OpenAI recently released its search product, ChatGPT Search, to compete against Google, but seems impractical for short queries.
Queries shorter than four words, which represent the bulk of Google searches, do not work reliably with ChatGPT Search.
It is unable to provide correct information on such queries and even hallucinates and makes up information at times.
ChatGPT Search seems successful with answering long written-out research questions.
Microsoft Bing is the search engine behind ChatGPT and is not comparable in quality to Google.
ChatGPT depends on large language models (LLMs) which are not suitable for short prompts, as it needs fully written-out questions to provide effective answers.
Perplexity's AI search tool is also facing similar issues with short queries.
Until these products improve capability for short queries, they cannot replace Google Search.
To replace the front door to the internet, OpenAI has to create a better one that can support all types of queries.
ChatGPT Search and Perplexity are covering a different part of search and are valuable in their own way.