arc is more valuable than any single component on its own.

Subscribe now

Carl Zimmer, NYT: How to Turn a Chicken Egg Into a Drug Factory. Neion Bio, a startup that emerged from stealth this week, is engineering chickens whose eggs produce pharmaceutical proteins, potentially replacing the Chinese hamster ovary (CHO) cell lines that currently dominate biologic drug manufacturing. The company claims 3,900 hens could meet global demand for Humira at a fraction of the cost of a CHO facility (Merck just broke ground on a $1B Keytruda plant, for comparison). Sven Bocklandt, Neion's chief scientific officer, was a colleague of mine at Colossal, where we worked on the dire wolf program together. Zimmer's writeup (great as usual) discussesthe history of how CHO cells became the default and why advances in primordial germ cell manipulation are finally making avian biomanufacturing viable.

New NIH Highlighted Topic: Advancing “Science of Science” Research to Understand and Strengthen the Biomedical Research Ecosystem. These are not NOFOs, but descriptions of scientific areas that NIH ICOs are interested in funding through existing parent announcements. This one encourages investigator-initiated applications on the “science of science,” the study of how the biomedical research ecosystem itself works. Topics include workforce retention, research capacity building, rigor and reproducibility, translation bottlenecks, and the economic returns of research investment.

Yet another new NIH Highlighted Topic: BRAIN Initiative: Advancing Human Neuroscience and Precision Molecular Therapies for Transformative Treatments. This one covers the BRAIN Initiative’s priorities in human neural circuit research, clinical neurotechnology, and precision molecular therapies (optogenetics, chemogenetics). 11 ICOs are listed as participating.

More NIH news: NOT-OD-26-064: Update of NIH Late Application Submission Policy and End of Continuous Submission. NIH is ending its Continuous Submission policy, which let PIs serving on review panels submit applications outside normal deadlines. Effective for due dates on or after May 25, 2026.

TIP Leadership Update. NSF's Erwin Gianchandani announces the retirement of Gracie Narcho, who served as deputy assistant director and directorate head for the Technology, Innovation and Partnerships directorate since its founding. Gianchandani credits Narcho with co-authoring the vision that became TIP before it had authorizing legislation, and with launching programs like the NSF Regional Innovation Engines and the I-Corps Hubs during a career spanning three decades and multiple NSF directorates.

Austin Dickey: How we chose Positron's Python type checker. Posit evaluated 4 open-source Python language servers (Pyrefly, ty, Basedpyright, Zuban) across features, correctness, performance, and ecosystem health, then chose Meta's Pyrefly as Positron's default. The most interesting section is the comparison of type-checking philosophies: ty follows a "gradual guarantee" where removing a type annotation never introduces an error, while Pyrefly infers types aggressively even in untyped code. Good overview of a space that's moving fast.

Mario Zechner: Thoughts on slowing the f*ck down. A year into production use of coding agents, Zechner argues that the compounding of small errors at machine speed, combined with agents’ inability to learn from mistakes and their low-recall search over large codebases, is producing unmaintainable messes far faster than human teams ever could. The prescription: treat agents as task-level tools with humans as the quality gate, write your architecture by hand, and set deliberate limits on how much generated code you accept per day.

Theo Roe: Why Learning R is a Good Career Move in 2026. A short, beginner-oriented pitch from Jumping Rivers (an R training company, so calibrate accordingly) making the case for R as a first language for data work. Nothing new for experienced practitioners, but a reasonable overview of where R still has a strong foothold: healthcare, pharma, government, academic research, and anywhere visualization and reproducible reporting are central. The honest caveat at the end is useful: if you want software engineering or large-scale production systems, you probably need Python.

at : Five Things: March 29, 2026: Anthropic temporary win, scheming, biodesign by LLM, White House advisors, Anthropic security.

Ryan Layer: What do I teach now?. Ryan has taught Software Engineering for Scientists at CU Boulder since 2019, and coding agents have forced him to rethink the whole course. In science, code is the method, so vibe coding is a reproducibility problem in addition to being a quality problem. He’s now rebuilding the class around open questions like who audits AI-generated analyses in ten years if no one learns to build from scratch.

The thought of my students building software by prompting and accepting the output without reading the code keeps me up at night. […] For science, where the code is the method, vibe coding is not an option.

at : Creating reproducible data analysis pipelines. A case against the “run everything from raw data” ideal of reproducibility. Claus argues that intermediate CSV files saved right before plotting are more durable than any end-to-end pipeline: pipelines break, Docker images rot,1 and students (and PIs!) lose afternoons rerunning everything to swap a violin plot for a boxplot.

rOpenSci News Digest, March 2026: dev guide, champions program, software review and usage of AI tools.

Joe Rickert: February 2026 Top 40 New CRAN Packages: AI, machine learning, biology, medical applications, physics, Buddhism, statistics, climate science, computational methods, data, surveys, ecology, time series, epidemiology, utilities, genomics, and visualization.

R Weekly 2026-W14: ggauto, alt text, scientific coffee.

Subscribe now

Max Kuhn: tabpfn 0.1.0. An R interface (via reticulate) to TabPFN, a pre-trained deep learning model for tabular prediction from PriorLabs (I wrote a this short summary of TabPFN last year). The model was trained entirely on synthetic data generated from complex graph models simulating correlation structures, skewness, missing data, interactions, and more. No fitting happens on your data; your training set primes an attention mechanism via in-context learning.

: Inside the NIH Forecast Graveyard. An accounting of NIH funding opportunities that were announced on grants.gov and then never published. Of 336 open forecasts, 205 have passed their promised posting dates with no explanation. The first wave of cancellations in April 2025 was keyword-driven (DEI, HIV, health disparities), but the later waves and the larger mass of silently expiring forecasts hit basic science, clinical infrastructure, and congressionally mandated programs like the BRAIN Initiative and Gabriella Miller Kids First. Ginexi, a former NIH insider, makes the dataset available for anyone to check.

Niko McCarty: Many Great Inventions Weren’t Made by “Serendipity”. Niko uses optogenetics as the central case for a broader argument: the breakthroughs we narrate as lucky accidents were usually preceded by years of deliberate preparation and systematic enumeration of possible solutions.

New papers & preprints:

• Towards end-to-end automation of AI research

• Toward next-generation machine learning and deep learning for spatial omics

• High-resolution metagenome assembly for modern long reads with myloasm

• The Age Illusion — Limitations of Chronologic Age in Medicine

• Accelerating coral assisted evolution to keep pace with climate change

• SNP calling, haplotype phasing and allele-specific analysis with long RNA-seq reads

• State AIDS Drug Assistance Programs’ Contribution to the US Viral Suppression, 2015–2022

• AlphaFold as a prior: experimental structure determination conditioned on a pretrained neural network

Subscribe now

Beal J. and Alexanian, T. 2025. “Creating Enforceable Biosecurity Standards for Nucleic Acid Providers,” Engineering Biology: e70003. https://doi.org/10.1049/enb2.70003.

The DNA synthesis industry faces a collective action problem: responsible companies that screen orders for biosecurity risks invest time and money into safety measures, while less scrupulous competitors can undercut them on price and speed by skipping these checks. This paper proposes a roadmap out of this dilemma through industry led development of enforceable standards.

Rather than waiting for governments to define every “sequence of concern”1 from scratch (something decades of advocacy has failed to produce) the authors propose that industry stakeholders work together to establish what they can agree on right now. Their Sequence Biosecurity Risk Consortium (SBRC) categorizes sequences into three groups: clearly benign (no flag), clearly dangerous (flag), and disputed or uncertain. By explicitly acknowledging the gray area, they reduce liability for responsible actors while enabling enforcement on sequences where consensus exists.

The International Gene Synthesis Consortium’s (IGSC) analysis of screening test sets shows that experts already agree on risk assessment for roughly 76% of sequences, with only 5% actively disputed. The remaining sequences are simply too poorly studied to categorize confidently. Rather than letting disagreement over that final 5-25% paralyze the entire system, the SBRC approach allows enforcement on the clear cases while research continues on the rest.

