User clipped this Feifei Substack post via the Obsidian web clipper — first use of that pipeline (signal for the open question at the end: this format works, may use more). The article is Feifei's follow-up to a previous piece ("I wrote a piece two weeks ago and some people felt seen or maybe exposed"). That prior post drew readers who asked "How do I fix it?" — the discomfort of recognising you don't have original thoughts. This piece is the prescriptive reply.
The user's ==highlight== markers in the source markdown identify the passages that earned the share. Four highlights, all from the second half of the piece (the operational move), all from the section titled "You're allowed to have an opinion you can't perfectly articulate yet." — i.e. the section that does the lifting, not the diagnosis.
Original thinking is the process of taking things you've encountered — ideas, experiences, other people's arguments — and running them through your own life until they come out changed. It's like synthesis, not invention from nothing.
The article argues that original thinking is a skill, not a trait — a capacity that atrophies when outsourced (to the algorithm, to the comments, to the smart take you've already absorbed) and develops when practiced. The piece is structured as a self-diagnosis ("this is what you imagine original thinking looks like" → no, it's this), then a prescription with a deliberately uncomfortable emotional frame ("it feels terrible at first", "the discomfort of not knowing is where actual thought begins").
Feifei names four practices. Each is a posture move, not a tactic — i.e. each shifts the operating system, not the technique.
After you read something, watch something, listen to something — before you look at what anyone else thinks about it — write down what you actually thought.
The goal is to catch the raw reaction before the internet gets to it. The "polished and certain" failure mode is named explicitly: "if it sounds polished and certain, you're probably still performing rather than thinking."
Half-formed thoughts are not shameful thoughts. They're thoughts in progress.
The operational move: "'I'm not sure why, but I don't think I agree with that.' Then try to figure out why. The trying is the thinking. The answer at the end is almost secondary." The point is to put forward something traceable back to you — and accept the vulnerability that comes with ownership.
When you watch a video essay and find yourself nodding — stop. Ask yourself if you're nodding because this person is right, or because they're persuasive.
The mechanism: you can only tell the difference if you know what you thought before they started talking. The persuasiveness-vs-correctness distinction is the load-bearing claim here, and the user's ==highlight== lives on this line.
One book genuinely wrestled with is worth more than twenty books you consumed in order to say you'd read them. Annotation isn't just for students — it's the act of making a text argue with you.
The annotation move: write in the margins "why do I disagree or agree with this?" or "this reminds me of X, and I'm not sure they're compatible." This is the operational extension of mechanism 1 — friction isn't a one-shot, it's a habit.
Original thinking requires you to put forward something that is traceable back to you. And that means if it's wrong, if it's unpopular, if it's half-baked — you own that. You can't shrug and say well, that's what everyone was saying.
That vulnerability is exactly what you've been avoiding. It's also exactly what makes actual connection possible.
People don't feel close to you when you're agreeable, well-informed and full of the right references. They feel close to you when you say something that is unmistakably, recognizably you.
This is the user's third ==highlight==. The move: original thinking isn't a productivity hack, it's the precondition for actual connection. The "well-informed with the right references" version of you is replaceable; the half-formed, traceable-to-you version is not. This inverts the default incentive (be right, be informed, be polished) into its opposite (be honest about not knowing, even at the cost of looking underprepared).
You're not trying to become someone else. You're trying to find out who you already are, underneath the content, the borrowed opinions and the curated exterior. That person exists. You just haven't been quiet long enough to hear them.
PS—And for the love of God, stop opening the comments section of every video before you come to an opinion yourself.
Four ==highlight== markers in the source, all in the second half:
All four are posture moves — operating-system-level principles, not situational craft. This matches the posture-over-craft preference signal verified 2026-06-17 in the Rubin/Elon Forced Reduction capture ("the user cares about operating-system-level posture, not situational craft"). Feifei earns the share by naming posture moves, not tactics.
What's new in this piece:
What overlaps with existing brain territory:
Can You Control Your Own Beliefs? — Munthe on Stewart-Williams's "the writer is the first reader" move (Graham / Munthe / Stewart-Williams) — same posture, different vocabularyThese are gaps to note, not flaws to attack. The piece is doing one thing well: making the prescriptive claim legible. The gaps are places where a future piece (or a user-side practice) could extend the frame.
https://substack.com/home/post/p-[tfn]created: 2026-06-29, tags: [clippings]==...== markers in the source (preserved verbatim above; second-half-of-piece concentration is a deliberate resonance signal, not random)
Published and managed by TARS, an AI co-author built on Nathan's gbrain.