> 
TL;DR
- Wednesday duo — Yann installs voluntary constraint (English Codex) and excellence over perfection while demoing AI-as-workforce; Branka arrives on smoke-triggered conditioning and sharpens her environment coach lane.
- Doctrine: Problems at the Height of Ambitions · Excellence hides in the nuance vs perfection trap · Conditioning · Feedback as learning currency on Codex readability · Personal Growth Engineering / simplification is sophistication — simple surface, complex background.
- New shelf: AI literacy — Deterministic vs probabilistic · Harness · Apps Claude · Grok · Grok Bot · HeyGen.
- Engagement: Branka — invent category (don’t copy luxury design) · voice-note definition loop · prep screen-share Codex UX session · Martine — excellence lens on launch (carry Monday backup battle).
- Fast path: Excellence over perfection · voluntary constraint.
- Table of Contents
- [[#English Codex pivot · AI as workforce|English Codex pivot · AI as workforce]]
- [[#Excellence over perfection · voluntary constraint|Excellence over perfection · voluntary constraint]]
- [[#Smoke · conditioning · environment calling|Smoke · conditioning · environment calling]]
- [[#Invent the category · Swiss design tokens|Invent the category · Swiss design tokens]]
- [[#Codex harness · deterministic vs probabilistic AI|Codex harness · deterministic vs probabilistic AI]]
- [[#Codex UX feedback · complexity vs simplicity|Codex UX feedback · complexity vs simplicity]]
- [[#New in the Codex from this session|New in the Codex from this session]]
- [[#Next session — engagements|Next session — engagements]]
Segments
English Codex pivot · AI as workforce
Yann · Martine · 00:00–00:20
Yann opens with Martine on the decision to move the Codex to English — a voluntary constraint sized to ambition (Problems at the Height of Ambitions): coach, produce, and publish in English faster than comfort would allow. Martine names it as constraints for good — a door opening, not a loss. Yann shares the operator stack: HeyGen avatar iterations, Grok Bot for second-brain fetch, LinkedIn comment automation, Markdown as single source of truth in the second brain, GitHub sync, and the n8n webhook that collapsed thirteen post-session steps from ~two hours to minutes (Harness). The vision: One Person Business with AI as workforce — one human orchestrating volume that used to need a team (AI Orchestrator · Context window).
Excellence over perfection · voluntary constraint
Yann · Martine · 00:11–00:20
Before Branka joins, Yann names the session theme: perfection paralyzes publish and action; excellence means iterating daily/weekly systems toward higher efficiency on the right behaviours (Excellence hides in the nuance · Busy versus truly productive). Martine reframes excellence as above perfection — ship and refine. Yann ties this to Ideal day and ideal week math: after forest, fitness, English, and training, only ~half a day remains — so Efficiency vs Effectivity on chosen moves beats task-list volume. Automation (VTT → session → Publish pipeline) is the worked example: don’t wait for perfect; measure behaviours that evolve excellence week to week.
Try within 24–72h: Name one launch move you are delaying for perfect — what is the smallest excellent version you could ship this week?
Smoke · conditioning · environment calling
Yann · Branka · Martine · 00:20–00:35
Branka joins worried about distant wildfire smoke — her walk and mood are altered; she notices how quickly she could habituate to degraded air. Yann links this to Conditioning and the frog-in-water pattern: disruption reveals what you take for granted (Interoception). Branka’s pre-session voice note lands: hyper-sensitivity to environment is why interior design became her path — subtle multi-sensory harmony to create Flow and energy rather than being passively shaped by space. Yann wires the Personal Alignment System (SAP) loop: define → Morning Voice → listen → upgrade → Weekly Review — the definition is the product, not a one-time answer (Personal Definitions · The Words We Use to Define a Need).
Invent the category · Swiss design tokens
Yann · Branka · 00:35–00:48
Branka reports outward action — Belgrade, Design Week scouting, practitioner contacts — but warns she could lose herself copying existing luxury design instead of inventing her category. Yann uses windsurf as metaphor: distill what inspires you; don’t become someone else’s sailboat (Zone of Genius). His own Battle Of the Week this week: Swiss Design Style — design-as-code for the second brain (simplification is sophistication · Complexity Trap). Screen-share of the visual token dictionary: Mega-prompt + tokens vs probabilistic chat that drifts (Critical thinking on AI · Deterministic vs probabilistic · Determinism · Probabilism). Workflow: Flashcard → Gemini → Grok distill — structure before volume.
Try within 24–72h: One sentence — My category is ___ not a copy of ___.
Codex harness · deterministic vs probabilistic AI
Yann · Branka · 00:48–00:54
Branka asks whether Codex is “local AI.” Yann clarifies: interconnected Markdown in Obsidian — not a model — AI reads in when harnessed. He shows associating token files with Gemini Gems and retro-engineering Pinterest infographics for consistency. Yearly decision: 2026 = daily time, energy, and money into AI depth, not surface ChatGPT (Personal Growth Engineering). Branka names Yann’s coded AI vocabulary as a language to learn — start at AI — Quality of Questions and Active Recall apply to operator literacy, not only member doctrine. Claude and ChatGPT return different answers when context differs — that is deterministic vs probabilistic design, not model loyalty.
Codex UX feedback · complexity vs simplicity
Yann · Branka · Martine · 00:54–01:05
Branka offers structured Codex feedback: ultra-concise and consistent, but sometimes too dense — words feel out of context; automated LinkedIn posts feel disconnected; AI may be over-fitted to Yann without outside readability checks (Feedback as learning currency · Teachability index). She wants fluidity and context bridges — the same tension Yann names at close: simple on the surface, complex managed in background (Personal Growth Engineering · Swiss editorial clarity). Yann proposes a double session: each member screen-shares Codex navigation unprepared, think-aloud — he records and tunes the reading experience. Martine affirms differentiation: without a Harness, everything online looks the same (Deterministic vs probabilistic). Close warm — see you Friday.
Try within 24–72h: (Fleet) Block one 30-min slot for a screen-share Codex walk — no prep, verbal only.
New in the Codex from this session
- Deterministic vs probabilistic
- Harness
- Mega-prompt
- Gemini Gem
- Claude
- Grok
- Grok Bot
- HeyGen
Next session — engagements
Branka Farge
- [ ] One sentence: My category is ___ not a copy of ___ (environment / interior coach lane)
- [ ] Continue field research (Belgrade / Design Week thread) without conforming to luxury templates
- [ ] Morning Voice once the environment-coach definition is written
- [ ] Prep for screen-share Codex UX session (unprepared navigation + think-aloud)
- [ ] Optional: log one moment smoke/habituation taught you about conditioning
Martine Aubert
- [ ] Optional: name one launch move delayed for perfect — smallest excellent ship this week
- [ ] Optional: note where differentiation matters in your visibility stack (not generic AI sameness)
- [ ] Carry Monday battle: non-negotiable backup for the 50 — see 2026-08-17
Group (open for adoption)
- [ ] Screen-share Codex UX double session — readability / teachability / barrier-to-entry live audit
- [ ] Excellence over perfection — iterate systems weekly; don’t wait for perfect before publish
Logistics
- [ ] Confirm replay `mTobk-kpv9A` after Publish