Researcher on doing science in the agent era: ask the hard questions, verify every line
CatAstro_Piyush · x · 2026-10-13
ML researcher Aman Sharma published a new essay, Research Taste in the Age of AI Agents, addressing juniors asking how to do research now that agents can read papers, run experiments and write them up.
Core claim: agents speed up the research loop, but two steps must never be handed over —
- Asking a hard question you truly care about, which sets the direction;
- Verify and debug: read every line the agent returns.
Everything in between can be delegated: reading and discussing, forming hypotheses, letting agents build and run code/training/inference, plus sharing publicly. The essay draws on a year of working with agents rather than against them, and on building research taste.
More from coding & agent
- Running tens of millions of experiments on useless programs with LLMs — bushibuilds · 2026-10-13
- Xiryn: Open-Source Personal Agent with Docker Sandboxes and Telegram Control — Scobleizer · 2026-10-13
- Witness Tracks Everything You Do and See So Your Agents Always Have Context — Scobleizer · 2026-10-13
- MecAgent 2.0 launches as its maker claims 'CAD is about to get solved' — Scobleizer · 2026-10-13
- Dev finds ~25% of AI-generated code introduces regressions in .NET 8+ — lucasmeijer · 2026-10-13
- Investor says coding agents left most software firms badly overstaffed, cut now or be disrupted — doodlestein · 2026-10-13