The paper outlines an ambitious three-phase roadmap for the United States, leveraging the May 2025 Executive Order on biological research safety.2 Phase 1 focuses on verifying minimal compliance, i.e. essentially proving that screening systems exist and function at a basic level. Phase 2 strengthens these requirements to ensure actual biosecurity effectiveness, incorporating comprehensive standards for both sequence and customer screening. Phase 3 aims for continuous improvement, including customer preapproval systems and anomaly reporting mechanisms similar to those used in aviation safety.

The approach attempts to align incentives at each stage of adoption. Early adopters (screening tool developers and major synthesis companies) are already participating because they see strategic value. The “early majority” will follow once certification provides competitive advantage and reduces regulatory uncertainty. Government enforcement becomes less risky once this critical mass has adopted standards voluntarily, and the early adopters actually benefit from enforcement that levels the playing field against laggards.

The authors emphasize that standards must be industry-led for several reasons. Governments lack the institutional capacity for systematic sequence-level risk assessment. Security-focused policymakers tend toward conservative standards that might stifle legitimate research. Synthesis is a global industry but regulation remains national: overly stringent early standards could encourage “provider shopping” across borders.

Some challenges remain underexplored in the paper. Customer screening standards are acknowledged as less mature than sequence screening, with decision-making about customer legitimacy involving “non-automated engagement between humans.” The roadmap’s timeline assumes coordination across synthesis providers, screening tool developers, certification bodies, and multiple government agencies. International harmonization receives appropriate emphasis but limited concrete detail.

Still, this represents a strong proposal for making DNA synthesis screening both effective and enforceable. By September 2025, the SBRC had already released its first “enforcement ready” standards covering pandemic-potential viruses and regulated toxins. Voluntary third-party testing became available in October 2025. The machinery is in motion.

The broader lesson extends beyond biosecurity: when voluntary guidelines fail to achieve universal adoption despite broad agreement on their necessity, the path forward may require industry to formalize what they can agree on, making enforcement both feasible and beneficial for responsible actors.

Subscribe now

The first, an introduction and overview of the report:

The second on the “if-then” approach to AI and Biosecurity (AIxBio):

The third on current capabilities across different levels of biological complexity.

Chapter 5 discusses how biological data are the leading indicator of emerging AI capabilities, but also reveals that high-quality, well-curated datasets are scarce outside of a few areas like protein sequences and structures. The report makes recommendations about treating biological databases as strategic national assets and creating AI-ready training datasets. This is less flashy than discussing what AI might do someday, but it’s arguably more important for both beneficial applications and risk management.

Subscribe now

One of the more subtle but critical insights from the National Academies report concerns data infrastructure. While much attention focuses on AI algorithms and computational power, the report emphasizes that high-quality biological datasets are the fundamental prerequisite for developing capable AI models. This realization has implications for both advancing beneficial applications and managing biosecurity risks.

The report notes that foundation models like AlphaFold demonstrate what’s possible when you combine powerful algorithms with comprehensive, well-curated datasets. AlphaFold was trained on approximately 200,000 experimentally determined protein structures deposited in the Protein Data Bank over decades of structural biology research. This combination of scale, quality, and standardization is exceptional in biology. Most biological data are fragmented, inconsistent in format, and lack the standardized contextual information needed for effective machine learning.

The scarcity of high-quality datasets has direct consequences for AI capabilities. The report’s assessment that current models cannot reliably predict viral transmissibility or design novel pathogens stems largely from the absence of appropriate training data. Viral sequence databases contain millions of sequences, but few are linked to detailed phenotypic measurements of infectivity, pathogenesis, or immune evasion. Even for well-studied viruses like SARS-CoV-2, comprehensive genotype-to-phenotype datasets exist for only a small fraction of possible mutations.

This data limitation is actually reassuring from a biosecurity perspective. It means that certain concerning capabilities remain out of reach not just because of algorithmic challenges, but because the requisite training data don’t exist. This is why the report identifies data availability as a leading indicator for emerging AI capabilities. Significant new datasets, particularly those linking genetic sequences to virulence or transmissibility, would signal potential capability advances worth monitoring.

The report argues that biological databases should be treated as strategic national assets, requiring robust data provenance mechanisms and careful curation. This recommendation serves dual purposes. High-quality, well-maintained datasets accelerate beneficial research in areas like drug discovery and vaccine development. Simultaneously, tracking what types of data are being generated helps with early detection of potentially concerning capability developments.

The report also emphasizes the need for substantial federal investment in biological data infrastructure. This includes funding efforts to create AI-ready datasets, supporting public-private partnerships for data generation, and maintaining existing repositories. The National Artificial Intelligence Research Resource pilot is cited as one potential model for connecting data resources with computational infrastructure.

The same data infrastructure that enables beneficial AI applications also provides the foundation for potential misuse. The report doesn’t shy away from this dual-use reality, but argues that the solution isn’t to restrict data broadly. Instead, it advocates for strategic investment combined with monitoring, ensuring that datasets critical for public health and biomedical research remain accessible while tracking indicators of concerning capability development.

The data infrastructure discussion reveals how biosecurity and scientific advancement are inextricably linked. Progress in both domains depends on the same fundamental resource: comprehensive, high-quality biological data properly managed and made accessible to legitimate researchers.

Subscribe now

Artificial General Intelligence Forecasting and Scenario Analysis. This report from Sarma, Bhatt, Jacob, and Steratore at the RAND Corporation Center on AI, Security, and Technology (CAST) synthesizes the current state of AGI timeline forecasting across expert surveys, prediction markets, compute-centric models, and trend extrapolation, finding that estimates have compressed toward the 2030s across independent methods while the infrastructure for validating those forecasts remains thin. The report describes a taxonomy of three “cruxes” driving disagreement (scaling sufficiency, diffusion speed, takeoff dynamics) and its argument that forecasts should be treated as scenario-structuring tools rather than point predictions. There’s an interesting disclosure here: the report was primarily drafted by LLMs (GPT-5.1, Gemini 3 Pro, Claude 4.5 Opus) with human direction and fact-checking, and the authors are candid about the recurring citation errors and factual inaccuracies they had to correct in the AI-generated drafts. Maybe the disclosure itself made me extra sensitive to all the LLM tells—it’s not X it’s Y, em dashes everywhere, and tricolons like this one.

The White House released their National Policy Framework on AI. The move that shouldn’t surprise anyone here is federal preemption of state AI laws: the framework explicitly calls for barring states from regulating AI development, which it frames as inherently interstate with national security implications. Beyond preemption, the document covers child safety (parental controls, age-assurance requirements), energy (residential ratepayers shouldn't bear data center electricity costs), IP (the administration believes training on copyrighted material is fair use but says courts should resolve it), and an anti-censorship provision aimed at preventing government from coercing AI providers on content. Notably absent: any new federal regulatory body for AI. The framework explicitly says Congress should not create one, relying instead on existing sector regulators and industry-led standards.

White House National Policy Framework On Artificial Intelligence

234KB ∙ PDF file

Download

Download

Technology Development for Genomics. A new NIH highlighted topic (open through August 2026) signaling investigator-initiated applications are welcome for new methods in sequencing, epigenomics, functional genomics, multiomics, spatial/temporal resolution, direct RNA/DNA sequencing, oligo synthesis, and high-throughput phenotyping of genetic perturbations. Participating ICOs include NHGRI, NIAID, NCI, NHLBI, and others. Not a specific funding opportunity announcement, but a signal of what program officers want to see come through the parent R01/R21 mechanisms. Particularly relevant context given the OMB hold on NIH spending just lifted this week; if new grants are about to start flowing again, knowing what the institutes are hungry for matters.

Posit: Working with Jupyter Notebooks in Positron.

UVA SDS Story of Us: Chapter 4: A Growing School. The latest chapter of UVA’s institutional history of its School of Data Science, covering 2024 to the present. The throughline is a school reaching institutional maturity: first PhD graduates hooded, a retooled master’s curriculum, an inaugural undergraduate class of 75 (now 125), a second building approved for design, and our faculty and staff growing from 30 to over 100. The chapter closes with the death of founding dean Phil Bourne earlier this month. Phil’s legacy is woven throughout, from insisting his office be no larger than any other faculty member’s to spearheading UVA’s Futures Initiative on AI in higher education. When asked what he was most proud of, Phil answered immediately: “Everybody.” A fitting capstone for a school that didn’t exist a decade ago. Phil’s memorial service was live streamed last week, and the recording is available here.

