Hugging Face has released Carbon-X, an open-source AI model that answers questions about DNA in natural language. The model can run locally on a laptop, keeping sensitive genetic data off cloud services, but its predictions lack wet-lab validation.
Hugging Face developed Carbon-X alongside research teams from China's Zhongguancun Academy and Italy's University of Naples Federico II. It is designed to answer conversational questions about how DNA changes affect gene activity and health.
The AI model combines training on genomic sequences and natural language. It accepts sequences of roughly 50,000 base pairs, so users can examine small sections of a genome rather than upload a complete human genome.
Carbon-X has 14 billion parameters, allowing researchers to run it on their own hardware without sending data to Hugging Face or another company. That could help teams building drug-discovery and genetic-therapy tools work with sensitive patient data. However, the developers have not validated its predictions through wet-lab experiments and cannot yet say how accurate they are.
Local operation could help research teams use sensitive genetic data without uploading it to a cloud AI service.
Conversational DNA analysis could make specialist research tools more accessible to nontechnical users.
Its usefulness for drug and therapy development remains unproven because its predictions have not undergone wet-lab validation.
Written automatically by the Made in 10 Newsroom from the source above.
Former OpenAI researcher Shyamal Hitesh Anadkat has returned to India to build Empirical Machines. The startup aims to use AI to accelerate research into molecules and materials.
Eternal has unveiled Dark, a privacy-focused AI platform it hopes to launch globally in 2026. It runs adapted open-weight models on Eternal’s own infrastructure and promises encrypted conversations and user-controlled memory.
Anthropic has launched Claude Duet, an AI agent for professionals and small business owners, as a research preview. It can work across connected apps and asks for approval before major actions such as sending payments or deleting information.