AI and LLMs in the lab and research
A practitioner's tour of how large language models and agentic AI tools fit into real biomedical research. It covers where LLMs genuinely help day-to-day (ideation, literature review, lab notes, analysis) and where they waste your time; AI co-scientists and the levels of agent autonomy; agentic coding with Claude Code (plan mode, skills, MCP, worktrees) and my actual setup; and — importantly — the data-privacy, reproducibility, and ethical precautions researchers need to take. The throughline: stay productive without producing AI slop, and keep your critical thinking alive.
- LLMs in everyday research — and their real limits
- AI co-scientists, agents, and levels of autonomy
- Agentic coding with Claude Code
- Personal precautions: data privacy, reproducibility, disclosure