Scientific Brainstorming

Problem

Research ideation often stalls in two places: a familiar framework hides challengeable assumptions, and too few cross-domain analogies let ideas only follow old paths. scientific-brainstorming treats the model as an equal research brainstorming partner rather than a lecturer that simply provides answers.

How It Works

It follows five stages. First, it clarifies the current project, interests, constraints, and unexplored angles. Then it drives divergent exploration with cross-domain analogy, assumption reversal, scale shifting, constraint removal or addition, interdisciplinary fusion, and technology speculation to generate many ideas. Next, it looks for shared threads and surprising connections. It then evaluates the most promising directions by asking about testable conditions, minimal experiments, and available resources. Finally, it synthesizes next steps such as literature search, pilot experiments, or collaboration.

Boundaries

This is a dialogue, not report generation. The researcher should carry at least 50% of the conversation while the model asks questions, expands ideas, and structures them. If the work shifts to multi-step reasoning, long-running workflows, large-document analysis, or multi-tool coordination, treat it as lightweight ideation support rather than a complete research execution plan.