LLMs tend to uncover traits or disclose tendences based on ongoing searches (something traditional search was unable to do beforehand).
If you dive into LLM deep thinking, you’ll note harnesses compressing and summarizing behaviors.
When I asked GPT 5.6 to summarize my personal aptitudes, it came up with “practical luxury”, “high-agency optimizer”, “leverage-seeking operator”, “coordination-tax intolerant”, “systems maximalist”, “optionality accumulator”, “control-through-visibility operator”. This summarizes approaches to life, family, philosophy, longevity, communication.
Opus agrees with “specification re-negotiator”, “Scope frontloader”, “Serial re-diagnostician”, “artifact delegating principal”. These are often relevant to how I brief, plan, structure, communicate, and iterate on work R&D, building teams, iterating on KPIs.
Never would I ever synthesize these the same way. But LLMs can do that when they see how I travel, what I focus on in my diet, how I do R&D, what briefs I download for my teams, how often I do reconciliation. There’s critical meta data hidden in it.
If you ask LLMs to profile you and uncover what deep thinking sessions find out and are willing to share, drop them in the thread.