Richard McElreath released lecture recordings and notes from his course, “Statistical Rethinking.” The course focuses on logical and critical statistical workflows, from basic probability theory to causal inference to reliable computation to sensitivity.

Subscribe now

Rahul Garg: Context Anchoring. Part of an ongoing “Reducing Friction” series on martinfowler.com, this one argues that developers cling to long AI coding sessions not because they’re productive but because the decision context lives nowhere else. The proposed fix is a lightweight “feature document” that externalizes the why behind decisions (not just the what), functioning as a living Architecture Decision Record (ADR) that lets you kill a session and cold-start a new one in 30 seconds. If closing your chat window makes you anxious, your context isn’t properly anchored.

Five Things: March 22, 2026. Politicians, security breaches, OpenFold3, the AI “consciousness cluster”, dog’s cancer cured? As always, Matt provides a great recap of things happening in the world of AI/biosecurity and related topics over the past week, and it’s one of the weekly newsletters I look forward to the most.

Charlotte Wickham: Quarto 1.9. The highlights for me: Typst now supports book projects and article layouts with margin content, there’s experimental PDF accessibility compliance for both LaTeX and Typst, and websites can generate llms.txt output so LLMs can read your docs cleanly. More on accessibility via alt text, and automating it with Claude Code:

: The NOFO Graveyard. A former NIH program officer (22 years) responds to NIH Deputy Director Jon Lorsch's blog post defending the agency's reduction in Notices of Funding Opportunities. Ginexi compiled 14 years of NOFO data: NIH averaged ~724 NOFOs per year in 2023–2024, posted 120 in 2025 (an 83% decline, not the mandated 50%), and is on pace for far fewer in 2026. Of 271 opportunities forecasted in 2025, only 120 were ever posted. Lorsch's claim that fewer NOFOs wouldn't mean fewer funded applications is contradicted by NIH's own data showing R01-equivalent awards fell 20% in a single year. Ginexi argues the real administrative burden on program staff isn't NOFO management but politically motivated grant screening, and that centralizing NOFO approval through NIH/OD, HHS, and OMB transfers scientific priority-setting from domain experts to political appointees.

Proteina-Complexa. NVIDIA's new protein binder design framework unifies generative modeling and hallucination-style optimization under one roof, using flow-based latent generation with test-time compute scaling. Over a million binder candidates screened across 133 targets yielding 63.5% hit rates on PDGFR with picomolar affinities and the first-ever de novo designed carbohydrate binders. Self-generated sequences outperformed ProteinMPNN redesign across the board, which, if it holds up broadly, is a shift away from the two-stage generate-then-redesign paradigm.

R Weekly 2026-W13: What’s new in dplyr, Atmospheric simulation.

Simon Willison: Using Git with Coding Agents. A chapter from Willison's growing “Agentic Engineering Patterns” guide. You don’t need to memorize Git's more arcane commands anymore, but you do need to know what's possible so you can ask for it. Good bits on using git log to seed a fresh agent session with recent context, and on how agents turn git bisect from a tool most developers avoid into something you'd reach for casually.

Diego Oliveira, Qian Huang, Teresa Woodruff, and Brian Uzzi: How the 2025 NIH Grant Terminations Varied by Researchers' Demographic Groups. A PNAS brief report documenting who got hit by the 2025 NIH terminations: 2,291 grants, $2.45 billion rescinded, with 52% of allocated funds already spent at the time of cancellation. Early-career investigators and women were disproportionately affected. Among assistant professors, 59.8% of terminated projects were women-led. Women's grants were smaller on average ($940K median vs. $1.4M for men) but had a larger share of unspent funds at cancellation (57.9% vs. 48.2%), meaning more ongoing work was interrupted per dollar. Training and transition awards (F31s, F30s, T34s) were frequently cut. The authors estimate $6.29 billion in unrealized economic output using standard NIH multipliers, though they're careful to frame that as a benchmark rather than a verified loss.

Emily Mullin, WIRED: A Billionaire-Backed Startup Wants to Grow ‘Organ Sacks’ to Replace Animal Testing. R3 Bio, a Bay Area startup backed by Tim Draper and a Singapore longevity fund, is pitching the idea of growing brainless primate (and eventually human) “organ sacks” as replacements for lab animals. The concept: use stem cells and gene editing to grow organized organ structures that lack any brain tissue, making them incapable of sentience or pain. The near-term application is drug toxicity testing in monkeys at a time when US primate supply is constrained after China’s 2020 export ban. The longer-term ambition is growing human organ sacks as a source of transplant organs. The company says it’s currently only working in monkey cells, though a job posting seeks a veterinarian in Puerto Rico to implant embryos in nonhuman primates. The science is plausible according to UC Davis stem cell biologist Paul Knoepfler, but everything about this is still highly theoretical, and a Stanford bioethicist notes the “yuck factor will be strong.”

Alexis Gallagher and Rens Dimmendaal: So Where Are All the AI Apps?. If AI coding tools make developers 2x (or 10x, or 100x) more productive, you should be able to see it in PyPI package creation and update rates. The authors looked, and mostly didn’t see it. Total package creation shows no inflection at ChatGPT’s release. Update frequency for the top 15,000 packages shows a modest secular trend that predates AI tools. The one place a clear >2x effect does appear: popular packages about AI, which jumped post-ChatGPT compared to popular non-AI packages. The authors’ two candidate explanations are that AI-tool builders are the ones most skilled at using AI to build (a skill effect), or that the flood of money into AI is simply paying for more work on AI packages (a funding effect). Probably both, but the data can’t distinguish them. Either way, the Cambrian explosion in all software hasn’t materialized; the measurable effect is concentrated in the AI ecosystem building tools for itself.

and , at : State of AI in the Commonwealth: Teaching & Learning with AI. Part of a rolling series from UVA’s Center for Management IT drawing on nearly 200 sources to map AI’s impact in Virginia. This installment focuses on education, and the most useful data point comes from a McIntire study of 356 students across 6 conditions: the best learning outcomes came from group work paired with a custom RAG tutor, while generic chatbots performed worst. The broader numbers frame the context: the global AI-in-education market jumped 46% in a single year (to $7.57B in 2025), 61% of faculty report using AI in teaching, but 88% of those describe their use as minimal.

And a recommendation: My SDS colleague YY Ahn maintains a public wiki that sits somewhere between a research notebook, a tool log, and a blog. It’s updated frequently and worth browsing if you’re into academic workflows, network science, or the ongoing project of making AI tools actually useful for research. A few recent entries: Claude Scholar plugin, a set of Claude Code skills for academic work (arXiv metadata fetching, BibTeX from DOIs, math verification via SymPy, pre-submission LaTeX checks); How I Use Claude Code, a living document on his terminal/editor/dictation setup; paper notes on topics ranging from LLMs inferring political alignment from conversations to community-centric citation dynamics; and a separate urbanism thread with entries on topics like the true cost of car ownership (read this one!). The whole thing runs on a kind of accumulating-notes-in-public model that I bookmark whenever I find them.

Subscribe now

New papers & preprints:

• Generalist biological artificial intelligence in modeling the language of life

• LLMs Can Infer Political Alignment from Online Conversations

• A harmonized benchmarking framework for implementation-aware evaluation of 46 polygenic risk score tools across binary and continuous phenotypes

• LazySlide: accessible and interoperable whole-slide image analysis

• How LLMs Distort Our Written Language

• ProteinMCP: An Agentic AI Framework for Autonomous Protein Engineering

• LLM agents for biological intelligence across genomics, proteomics, spatial biology, and biomedicine

Subscribe now

Emil Hvitfeldt published a blog post with a Claude Code skill to create alt text in Quarto documents.

I have some personal skills in ~/.claude/skills and I added this one to it. I.e., I added a file at ~/.claude/skills/write-alt-text/SKILL.md with the skill text Emil provides in his blog post.

