The Presence Vector
A companion read on the one cluster the breadcrumbs post only gestured at: roughly forty items in thirty days about how I land on other people — as large as the AI cluster, and actually three different projects sharing a shelf.
A companion piece to The Algorithm Noticed First — a closer read on the one cluster that post only gestured at.
1. It's bigger than it looks
Pulled apart by hand rather than by regex, the cluster is ~33 items (26 long-form, ~7 Shorts) out of 373 — about 9% of the month.
But that undercounts it. Sitting immediately adjacent is a set the categorizer filed under "productivity" that is really the same project: Andrew Bustamante's "CIA Framework You Can Use To Read Anyone in 4 Minutes," Jessica Chen's "The Real Secret to Answering Questions on the Spot," Codie Sanchez's "Stop Rambling: The 3-2-1 Speaking Trick That Makes You Sound Like A CEO," William Fitzpatrick on writing clearly, Sandeep Swadia on first-principles thinking. Add those seven and you get ~40 items — roughly the same size as your entire AI/technical cluster (46).
That's the headline finding. You spent about as much of the last thirty days studying how you land on other people as you spent studying AI. Nothing in the 2025 post's three-pillar taxonomy has a slot for it.
2. It's not one vector. It's three, and they behave completely differently.
A. Performative presence — ~12 items
Charisma on Command (8), Jessica Chen, Codie Sanchez, Bustamante's read-anyone framework, Fitzpatrick on clear writing.
The titles here are instruction-shaped: "Give Me 7 Minutes and I'll Make You Dangerously Good at Small Talk." "How To Be Confident Under Pressure (Even if You're Shy)." "Stop Rambling: The 3-2-1 Speaking Trick." "This Is Not Confidence." "Identity Protection Is Killing Your Confidence, Do This Instead." Each one contains a drill you could run tomorrow and know within a week whether it worked.
One detail worth flagging: "If You Only Watch One Charisma Video, Make It This" is the only repeat in the entire self-development half of your thirty days. You watched it Aug 19 and again Sep 2. Everything else in this space — Nate B Jones' ten videos, Gentle Path Forward's nine, Tina Huang's three — is all distinct titles. Music has plenty of repeats (the Deep Persian track four times, Bernstein/Mahler twice). Self-development has exactly one, and it's this.
That's your subconscious bookmarking something.
B. Attraction and self-concept — ~18 items
Gentle Path Forward (9), Marni Your Personal Wing Girl, Caitlin V, Francesca Psychology, QOVES facial analysis, plus the appearance Shorts ("Beautiful or Hot?", "How The Eyes Change Your Face", "Why Do Asians Generally Look So Much Younger").
This is the largest sub-vector, and it's structurally different from A. Look at the titles as a set:
- "You're Oblivious To How Attractive You Really Are"
- "Attractive Men are Intimidating"
- "5 Signs You Are Intimidating, Not Unattractive"
- "Signs You Are A Polymath (Rare Intelligence)"
- "The Male Body Part Women Always Notice"
- "How Women Act Around Attractive Men (Based on Behavioral Science)"
These are reassurance-shaped, not instruction-shaped. The proposition isn't "here is a thing to practice" — it's "here is evidence you are already the thing, and the reason it hasn't registered is other people's perception, not your execution." Four of the nine Gentle Path Forward videos run that exact play.
"Signs You Are A Polymath (Rare Intelligence)" is the cleanest tell in the dataset. It's a flattery-shaped title sitting inside a channel whose other eight videos in your history are about attraction. The algorithm knows what that channel is for, in your account, and it isn't information.
The problem with B isn't that it's embarrassing. It's that it has no feedback loop. There is no rep, no output, no falsification. You can watch this category indefinitely and never find out whether anything changed — which is precisely what makes it comfortable, and precisely why it's the biggest of the three.
C. Sexual health, clinician-led — ~5 items
Rena Malik, M.D. (4 videos, including two on Aug 31), DrRachael Ross.
I'd argue this doesn't belong in the charisma cluster at all. It's credentialed, clinical, and behaves exactly like your Huberman / FoundMyFitness viewing — medication interactions, physiology, evidence-based. Same information-seeking posture as the mitochondria and sleep videos, different subject.
File it under health and the remaining picture gets much clearer: you have a presence-and-communication project, and you have a self-concept project, and they've been sharing a shelf.
3. Why this matters for the creator loop specifically
You've described the goal as a compounding loop: build → write → speak → reputation.
Sub-vector A is the "speak" leg. Not analogous to it — literally it. Twelve videos in thirty days on presence, delivery, small talk, answering under pressure, and thinking clearly out loud is a person preparing for a thing they intend to do.
Which makes the parallel to the main post's finding almost too neat:
47 gear videos, no practice sessions.
12 presence videos, no recorded reps.
Same structure, different domain. The on-ramp genre is being consumed correctly. The conversion step is missing in both.
The difference is that A has a cheap conversion step. A piano requires the piano. Presence requires a phone camera and five minutes: record a 90-second explanation of what you're building with the knowledge graph, watch it back, note one thing. That's a rep. Do eight of those and you'll have learned more than the next twelve videos will teach you — and you'll have generated raw material for the "write" leg at the same time.
4. The honest read on B
I'm not going to tell you to stop watching it. Two observations instead, and you can do what you want with them:
One. B and A are competing for the same slot, and B is winning 18 to 12. They feel like the same project — "become someone who lands well" — but only one of them can be practiced. Every hour in B is an hour that felt like progress on A and wasn't.
Two. B tends to shrink on its own once A starts producing reps, because the reassurance need is downstream of not having evidence. Once you have footage of yourself explaining something well, "You're Oblivious To How Attractive You Really Are" stops being interesting — you have data, and the video doesn't. This is worth trying before willpower, because willpower against a reassurance loop generally loses.
5. What this data can't tell you
Worth stating plainly so you don't over-read your own log:
- Watch ≠ watched. No completion data. A 14-minute video in the history might be 40 seconds. This is especially true of the ~7 Shorts here, which are substantially autoplay.
- Push vs. pull is invisible. Once you watch two Gentle Path Forward videos, YouTube will show you seven more whether you wanted them or not. Some of that nine is your interest; some is the recommendation engine finding a groove. The log can't separate them.
- Absence isn't evidence. Nothing here says anything about what you're doing offline, only about what you're watching. The gap between the two is the entire subject of the main post, and it cuts both ways.
6. If you want one number to track
Ratio of A to B — instruction-shaped to reassurance-shaped — in the next thirty days. Right now it's roughly 12:18, i.e. you're consuming 1.5x more content about already being enough than about getting better at a specific thing.
Flip it to 2:1 the other way and add any recorded reps at all, and this stops being a consumption vector and starts being the "speak" leg you've been building toward.
Sources: YouTube watch history, Aug 9 – Sep 8 2026, 373 items. Hand-classified; sub-vector counts are my judgment calls and the boundaries between A and B are genuinely fuzzy in a few places (Bustamante's "Why You Feel Lost" could go either way).