Algorithm, Search & Distribution
Semantic search
Also called: meaning-based search
Older systems matched strings and the query had to appear in the document. Semantic search represents both the query and the page as meaning, using entities and vector embeddings, so someone asking why their camera keeps hunting in low light can be shown a page about autofocus.
This changes what optimisation looks like. Repeating a phrase eight times adds nothing. Covering the whole question in the language your audience actually uses does. Exact wording still earns its place in the title and the first line, where a literal match is what a human scanning a results page recognises as the answer.