After writing a Quarto doc, I can just ask Claude Code, “Use the write-alt-text on the file @my-doc.qmd” or I can directly invoke the new slash command available to me with the skill now installed: /write-alt-text @my-doc.qmd.

Here’s what a demo Quarto doc looks like before with some plots using the penguins, mtcars, and diamonds data.

---
title: "Testing alt text"
format: html
---

```{r}
#| message: false
library(ggplot2)
```

## Penguins: Bill Length vs. Flipper Length

```{r}
ggplot(
  penguins,
  aes(x = flipper_len, y = bill_len, color = species)
) +
  geom_point() +
  labs(
    title = "Penguin bill length vs. flipper length",
    x = "Flipper length (mm)",
    y = "Bill length (mm)",
    color = "Species"
  ) +
  theme_minimal()
```

## mtcars: Fuel Efficiency by Cylinders

```{r}
ggplot(mtcars, aes(x = factor(cyl), y = mpg, fill = factor(cyl))) +
  geom_boxplot(show.legend = FALSE) +
  labs(
    title = "Fuel efficiency by number of cylinders",
    x = "Number of cylinders",
    y = "Miles per gallon (MPG)"
  ) +
  theme_minimal()
```

## Diamonds: Cut vs. Price

```{r}
ggplot(diamonds, aes(x = cut, y = price, fill = cut)) +
  geom_violin(show.legend = FALSE) +
  scale_y_log10(labels = scales::dollar) +
  labs(
    title = "Diamond price distribution by cut",
    x = "Cut",
    y = "Price (log scale)"
  ) +
  theme_minimal()
```

After running that new skill, Claude Code confirms what it did:

And my new quarto doc looks like this, now with alt text on each chunk:

---
title: "Testing alt text"
format: html
---

```{r}
#| message: false
library(ggplot2)
```

## Penguins: Bill Length vs. Flipper Length

```{r}
#| fig-alt: |
#|   Scatter chart of penguin bill length (mm) on the y-axis versus flipper
#|   length (mm) on the x-axis, with points colored by species (Adelie,
#|   Chinstrap, Gentoo). The three species form distinct clusters. Gentoo
#|   penguins tend to have the longest flippers but moderate bill lengths,
#|   while Chinstrap penguins have shorter flippers but relatively long bills.
#|   Across all species, longer flippers are generally associated with longer
#|   bills.
ggplot(
  penguins,
  aes(x = flipper_len, y = bill_len, color = species)
) +
  geom_point() +
  labs(
    title = "Penguin bill length vs. flipper length",
    x = "Flipper length (mm)",
    y = "Bill length (mm)",
    color = "Species"
  ) +
  theme_minimal()
```

## mtcars: Fuel Efficiency by Cylinders

```{r}
#| fig-alt: |
#|   Box plot of fuel efficiency (miles per gallon) by number of cylinders
#|   (4, 6, or 8) for cars in the mtcars dataset. Cars with 4 cylinders have
#|   the highest median MPG, followed by 6-cylinder cars, with 8-cylinder cars
#|   showing the lowest and most variable fuel efficiency. The relationship is
#|   clearly monotonically decreasing: more cylinders corresponds to worse fuel
#|   economy.
ggplot(mtcars, aes(x = factor(cyl), y = mpg, fill = factor(cyl))) +
  geom_boxplot(show.legend = FALSE) +
  labs(
    title = "Fuel efficiency by number of cylinders",
    x = "Number of cylinders",
    y = "Miles per gallon (MPG)"
  ) +
  theme_minimal()
```

## Diamonds: Cut vs. Price

```{r}
#| fig-alt: |
#|   Violin plot of diamond price (log scale, shown in dollars) by cut quality
#|   (Fair, Good, Very Good, Premium, Ideal) for the diamonds dataset.
#|   Counterintuitively, higher cut grades do not correspond to higher prices:
#|   Ideal-cut diamonds show a distribution concentrated at lower prices, while
#|   Fair and Premium cuts have heavier tails toward higher prices. This likely
#|   reflects confounding with carat size, as lower-cut diamonds tend to be
#|   larger. All distributions are right-skewed on the log scale.
ggplot(diamonds, aes(x = cut, y = price, fill = cut)) +
  geom_violin(show.legend = FALSE) +
  scale_y_log10(labels = scales::dollar) +
  labs(
    title = "Diamond price distribution by cut",
    x = "Cut",
    y = "Price (log scale)"
  ) +
  theme_minimal()
```

Subscribe now

Workshops for Ukraine

Back in January I gave a 2-hour workshop on R Package Development in Positron, covering package development with devtools, usethis, covr, GitHub actions, pkgdown, writing functions, documentation, exported/installed package data, programming with dplyr, and writing unit tests with testthat and lots of help from Claude via Positron Assistant.

All the materials are online, and the workshop recording is below.

• Rendered pkgdown website: https://stephenturner.github.io/rpkgdemo/

• Full tutorial: https://stephenturner.github.io/rpkgdemo/articles/rpkgdemo

• Source code for the built package: https://github.com/stephenturner/rpkgdemo

Subscribe now

Related posts:1

Share

That’s my typical red line. However, some writing exists purely as bureaucratic packaging, and I’d rather spend zero minutes on it.

Cover letters for journal submissions are the obvious case IMHO. There’s really no reason for these at all. You state the title, summarize the contribution in two sentences, confirm it’s not under review elsewhere, and move on.5 Other entries in this category:6 data management plans, facilities and equipment descriptions (given the specs you already have on file), budget justifications given a budget spreadsheet and a research strategy (“The PI requests 1.2 calendar months of effort to provide scientific direction and oversight and blah blah blah...”). And so on.

For these tasks I do use AI to write my first draft and edit as needed. But default LLM output usually reads like default LLM output. The tells are well-documented at this point: filler transitions, “the ever-evolving landscape,” em dashes on every line, self-answered rhetorical questions, it’s not X it’s Y, let that sink in. Let an LLM write your cover letter unedited and the reviewer (if they’re human) can probably tell.

I built a Claude skill called deslop to strip these patterns out. It teaches Claude to identify and remove AI writing tropes across about 30 categories, scores output on a 50-point rubric (directness, rhythm, trust, authenticity, density), and includes before and after examples for scientific writing, blog posts, and grant narratives. The deslop GitHub repo has the full catalog, along with credit where it’s due (see the README).7

Claude Skill: Deslop

As a demo, I had Claude write a cover letter for a paper that I’ve already published (Nagraj et al. 2025, PLOS ONE, blog post here), as if I’m submitting it to PLOS ONE, with every AI trope cranked up. It scored 8/50. Then I ran the deslop skill on the same content. The rewrite scored 43/50. Specific numbers replaced vague claims, filler transitions disappeared, and it read like a person wrote it for a purpose rather than a model producing plausible cover-letter-shaped text.

The skill works on any chat on claude.ai, Claude Code, or the desktop app. Clone the repo, install the skill, ask it to deslop or write like a human, and the patterns get filtered at generation time instead of in post-hoc editing.

Subscribe now

Here’s the before. Yes, it’s utterly ridiculous (yet reads like most X/LinkedIn posts these days).

It’s worth noting that methods for reviewing epidemiological signals serve as a critical component of the ever-evolving landscape of public health surveillance. Here’s the thing: implausible forecasts and faulty surveillance data don’t just bias ensemble models — they erode trust in the very systems designed to protect public health. We are pleased to submit our manuscript, “PLANES: Plausibility Analysis of Epidemiological Signals,” for consideration in PLOS ONE. In this paper, we delve into a novel, robust approach for assessing forecast plausibility and surveillance data integrity — one that is multi-dimensional, flexible, and open-source. The result? A scoring system that fundamentally reshapes how modelers and public health stakeholders navigate the challenges of data quality review. Our analysis of 2022-23 FluSight forecasts highlights that forecasts flagged by multiple PLANES components were remarkably less accurate on average, underscoring the approach’s practical utility for operational forecasting — from pre-submission quality control to ensemble weighting to surveillance system monitoring.

