"The Java performance industry has moved in two ways to support AI: firstly to provide contextual information to the AI from the outcome of a tool's investigation, and secondly to provide information to the AI on how to use a tool to analyse a performance issue"
"Stream JFR events to a database, then process them into statistical aggregations (like counts, averages, and maximums) and send to an AI for reasoning and decision making"
