researcherlisted

# Researcher Most "research" fails for one reason: it searches to confirm, not to falsify. This skill exists to produce research that survives someone actively trying to break it — a patent examiner, a peer reviewer, a competitor's lawyer, a skeptical cofounder. ## Step 1 — Scope before searching Pin down, in writing, before running any search: - The exact claim or question being tested (not "is my idea good" — "does a system with properties X, Y, Z already exist, and if a close variant exists, what exactly differs"). - The domain boundaries (industry, geography, time window — a 2019 paper doesn't tell you what shipped in 2026). - What would count as a **kill** — the single strongest form of prior art or counter-evidence that would invalidate the premise. Name it before searching for it. If you can't name what would kill the claim, you don't understand the claim yet. ## Step 2 — Search across every layer that matters, not just the easy one Skipping a layer is the single most common way research gets blindsided later. Cover what's relevant to the question: - **Academic**: arXiv, IEEE Xplore, ACM DL, Google Scholar — for SOTA techniques and named prior work. - **Patents**: Google Patents, Espacenet, WIPO PatentScope, USPTO full-text search — required whenever novelty or patentability is even adjacent to the question. A literature search with no patent search is not a novelty search; patents are often the closest prior art and never show up in a plain web search. - **Comme