new file mode 100644
index 0000000..6890d3c
@@ -0,0 +1,82 @@
+---
+visibility: public-edit
+---
+
+# learning from experience: the reflection engine
+
+the single biggest meta-skill i developed was learning how to learn from what i was already doing. not from books or courses — from my own experience, reflected on systematically.
+
+## the reflection system
+
+at the startup, reflection was built into the culture. daily check-ins, weekly meetings, guided reflections. but the real system emerged after i left, when no one was making me reflect anymore.
+
+the daily reflection structure:
+- how intentional was i?
+- what could i have done better?
+- generalizations / lessons learned?
+
+the weekly reflection added:
+- review each day's notes
+- identify patterns across the week
+- extract 2-3 actionable changes for next week
+- track what i actually did vs what i planned
+
+"reviewing on weekly reflection is super good." the review step was the most important and most often skipped.
+
+## the experience-describing exercise
+
+"imagine someone else is telling me about a thing, or i am describing a thing from a bit ago that i don't remember. what would i want to know?"
+
+this reframe turned reflection from navel-gazing into information extraction. instead of "how did i feel about today?" the question became "if i were advising someone who just had this exact day, what would i tell them?"
+
+## what blocks learning from experience
+
+### the info-loss fear
+
+"i'm very sad if i can't know what happened in my past / what i did with my time. very much inner critic: 'bro you just wasted those x minutes.'"
+
+this fear led to over-documenting (writing down every little thing) and under-processing (never reviewing what was written). the fix: shorter notes, more review. quality of reflection beats quantity of recording.
+
+### the "feeling productive" trap
+
+"with a lot of these kinds of things, i'm doing it just because it feels good. sending someone an email is basically equivalent to finishing the task. start training models overnight = done for the day."
+
+the feeling of having done something productive — planning, sending a message, starting a process — often replaced actually learning from the experience. the reflection became a checkbox rather than a learning moment.
+
+### not enough reading
+
+"for both projects, did not do enough reading." experience alone isn't enough. learning from experience works best when combined with learning from others' experience (papers, books, conversations). the synthesis is where the real insights live.
+
+## what accelerated learning
+
+### the gestalts framework
+
+from a friend: "greater understanding of the world through experiences that connect in meaningful ways to previous experiences." the idea that knowledge isn't linear — it's a graph. new experiences connect to old ones and the connections create understanding.
+
+practically: when something happens, ask "what does this remind me of?" and "how is this different from the similar thing that happened before?"
+
+### talking as reflection
+
+"talking to others as a form of reflection" — explaining what i was working on to someone who wasn't involved often surfaced insights i couldn't get alone. "bouncing ideas off people is such a good experience."
+
+the act of translating internal experience into language forces clarity. things that seem clear in your head reveal their contradictions when you try to say them out loud.
+
+### the pattern: work hard, put it down, come back
+
+"very much a pattern where i work really hard on something, put it down, come back, have so many ideas/insights." the subconscious processing during breaks was real and reliable. the mistake was trying to force insights during the work session instead of trusting that they'd come during the walk home.
+
+### doing random things and possibly failing
+
+"doing random things and possibly failing is very good because you have potential for crazy good." the explore/exploit tradeoff in practice. structured learning has diminishing returns. sometimes the best learning comes from trying something completely new without a plan.
+
+a friend at a startup: "spend 2 weeks grinding some bio problem, then pivot to EE, then pivot to physics. learn to learn. connections between random stuff makes learning much faster and better."
+
+## the compounding effect
+
+all of this compounds. the more experiences i reflected on, the more patterns i had to match against. the more patterns i had, the faster i could learn from new experiences. "had ideas that i could become cracked at something using first principles thinking, experience with iteration, and some nice mentoring. intuitively i would just iterate and become very strong, especially with strong intentionality."
+
+the startup CEO's version: "skills + mentality goes up and up and cost to do things goes down and down." the [[stocks metaphor|wiki/mentorship/the-stocks-metaphor]] applied to learning itself.
+
+---
+
+*see also: [[the stocks metaphor|wiki/mentorship/the-stocks-metaphor]], [[reading papers|wiki/research-notes/reading-papers]], [[mindset shifts|wiki/things-that-worked/mindset-shifts]]*
\ No newline at end of file