Despite these challenges in the current state of plausibility assessment, we believe PLANES fills an important gap. Not a replacement for human judgment. Not a fully automated system. A decision-support tool that quietly reduces the burden of manual review while leveraging historical data to codify nuanced expectations about signal behavior. It bears mentioning that the accompanying rplanes R package — available on CRAN and GitHub — makes the approach accessible to practitioners across federal, state, and local jurisdictions. Importantly, we have already piloted rplanes in operational settings, including the 2023-24 FluSight season, reflecting broader trends toward more standardized and reproducible forecasting workflows. In summary, we believe this work represents a meaningful contribution to the growing body of literature on epidemiological data integrity, and we look forward to the opportunity to share it with the PLOS ONE readership. Let’s dive in.

Here’s how it scored:

Here’s what the text looks like after running /deslop.

We submit “PLANES: Plausibility Analysis of Epidemiological Signals” for consideration in PLOS ONE. The manuscript describes a multi-dimensional scoring method for assessing the plausibility of infectious disease forecasts and surveillance data, delivered through an open-source R package (rplanes, available on CRAN). PLANES uses historical observed data to seed location-specific baseline characteristics, then evaluates signals across seven components, each producing a binary flag that rolls up into an ordinal score. In a retrospective analysis of over 29,000 FluSight forecasts from the 2022-23 season, we found that forecasts with two or more components flagged had significantly higher weighted interval scores (r = 0.267, p < 0.0001), indicating worse predictive accuracy.

No codified plausibility heuristics currently exist for epidemiological signals; review, where it happens at all, is ad hoc and undocumented. PLANES addresses this gap with a method that works for both forecasted and observed signals at daily, weekly, or monthly resolution. We piloted rplanes operationally during the 2023-24 FluSight season to review both our submitted forecasts and the NHSN surveillance data used to train our models. The package is designed for human-in-the-loop use: it flags signals for review rather than censoring them automatically. We believe this work will be of interest to PLOS ONE readers working in disease forecasting, surveillance data quality, and public health decision support.

And the score:

Get the skill here: https://github.com/stephenturner/skill-deslop.

Subscribe now

and at : AI Exchange @ UVA Podcast with Piers Gelly and T. Kenny Fountain, faculty members in the University of Virginia’s Department of English, discussing how generative AI is reshaping the teaching and practice of writing. Piers Gelly wrote a fantastic essay last year, “What Happened When I Tried to Replace Myself with ChatGPT in My English Classroom” and if you haven’t read it yet, I recommend it.

How bioRxiv changed the way biologists share ideas. A Nature news piece pegged to a new analysis from the openRxiv team marking bioRxiv's first 13 years: 310,000+ preprints posted, 4 million downloads a month, and 80% of preprints eventually published in journals. The survey data on why researchers post (visibility, priority, feedback) is a useful corrective to the outdated assumption that preprints are a workaround for researchers with something to hide.

More from and at : Exploring AI Readiness in Virginia. Big Question: Are we preparing every Virginian, not just technologists, to thrive in an AI-enabled economy? AI literacy is still uneven and demand is rising faster than readiness.

Charlie Marsh, Astral: Astral to Join OpenAI. Astral, the company behind Ruff, uv, and ty (tools that have become close to default infrastructure for modern Python development), is being acquired by OpenAI and will join the Codex team. Marsh writes about this as a natural extension of the mission to make programming more productive, now at the frontier where AI meets developer tooling. OpenAI says it will continue supporting the open source tools, and Astral says they'll keep building in the open. The obvious question for anyone who depends on uv or Ruff is whether “continue supporting” holds up once Astral's attention is split between maintaining open source Python tooling and building product for OpenAI's coding agent. Worth watching closely.

Job ad: arXiv CEO. arXiv is becoming an independent (nonprofit) company and they're looking for a CEO. Salary $300,000.

RAND Europe: Cost-Benefit Analysis for Synthetic Nucleic Acid Screening in the European Union. The first EU-level quantitative CBA comparing three policy pathways for DNA synthesis screening: voluntary guidance, conditionality tied to research funding, and mandatory regulation under the proposed Biotech Act. The headline finding is that mandatory screening yields about €4.6 billion in average annual net benefits over ten years, with a 6:1 return on costs, driven largely by the non-linear relationship between screening coverage and risk reduction. The report was partly funded by Sentinel Bio, and the cost estimates are described by the authors themselves as conservative, with real-world screening costs potentially 75% lower than modeled.

Schedule Released: 2026 Applied Machine Learning Conference. Friday, April 17 features keynotes from David Luebke (Co-founder of NVIDIA Research) and Vicki Boykis (founding ML engineer at Malachyte), plus four tracks of talks from speakers at organizations like Anthropic, Amazon, Microsoft, McAfee, S&P Global, DARPA, IBM, PagerDuty, and many more, as well as leading universities including UVA, Virginia Tech, Duke, and NC State. Saturday, April 18 is all hands-on, with three tracks of 90-minute tutorials that dive deeper into topics such as building production-grade AI agents, agentic search, knowledge graph construction, MCP integration with real databases, probabilistic programming, physics-informed neural networks, and more. Register here (early bird pricing ends today).

Register

: Five Things: March 15, 2026: Anthropic v. Pentagon latest, human v. LLM text, cybercrime updates, protein production paper, reviews of LLM bio uplift.

Bio-attribution Challenge. DARPA's Biological Technologies Office is running a two-round computational competition with $180K in prizes for teams that can detect and attribute engineered pathogens at petabyte scale in near real-time. No wet lab work involved: all data is LLNL-curated synthetic sequences run in a government-controlled environment, which makes this an accessible entry point for bioinformatics and ML teams who want to work on biosecurity problems without navigating biosafety infrastructure.

: Reasons to be pessimistic (and optimistic) on the future of biosecurity. A long, interview-dense survey of the biosecurity landscape covering the disappearing DNA synthesis chokepoint, the surprisingly underappreciated risk of agricultural bioterrorism, and the funding cliff facing BARDA and CDC under the FY26 budget. Abhishaike takes the AI uplift threat seriously while also citing the Active Site RCT (my own blog post on this here) showing that frontier models even with safety classifiers off gave novices no statistically significant advantage over a Google search in a wet lab. Well worth the long read for anyone interested in this space. Pairs naturally with the DARPA bio-attribution challenge above.

Sara Altman + Simon Couch / Posit blog: 2026-03-13 AI Newsletter.

Simon Willison: What is agentic engineering? (It’s not just vibe coding).

Julie Bedard et al: When Using AI Leads to "Brain Fry". A BCG research team surveyed 1,488 U.S. workers and found that intensive AI oversight (not AI use per se) drives acute cognitive fatigue, while using AI to replace repetitive tasks actually lowers burnout scores. The distinction between burnout (emotional exhaustion) and “brain fry” (attentional overload from managing agents, c.f. brain rot) is the useful, even if the prescription (e.g. clearer org communication, holistic job redesign) reads like standard change management advice.

"Shall I implement it?" User says no. OpenClaw thinks for a bit, and decides that the user actually meant yes, do it without asking.

New NBER working paper: Attention (And Money) Is All You Need: Why Universities Are Struggling to Keep AI Talent. “The top 1% of publishing industry scientists now earn $1.5 million more annually than comparable academics, a fivefold increase since 2001.” In one image:

David Adam, Nature Medicine: The AI Co-Scientist Is Here. A survey of where AI-driven hypothesis generation stands in biomedicine right now, anchored by Google’s AI co-scientist tool and Insilico Medicine’s rentosertib (possibly the first fully AI-discovered drug to reach phase 3). An interesting tension in the piece is between the flashy anecdotes (Gary Peltz’s drug repurposing hits, José Penadés’s lab having its unpublished findings independently rediscovered by the model) and the less-flashy caveats: Google won’t say how many testers got less impressive results, the Stanford ideation-execution gap study found human ideas still won after implementation, and LLM-based tools are limited to open-access literature.

R Weekly 2026-W12: Sharing across shiny modules, F1 data.

Isabella Velásquez & Libby Heeren, Posit Data Science Lab: How to use Positron’s GitHub integration.

