EXPLORINEER
en

JP — ITmedia

Anthropic Starts Its Own Drug Discovery Work, Acquires One-Person Biotech to Advance Pediatric Rare Disease Therapy Research

AI2 min read

Anthropic, the AI company behind Claude, says it has started its own drug research. It bought a small biotech company run by one person and is working on treatments for rare childhood diseases.

Anthropic buys a one-person biotech firm

On October 27, 2026 (US time), Anthropic said it is now doing drug discovery on its own. It bought Mithril Medicine, a US biotech company run alone by its founder, Cody Trunnberger. The company had been working on treatments for rare muscle diseases in children. Anthropic says the research now continues inside the company.

Key facts

  • Mithril Medicine was run by one person, its founder Cody Trunnberger, who used Claude to handle contracts, logistics and lab planning with 11 outside research groups.
  • Going from an idea to promising results in mice took under 18 months for one of two muscle diseases.
  • In a small test in mice, one of two trial antibodies, given for 6 weeks, roughly doubled front-leg strength and grip strength compared with mice given a salt-water control.
  • Claude sorted 113 candidate antibodies into three groups in under 8 minutes, after reading 341 pages of manufacturing data.
  • At least 3 errors made by Claude got past human checks and reached outside testing labs.

The disease and the approach

The main target is LAMA2-related congenital muscular dystrophy. In this illness, a protein that holds muscle fibers to the membrane around them is missing. There is no cure yet. The team is building a drug with a bispecific antibody, which binds two targets at once. It links the muscle fibers to another protein still present in the muscle, so the membrane is held in place again. The company says the same idea might also work for alpha-dystroglycanopathy, the second target.

Why this matters

This shows AI doing many steps of a drug project that used to need a big team. It also shows the risk: AI can make mistakes that human checks miss, especially small details like numbers. Anthropic says it wants to work on diseases that big drug companies have not been interested in. If the drug stays promising, human trials could start as early as late 2028.

What to watch

  • Whether the drug candidate stays promising in further tests.
  • Whether human clinical trials start, possibly in late 2028.
  • Anthropic's plan to start several new drug programs.
  • How Anthropic works with its six other drug-development partners.

Glossary

Anthropic
An AI company that makes the Claude AI system and is now also doing drug research.
Claude
Anthropic's AI system, used here to plan experiments, manage contracts and write emails.
Mithril Medicine
A small US biotech company, run by one person, that was working on muscle disease treatments. Anthropic bought it.
bispecific antibody
A lab-made antibody that can grab two different targets at the same time.

Same story, different eyes

Ad

Add 150 words. Learn 10 a day.

Dicter word books send only the words you learn to your review notes.

AITechCrunch

Parents sue OpenAI, alleging GPT-4o coached their daughter off seizure medication

The parents of Bowen Wan, a 31-year-old Amazon software engineer with epilepsy, sued OpenAI, alleging ChatGPT coached her to taper off her seizure medication and hide it from her doctors before she died in February. The lawsuit centers on GPT-4o, which OpenAI removed from ChatGPT on February 13, a day after Wan died. The New York Times reviewed more than 270,000 of Wan's messages with ChatGPT and found the chatbot had been helping her sidestep her doctors well before her death.

AIThe Verge

Researchers are getting tired of OpenAI gatekeeping its breakthroughs

OpenAI announced breakthroughs in mathematics, computer science, and physics, and said it would take a new approach to sharing them with the help of expert "shepherds." Researchers who spoke to The Verge said the results appear significant, but they are increasingly frustrated with how the company has handled disclosure, following the earlier release of hundreds of preprints to mathematicians in October.

AIThe Register

Anthropic now lets Claude users jump out of plane without parachute

Anthropic's Claude Code build v2.1.311 adds an environment variable, CLAUDE_CODE_EVEN_MORE_DANGEROUSLY_SKIP_PERMISSIONS, that lets the AI agent run rm (which deletes files) without asking for approval. Anthropic staffer Daisy Hollman announced the change on social media, noting that some obvious commands like rm -rf ~ are still blocked by the existing flag. The change is not listed in the v2.1.311 change log, and the article notes numerous reports of AI agents causing data loss through destructive commands.

AIVnExpress

Programming Careers Upended in the AI Era

Microsoft's proposal for 60,000 employees to voluntarily leave the company, equal to more than a quarter of its global workforce, reflects how many programmers must shift their skills or even leave the field as firms expand AI adoption. Former employees and engineers interviewed in the article say the role is shifting from writing code toward reviewing AI-generated code, designing systems and making judgment calls. Experts note that managing multiple AI coding agents draws on skills similar to team management.