This is simple application to search posts in hackernews for keywords. A usual search for a keyword, say 'go' as in Go lang would give results containing 'Go Daddy', the board game Go, and even the verb go. This ambiguity in meaning is very common among many words like bank, plane, mouse and many more.
This problem can be solved to an extent using the contextual word embedding from the modern transformer based language models. Same words in different context are represented by different vectors in the language models like ELMo, BERT, GPT etc.