: Your Bioinformatics Tools Need to be AI-Ready. A sharp argument that bioinformatics tools should be designed from the start to emit rich, structured metadata suitable for ML training, not just the summary QC metrics we’ve been building for human eyeballs. Homer draws on his experience building Picard and fgbio to make the case that the gap between “useful QC output” and “ML-ready feature vectors” is small but real, and that closing it retroactively across thousands of samples is far more painful than doing it right the first time. Nils lays out some design principles:

Observable. Emit structured, semantically rich metadata at every step. Not just final answers, but the evidence trail. Per-read, per-site, per-family, per-molecule. Use well-defined tags, clear column headers, and machine-readable formats.

Composable. Consistent, well-documented interfaces that both humans and AI agents can discover, chain, and reason about. Standard formats. Clear help text. Predictable behavior. Outputs parseable by a post-doc who happens to be an LLM.

Trainable. Feature-rich representations that serve as training data for downstream ML. Emit the features, not just the conclusions.

Subscribe now

New papers & preprints:

• Supporting Workflow Reproducibility by Linking Bioinformatics Tools across Papers and Executable Code

• Sassy2: Batch Searching of Short DNA Patterns

• GeneExt: a gene model extension tool for enhanced single-cell RNA-seq analysis

• Accelerating minimap2 for whole-genome alignment

• Eduomics: a Nextflow pipeline to simulate -omics data for education

• The archaeal roots of eukaryotic life

• Finding low-complexity DNA sequences with longdust

• OmicClaw: executable and reproducible natural-language multi-omics analysis over the unified OmicVerse ecosystem

Subscribe now

Subscribe now

While reading the BDL proposal in the Science paper, I kept thinking: how would this have played out during the early days of the COVID-19 outbreak?

I imagine the BDL framework may have classified much of the early SARS-CoV-2 research data under restricted tiers. Functional studies mapping spike protein mutations to ACE2 binding affinity, cell entry efficiency, and antibody evasion are the kind of genotype-to-phenotype data the framework targets at BDL-3 or even BDL-4. Deep mutational scanning of spike receptor binding, pseudovirus neutralization assays, and serial passage experiments generating data on transmissibility and immune escape would all potentially trigger higher-tier controls.

To their credit, the authors anticipate this. The paper explicitly calls for exceptions during disease outbreaks, when real-time data are especially important for public health.

We recommend that governments establish diverse expert panels [that could] provide for relevant exceptions, such as lifting restrictions on relevant data during a disease outbreak, when real-time data are especially important to inform public health measures.

But the early months of a pandemic are exactly when this exception would be hardest to implement well. In January 2020 the scientific community didn’t yet know SARS-CoV-2 would become a pandemic. The first genome sequence was shared openly on January 10, before most governments had begun emergency responses. Within weeks, structural studies and functional characterization of ACE2 binding were underway globally. Vaccine design efforts at Moderna and BioNTech were already using these data to select candidates before the WHO declared a pandemic in March.

A BDL framework as proposed here would have required an expert panel to decide, in real time, whether an emerging respiratory virus of uncertain scope warranted lifting data restrictions, before the scale of the threat was clear, through committees that may not yet have convened. Even modest delays during this period could have slowed vaccine development, diagnostics, and the epidemiological modeling that informed early public health responses.

None of this is a criticism of the BDL framework. It’s thoughtful, and it’s concisely described in just 4 pages (really, it’s good! Just go read it). The authors are spot on that proactive governance beats reactive restrictions imposed in a crisis.

The COVID-19 experience illustrates a potential implementation challenge: the moments when open data access matters most are also the moments when the biosecurity case for restricting it is strongest. The BDL framework proposed here or one like it will need outbreak exception mechanisms that are fast and that don’t depend on bureaucratic deliberation while a deadly virus is spreading.

Bloomfield, ... & Pannu, J. (2026). Biological data governance in an age of AI. Science, 391(6785), 558-561. DOI: 10.1126/science.aeb2689.

⁂

The senior author on this paper, Jassi Pannu, was interviewed last week on the Cognitive Revolutions podcast. It’s a really nice interview, full of insightful commentary from Dr. Pannu on the current state of biosecurity and the proposed BDL framework. You can watch it here:

Or listen to it here:

Subscribe now

Claude Code can erode your work-life balance if you’re not careful. I’m generally pretty good about turning off and not working at home in the evenings.1 But Claude Code has opened up a loophole in my own discipline. Typing a few prompts and walking away doesn’t feel like working. It’s not like sitting down to write code for an hour. You fire off a planning prompt, go do something else, come back in 20 minutes to review the output, nudge it in a new direction, and walk away again.2

The problem is that you can do this indefinitely. I’ve caught myself kicking off plan-and-implement cycles in the evenings, trying to squeeze everything I can out of my four-hour session window.3 A prompt or follow-up takes 30 seconds. Each response might take 30 minutes. In between I’m technically not at my desk. But my laptop is open, my attention drifts back to it, and my evening has become an extension of my workday with the “just one more prompt.”

Claude Code genuinely does make me more drastically more productive. It’s a helpful tool, and I’ll continue to use it. But I’m starting to think the ease of it is what makes it hard to put down.

Subscribe now

: The Shape of the Thing. Mollick’s latest state-of-play piece argues we've crossed into a new phase where AI agents take on hours of work autonomously, and that recursive self-improvement an explicit item on the roadmap of every major lab. A single chaotic week in late February as a preview of the near future is useful: stock market swings, mass layoffs attributed (dubiously) to AI, and a public Pentagon-Anthropic conflict.

Anthropic launches the Anthropic Institute. Anthropic is consolidating its Frontier Red Team, Societal Impacts, and Economic Research groups under a new umbrella called the Anthropic Institute (anthropic.com/institute), led by co-founder Jack Clark. The announcement reads like a company that believes the next two years will be decisive and wants a dedicated apparatus for telling the world what it’s learning. Anthropic Institute has an open position for an analyst ($295,000 - $345,000 annually).

Niko McCarty: eLife Fallout. An account of the internal tensions that led to Michael Eisen’s firing as EIC of eLife, and what happened to the journal’s radical publishing reforms afterward.

Armin Ronacher: AI And The Ship of Theseus. Interesting line of thought on what AI coding agents mean for re-implementing software for the purposes of re-licensing (e.g. GPL to MIT). See also the discussion on HN.

and : State of AI in the Commonwealth - Global & VA Economic Impact. A new report from the UVA McIntire’s Center for Management IT surveying AI's economic footprint in Virginia, from data center dominance in Northern Virginia to workforce displacement risk and energy grid strain. The policy recommendations are fairly standard, but with a Virginia-specific framing. Read the full report here.

Måns Thulin: Why I don’t use tidymodels. Strong critique of tidymodels: its bootstrap p-values and confidence intervals can contradict each other, and simulations show its bootstrap test actually performs worse than a standard t-test under non-normality. Also missing LOOCV and residual resampling.

Brookings: The train has left the station: Agentic AI and the future of social science research.

New NIH Highlighted Topic: Technology Development for Genomics.

Expanding our analysis of biological AI models. Epoch AI’s expanded database of 1,100+ biological AI models reveals a stark safeguards gap: only 3.2% have any documented safeguards, and just 2.5% underwent pre-release risk assessments, with frontier LLMs accounting for the majority of both. The field remains overwhelmingly academic, dominated by protein engineering and small molecule design, with most models building on a handful of foundation models (ESM-2 chief among them) and progress appearing more data-constrained than compute-constrained.

Joe Cheng, Nick Rohrbaugh, and Sara Altman (Posit): Introducing AI in RStudio. Posit is embedding an AI agent (Posit Assistant) and inline code completions (Next Edit Suggestions) directly into RStudio, backed by a zero-data-retention agreement with Anthropic, at $20/month. The framing is notably honest about their initial skepticism and why they changed their minds. Worth reading if you’re trying to think through when AI assistance in data analysis is and isn't appropriate.

Isabella Velásquez: posit::glimpse() Newsletter – March 2026. This one’s packed full of goodies. Posit AI, built-in PDF viewer in Positron, mirai, Quarto 1.9, Pointblank for Python, learning & community, Claude Code at next week’s Data Science Lab.

Subscribe now

Matthew Carter: The Legibility Problem. A thoughtful piece on what happens when AI systems produce scientific discoveries that humans can’t understand or interpret, framed around the chess analogy (we trust engines even when their moves make no sense to us), and what it would take to build the translation infrastructure between AI-generated knowledge and human science.

