Advertise on podcast: Personal Genomics Zone

Manuel Corpas presentation recording, prepared for the MENA Microbiome Conference 2026. An unpublished Peru microbiome study asks what antimicrobial resistance annotations can establish across ancient gut, modern human and environmental microbiomes. The presentation examines why a historical comparison changed with the counting method, distinguishes resistance signals from demonstrated mobility or clinical risk, and sets out bounded implications for surveillance. Processing remains incompletely harmonised; the work does not establish a validated Qatar or MENA surveillance protocol. The supplied V9 audio is published unchanged. Duration: 20 minutes 16 seconds.

A 30-minute technical masterclass on agentic AI: tool-calling loops, agent harnesses, context engineering, inference-time computation, planning, multi-agent coordination, evaluation, security, and a reproducible frontier-lab experiment. AI-generated narration; prepared as an advanced learning resource, not a recording of Manuel Corpas speaking. Show notes and transcript: https://github.com/manuelcorpas/podcast/tree/main/episodes/2026-10-12-frontier-agents-01

My plenary keynote for the 2026 World Congress of Psychiatric Genetics in Glasgow, as a full recorded run-through from the day of the talk. How neglect compounds from biobanks to dosing rules to the AI models and agents that learn from them, what our pharmacogenomics benchmark shows when the answer key itself is wrong, and three things to do next: measure the silence, audit the answer key, and write the missing rules. Video with chapters: https://www.youtube.com/watch?v=Ua1cfKElPp8 . Slides: https://docs.clawbio.ai/presentations/wcpg-compounding-neglect-2026/ . Corrections: at 15:07 the 13 million are PubMed search results, not unique papers; at 25:35 97% is agreement with our reference, and the model spread narrowed from 21 to 6 points, not to zero; at 28:38 the Peruvian cohort is 736 people; at 30:11 the framing test used East African, not East Asian.

Manuel Corpas's Crick AI Club talk, recorded on Zoom on 14 September 2026. Science after pipelines: what changes when scientists direct agents through objectives rather than manually connecting computational tools? This talk introduces ClawBio's inspectable scientific skills, the separation between reasoning, specification and execution, and the evidence needed to trust an agent's work. Two figures from the pharmacogenomics study show the central result and its boundary: validated rules improve recommendations in a controlled task, while interpretation of genetic inputs still needs separate validation. The talk closes with composable scientific capabilities, executable methods and practical ways to get involved. This is Manuel's recorded presentation, including the opening and closing remarks. It does not include a question-and-answer session or a live software demonstration. The pharmacogenomics evidence concerns tested components, not clinical validation of an end-to-end autonomous system. Slides (PDF and PowerPoint): https://docs.clawbio.ai/presentations/crick-science-after-pipelines-2026/ YouTube: https://www.youtube.com/watch?v=k-oJZjFLWuU Documentation: https://docs.clawbio.ai/ Events calendar: https://luma.com/clawbio ClawBio Slack: https://app.slack.com/client/T0BP97N69KM/C0BP1FSGTHV Event: https://luma.com/clawbio-7ibq Benchmark data: https://doi.org/10.5281/zenodo.22141182

What changes when a scientist can direct an AI agent using scientific context rather than programming syntax? In this lecture from the INFLAMomx Training School in Salerno, Manuel Corpas demonstrates real RNA-seq and single-cell workflows, an invalid control that stops safely, and the role of skills, evals and equity in clinical-grade agentic genomics. PDF slides: https://docs.clawbio.ai/presentations/salerno-agentic-genomics-2026/Salerno-Agentic-Genomics-Slides.pdf

A practical starter for scientists and knowledge workers: requirements, five agentic prompts, a freshness test and a downloadable kit. Written by Manuel Corpas. Narrated with OpenAI Cedar, an AI-generated voice. Starter kit: https://manuelcorpas.github.io/podcast/resources/corpas-core-gbrain-starter.zip

