A little premature to be calling such a prediction "silly". I think it's safe to assume some sort of LLM-based tech will be part of the most successful search engines within a relatively short period of time (a year tops). And if Google dallies its market share will definitely suffer.