Julia Silge: Outgrowing your laptop with R and Positron. More details at slides available on the first post on the new Positron blog.

: How Do Scientists Use Claude Code? A clever use of Claude Code’s default co-authorship commits and ORCID-GitHub profile links to measure AI coding tool adoption among ~16,000 computationally active scientists: about 2.1% as of February 2026, with a U-shaped adoption curve by career stage and surprisingly uniform rates across scientific fields. The sample size is small enough that the breakdowns are directional rather than definitive, but this is the kind of empirical work the AI-for-science discussion has been missing. Code for this analysis is here on GitHub.

: Five Things: March 8, 2026. America vs. Anthropic, Alibaba paper, Ginkgo splits up, biosecurity screening paper, GPT-5.4. I was going to do a short blog post on that biosecurity screening paper, but Matt did so, so I’m not going to. From Matt’s post:

In October, Science published an article led by Bruce Wittmann from Microsoft (and several DNA synthesis companies) demonstrating that AI can design dangerous sequences that evade those companies’ security screening protocols. This past week he and his team followed-up on this work with a new preprint: “The Limits of Sequence-Based Biosecurity Screening Tools in the Age of AI-Assisted Protein Design.” The setup: take dangerous proteins (officially called “proteins of concern”) and use AI to generate “reformulated synthetic homologs”: sequences that do the same thing as the dangerous protein but look different enough to potentially evade screening. This time, they fragmented the sequences into smaller segments (the way someone might actually order them from a synthesis company to avoid detection) and run them through four major biosecurity screening software (BSS) tools.

The results are a mixed bag. Two of the four tools were capable of “robustly detecting fragments as short as 50 nucleotides” — actually exceeding the requirements in the U.S. Framework for Nucleic Acid Synthesis Screening. The other two improved with upgrades. All in all, this is very good news! Of course AI technology is improving constantly, and at some point it might be able to design novel proteins from scratch. Once that occurs, we will need to step up the defensive AI tech to use the same methods to screen novel compounds as one might do to design them. Hopefully someone is working on that.

: The Real Cost of Running AI. A thorough walkthrough of inference economics, from the FLOPs-per-attention-head level all the way up to the crossover math on edge vs. cloud deployment. Decode is memory-bound, not compute-bound, and that structural mismatch only gets worse with each GPU generation. Long, but helps to understand why output tokens cost more than input tokens, why the KV cache is the real scaling bottleneck, and why every novel architecture is ultimately trying to kill the same quadratic term.

and : Towards a Science of AI Agent Reliability. Borrowing frameworks from aviation and nuclear safety, this post decomposes agent reliability into 12 metrics across 4 dimensions (consistency, robustness, calibration, safety). ~2 years of rapid capability gains have produced only modest reliability improvements. Agents that can solve a task often fail on repeated attempts under identical conditions, and their self-reported confidence is barely better than chance. A good frame for understanding why benchmark progress hasn't translated into economic impact as fast as the vibes would suggest.

R Weekly 2026-W11: Claude skills for R users, beeswarm charts in R, AI in RStudio.

Brian Naughton: How to Design Antibodies: A step-by-step guide to making de novo binders.

Tool            Commercial  Full Ab  VHH  Mini-binder  Peptide
-----------------------------------------------------------------
BoltzGen            ✅        ✅      ✅       ✅           ✅
Germinal            ❌        ✅      ✅       ❌           ❌
mBER                ✅        ❌      ✅       ❌           ❌
RFantibody          ✅        ✅      ✅       ❌           ❌
Mosaic              ✅        ✅      ✅       ✅           ✅
BindCraft*          ❌        ❌      ❌       ✅           ✅

: PLMs are mostly just memorizing stuff.

Subscribe now

New papers & preprints:

• Rubrics, Not Vibes: Structured AI Integration as Quality Control for Peer Review

• snputils: A High-Performance Python Library for Genetic Variation and Population Structure

• RNA-seq analysis in seconds using GPUs

• geneSTRUCTURE: A Modern Platform for Visualization of Gene Structures

• The Limits of Sequence-Based Biosecurity Screening Tools in the Age of AI-Assisted Protein Design

• Artificial intelligence-enabled multi-scale virtual cell: perspective, challenges, and opportunities

• Gene regulatory networks: from correlative models to causal explanations

• Towards a Science of AI Agent Reliability

• Bringing the genetically minimal cell to life on a computer in 4D: Cell

Subscribe now

Subscribe now

Measuring Mid-2025 LLM-Assistance on Novice Performance in Biology

The first paper:1

Hong SZ et al. Measuring Mid-2025 LLM-Assistance on Novice Performance in Biology, arXiv:2602.16703. Preprint, arXiv, 18 Feb. 2026. https://doi.org/10.48550/arXiv.2602.16703.

Hong and colleagues at Active Site ran a pre-registered, investigator-blinded randomized controlled trial in a BSL-2 laboratory over eight weeks, enrolling 153 undergrads with minimal prior lab experience. Half had access to frontier LLMs from Anthropic, OpenAI, and Google DeepMind; the other half had the internet only. Participants worked independently to complete tasks modeling a viral reverse genetics workflow: cell culture, molecular cloning, virus production, and RNA quantification. The headline result was that LLM access did not significantly improve completion of the full workflow.

Only about 5% of either group made it through. The more nuanced picture suggested a roughly 1.4-fold improvement in step-by-step advancement, statistically uncertain but directionally consistent. Notably, participants with LLMs succeeded at cell culture six days faster than controls. Less encouragingly, both groups rated YouTube more helpful than any single LLM, suggesting that tacit embodied skills like aseptic technique are still better transmitted through demonstration than text.

LLM Novice Uplift on Dual-Use, In Silico Biology Tasks

The second paper:

Zhang CBC et al. LLM Novice Uplift on Dual-Use, In Silico Biology Tasks, arXiv:2602.23329. Preprint, arXiv, 26 Feb. 2026. https://doi.org/10.48550/arXiv.2602.23329.

Zhang and colleagues at Scale AI and SecureBio asked a related but distinct question: can LLMs uplift novices on in silico biology tasks with dual-use relevance? Unlike the study above, this one wasn’t a wet lab experiment but a human benchmark study, pitting novices with or without LLM access against biosecurity-relevant question sets including the Virology Capabilities Test, Human Pathogen Capabilities Test, and several others. Here the results were considerably more striking. LLM-assisted novices were roughly 4x more accurate than controls and outperformed human domain experts on three of four benchmarks where expert baselines were available. Nearly 90% of participants reported no difficulty obtaining dual-use relevant information despite model safety filters.

Considering both studies

These two studies are complementary, measuring different things. Hong et al. probe physical execution, where tacit knowledge, manual dexterity, and equipment fluency remain significant barriers. Zhang et al. probe knowledge retrieval, reasoning, and protocol design, domains where LLMs already operate near or above expert level. Together they paint a picture in which the cognitive bottlenecks to dangerous biology are breaking down faster than the physical ones.

What still remains relatively unexplored is the hybrid scenario that’s relevant to near-term risk: a novice using AI to design an experimental protocol, which is then executed by trained personnel or automated liquid-handling systems in a cloud laboratory.2 In that threat model, the wet lab barriers documented by Hong et al. largely disappear, and the in silico capabilities documented by Zhang et al. become directly actionable. Evaluating that design-to-execution pipeline, with rigorous controls and physical outcome measures, is the next important experiment.

Subscribe now

Phil arrived at UVA in 2017 as Director of the Data Science Institute, succeeding Don Brown, who had founded the institute. In 2019, backed by the largest private donation in UVA history, Phil shepherded the institute into UVA’s newest school, becoming its first dean. He called it “a school without walls,” and he meant it. The School of Data Science under Phil’s leadership was meant to work alongside other schools, bridging disciplines and elevating the university as a whole.

Case in point: About a month after Phil arrived at UVA in 2017, I met with him while I was still a faculty member in the School of Medicine. His office was in a trailer because the Data Science Institute had no home yet. He had been on Grounds for only weeks, and already he was reaching out across schools. When I met with him he pressed on me the case that the Data Science Institute and the School of Medicine needed stronger collaborative ties. That kind of bridge-building was instinctive for Phil. It’s a mission that, in my current role at SDS, I’m still charged with carrying out.