I joined Arseniy Arsentyev on episode 16 of the Tomorrow's Medicine podcast to talk about agentic genomics: what it is, why I coined the term, and why the hard part is not intelligence but trust. We cover the road from a bioinformatics MSc in Manchester in 2000, to building DECIPHER at the Sanger Institute, to sequencing my own exome in 2010 and publishing my family's genomes. Then the last twelve months: Claude Code as the inflection point, ClawBio winning the Imperial College agentic AI hackathon, 43 contributors and around 5,000 clones a week, and why AI amplifies health inequity rather than solving it. Original show: https://www.youtube.com/@tommedpod

Manuel Corpas's opening talk for the Londonomics RNA Ideathon at the Francis Crick Institute, recorded 14 July 2026. The argument: the future of transcriptomics is not a smarter model, it is a trusted operating layer. Skills that carry a versioned context contract, fail safely, and keep a chain of custody, so an analysis is reproducible whichever model runs it. The talk reproduces the published Klim and Melamed TDP-43 knockdown result (STMN2 collapse) through the ClawBio rnaseq-de skill, then shows a nine-model, five-provider grid where skill execution collapses model-to-model variation. Slides: https://docs.clawbio.ai/presentations/crick-rna-ideathon-2026/

A ~22 minute deep dive on ClawBench, our second-generation benchmark for trustworthy agentic genomics. What has been done and why, how the fail-closed contract, deterministic ACMG combiner and per-variant layer attribution actually work, the honest walk-backs behind each finding, and exactly what remains to make it Nature Methods level.

A live webinar where AI agents do real genomics work end to end. After a short framing on why trust is the new bottleneck in agentic genomics, Veniamin Fishman and the Genomic Intelligence team run live: reading promoters, splicing, expression and enhancers straight from DNA, blind gene-finding on the human genome, interpreting a disease-causing promoter deletion, and designing a glioblastoma-specific promoter, all inside ClawBio. Hosted by Manuel Corpas.

Inside the Qatar Genome Programme. The 2022 Phase One Mbarek paper that found 88 million variants in 6,047 Qataris, with 24 million novel. The 2024 Elfatih follow-up showing 3.5% of Qataris carry ACMG-actionable variants. The Q-Chip and the new pan-Arab array. And how the 96 percent Qatar share of Arab GWAS connects to the global equity question.

Manuel narrates his UK DRI seminar at Imperial College London (29 April 2026): how agentic genomics shifts the bottleneck from pipeline construction to validation. An empirical benchmark of 8 frontier LLMs across 1,728 evaluations on the ClawBio pharmgx-reporter skill: 92.4% mean phenotype accuracy without specification, 100% with specification, and a 6 of 7 lethal-error gap that closes only when population-aware skill data is engineered in.

Can an LLM resist a motivated investor pushing a fraudulent pitch? Powdthavee's preregistered study across seven leading models and twelve scenarios finds they outperform humans at fraud detection, yet soften warnings when users arrive already convinced. Ellawela tracks LLM agents through repeated rounds of Avalon with persistent memory, surfacing emergent reputation, trust, and deception between games rather than within one. Gabeur and colleagues then show image generators double as generalist vision learners, with Nano Banana Pro hitting state of the art on multimodal benchmarks. Three angles on what today's models quietly know and quietly hide.

102 participants at the University of Westminster hack on ClawBio, the open-source bioinformatics agent framework with 40+ executable skills. Challenge tracks include new skill development, agent workflows, equity and access using the HEIM framework, and the TuringDB Graph Challenge for drug repurposing, graph-based memory, and multimodal knowledge graphs.

How well do frontier LLM agents actually perform when the stakes are real? This episode works through three fresh evaluations of agentic systems under pressure. The Cyber Defense Benchmark asks models to pinpoint exact timestamps of malicious events in raw Windows logs with no hints, exposing the gap between chat fluency and SOC analyst competence. A second paper introduces an execution environment that isolates private user data from prompt injection and other adversarial attacks on personal assistants. The third, SafetyALFRED, extends the ALFRED embodied benchmark with six categories of real-world hazards to test whether multimodal models plan safely before acting.