+a search engine that filters specifically for high-quality content — not most popular, not most recent, but actually best. the motivation is that Google search results have degraded: SEO-optimized content dominates, real expertise gets buried under content farms, and you have to know in advance which sources to trust. the experience of searching for nuanced technical or intellectual topics and getting mostly mediocre results is now common enough to be a cultural complaint.

+

+the filtering approach: a combination of source reputation signals (which domains consistently produce substantive content), quality indicators that aren't easily gamed (genuine engagement in comments, citation by other quality sources, structural signals like depth and specificity), and potentially a community layer where people with demonstrated expertise in a domain can curate what's good. the last part is interesting — a small number of trusted curators per domain would dramatically improve results for that domain, and the curation work could be distributed rather than centralized.

+

+this connects to [[universal-agentic-searcher|universal agentic searcher]] which attacks a related problem (deep research across sources) but from the automation side rather than the filtering side. [[dense-info-generator|dense info generator]] is downstream — it assumes quality sources have been found and synthesizes them. [[life-search|life search]] is the personal analog: searching your own high-quality prior knowledge rather than the web. [[idea-tester|idea tester]] could use quality-search as its underlying retrieval engine when doing market research and competitive analysis.