Before UVA, Phil had already built a career that spanned continents and disciplines. Phil spent two decades at UCSD where he co-directed the Protein Data Bank, a resource that has underpinned essentially every drug discovery effort of the past 30 years. He then served as the first Associate Director for Data Science at the National Institutes of Health, leading the Big Data to Knowledge initiative. Phil has co-authored hundreds of papers and and accumulated over 100,000 citations, placing him among the most-cited scholars in UVA history and in science generally.

Phil was also the founding Editor-in-Chief of PLOS Computational Biology, where he created the beloved Ten Simple Rules series. He was so prolific in the series that the meta-article “Ten Simple Rules for Writing a PLOS Ten Simple Rules Article” literally includes, as Rule 4: “Be Philip E. Bourne.”

I left UVA in 2019 to spend some time in industry. In 2023, I found myself in a difficult spot at a biotech startup, having just watched my boss and more than half my team get fired after joining only a couple months prior. I sent Phil an email, not quite sure what I was looking for. He invited me to his office over Christmas break in 2023 and spent the better part of an afternoon catching me up on the present and future of the School of Data Science. Phil, along with Don Brown, was the reason I came back to UVA. Phil and Don both reminded me why academic data science was worth dedicating a career to.

Just last month Phil had enthusiastically agreed to join me as a co-author a Ten Simple Rules article I’ve been working on about moving between academia and biotech. I will miss his contributions to that paper. But more than that, I’ll miss the kind of person who, after decades of building institutions and extremely busy running a new school like a scrappy startup, still answered a cold email from a former colleague going through a rough patch and said, “Come to my office. Let me tell you what we’re building.”

Phil’s vision and leadership were legendary. He leaves enormous shoes to fill, and I don’t envy the next SDS Dean who has to try. But Phil also leaves behind something durable here at the University of Virginia: a unique school with an inspiring origin story, a clear identity, an extremely talented faculty and staff who motivate me every single day, extremely bright students who keep me on my toes, and a mission that resonates more now than ever. For myself and my stellar colleagues here in the School of Data Science: the best tribute we can leave to Phil is to keep building the School Without Walls he envisioned.

Onwards, Phil.

Related, late last year I wrote about how unregulated individual adoption of AI in review can introduce its own biases when reviewers selectively prompt and cherry-pick AI outputs.

A new preprint I co-authored with Agnieszka Swiatecka-Urban (UVA School of Medicine) and Arjun Krishnan (CU Anschutz), now available on SSRN, develops these arguments more fully, with a particular focus on AI as a quality control and consistency tool for peer review.

The paper surveys decades of research documenting low inter-rater reliability in both manuscript and grant review, all of which supports the idea that reviewer assignment rather than scientific merit was the primary driver of outcomes.

Read the Paper

At the same time, the peer review system is under enormous structural pressure. Submission volumes continue to grow, reviewers are overburdened, and editors increasingly struggle to find willing reviewers. I see us entering a self-reinforcing cycle in which declining review quality encourages more speculative submissions, further taxing an already strained system.

Against this backdrop, recent large-scale studies show that AI systems achieve very high concordance with human reviewers when evaluation criteria are explicit, and that structured AI feedback improved review quality in the vast majority of assessed cases in a randomized trial.

The paper proposes a framework distinguishing where AI can help (systematic rubric-based evaluation, quality control of the reviews themselves) from where human judgment remains essential (novelty assessment, feasibility, recognizing creative leaps). Surveys of researchers themselves support this kind of complementary model over full automation.

The paper includes two supplementary tables cataloging common failure modes in grant and manuscript review that are amenable to AI-assisted quality control, from score-comment misalignment and implicit demographic bias to citation coercion and unprofessional tone, along with specific AI approaches for flagging each one.

We also genuinely engage with the risks, including confidentiality concerns, automation bias, gaming, and the worry that AI tools give overcommitted reviewers an excuse to skip the hard cognitive work of actually reading the manuscript. AI is a useful tool, but we try to be honest about the limitations.

You can read the full paper here. Citation details below. The paper is in peer review right now. I’d welcome any thoughts or comments you have (after you read the paper, not just the title/abstract). You can leave a comment here or find me on Bluesky.

Read the Paper

Turner, S., Swiatecka-Urban, A., & Krishnan, A. (2026). Rubrics, Not Vibes: Structured AI Integration as Quality Control for Peer Review. Social Science Research Network 6314421. https://doi.org/10.2139/ssrn.6314421

Subscribe now

Conference registration for the 2026 Applied Machine Learning Conference is now open! The conference will take place on April 17–18 here in Charlottesville, Virginia — now just 6 weeks away. We’re offering Full Conference, Friday Only, and Saturday Only tickets at employer-paid, personal, and student price tiers. Early Bird pricing is available until March 20.

: 30 great essays about biology. I’ve admired Niko’s writing for years, as you can probably tell from all the links to Asimov Press and Niko’s own blog I’ve linked to here over the years.

Niko is also giving out a $10,000 fast biology bounty. TL;DR: $10,000 in prizes for ideas on how to speed up wet-lab experiments. Prizes will be given for ideas that are highly original and technically tractable. A few paragraphs will suffice.

Isabella Velásquez: Claude Skills for R Users.

PDF Accessibility with LaTeX and Typst with Quarto 1.9.

What Your DNA Reveals About the Sex Life of Neanderthals (gift link). Another great story by Carl Zimmer at NYT covering the Science paper showing that interbreeding between Neanderthals and modern humans was sex-biased, with Neanderthal men mating more often with modern human women than the reverse. The word “preference” was used several times to refer to modern human womens’ “preference” for Neanderthal men.

: Agentic Engineering Patterns.

Andrew Gelman: The 80% power lie.

and : AI Won’t Automatically Accelerate Clinical Trials. A response to Dario Amodei’s recent interview with Dwarkesh Patel.

WTF Happened in 2025? A collection of datapoints in or around 2025 which we may look back as a historical inflection point. Open source & curated in realtime by Latent Space.

Ginkgo Bioworks launches Cloud Lab: https://cloud.ginkgo.bio/. Protocols that run in a fully autonomous lab. Currently 3 protocols are available, more on the way.

Science Policy Forum: The science and practice of proportionality in AI risk evaluations.

Private money cannot replace public funding of science.

Sara Altman & Simon Couch: 2026-02-27 AI Newsletter: LLM code review in a Google Docs-style interface, local models vs cloud models.

: There’s a realistic possibility you’ll read this.

Science News: How will we know if AI is smart enough to do science?

A Decade of Docker Containers. For the past decade, Docker has provided a robust solution for building, shipping, and sharing applications. But behind its simple “build and run” workflow lie years of complex technical challenges.

: Academics Need to Wake Up on AI. There are lots of nits I can pick with this one, but overall some good points throughout that are difficult to disagree with.

R Weekly 2025-W10: tidymodels, tidyplots, ggvariant.

rOpenSci: Software Review in the Era of AI: What We Are Testing at rOpenSci.

Subscribe now

New papers & preprints:

• Genome modelling and design across all domains of life with Evo 2

• DECODE: deep learning-based common deconvolution framework for various omics data

• LLM-assisted systematic review of LLMs in clinical medicine

• LRSomatic: a highly scalable and robust [Nextflow] pipeline for somatic variant calling in long-read sequencing data

• Empowering biological knowledgebases: advances in human-in-the-loop AI-driven literature curation

• How do authors want to use AI for review?

• Pangenome-guided sequence assembly via binary optimization

• Artificial Intelligence agents for biological research: a survey

• Charting single-cell lineages with synthetic and natural barcodes

• bioRxiv: the preprint server for biology

• SummArIzeR: Simplifying cross-database enrichment result clustering and annotation via large language models

• Using semantic search to find publicly available gene-expression datasets

• Interbreeding between Neanderthals and modern humans was strongly sex biased

• Torsional flexibility of the thoracic spine is superior to that of the lumbar spine in cats: Implications for the falling cat problem

Subscribe now