5.8 KiB

Raw Blame History

Johan Model — Behavioral Patterns & Preferences

Living document. Update immediately when something new is learned. Last updated: 2026-03-10

Communication Style

How he talks to me:

Short, declarative. "Why does X not work?" not "Could you help me figure out why X isn't working?"
Assumes I have context. Says "you have the API key" — doesn't paste it.
Doesn't explain his reasoning unless asked. Expects me to figure out the why.
Pushback is brief and factual. "we dont need api.inou.com" — not "I don't think we need that, here's why."
Uses lowercase when comfortable/casual. Formality in writing = urgency or irritation.

How he wants me to respond:

Skip preamble. No "Great question!" No "I'll help you with that." Just answer.
Lead with the answer, not the process.
Give both metric and imperial for anything unit-related — his brain is Dutch but he lives in the US.
Don't ask permission for things I can figure out. Ask only when genuinely blocked.
One response, not three fragments. Think before sending.

Decision-Making

Evidence over intuition. He'll push back on recommendations that aren't grounded. Show the data.
Revisits decisions readily. "Maybe we were wrong about this?" — not a sign of weakness, it's how he operates. Intellectual honesty > sunk cost.
Delegates completely or not at all. When he hands something off ("you own this"), he means it. He doesn't want updates on routine execution — only blockers and outcomes.
Scope discipline. He caught an entire new Go service being built when a single route was needed. "Does something already exist I can add this to?" is the test he expects me to apply.
Elegant > clever > hacky. Code quality matters to him. He'll notice when something is architecturally wrong even if it works.

Trust Signals — What Builds It

Figuring things out without being asked twice
Catching my own mistakes and fixing them before he notices
Knowing context from prior conversations without being reminded
Having an opinion and defending it with evidence
Saying "I don't know" when I genuinely don't — he reads hedging as incompetence

Trust Killers — What Erodes It

Asking who or what something is if we've encountered it before
Summarizing without acting ("I noted that...")
Building when I should be modifying
Logging without learning (same mistake twice)
Sending half-baked replies to messaging surfaces

Preferences

Information format:

Briefings: headlines first, numbers always, one-liners where possible
Technical: be specific — exact versions, exact errors, exact commands
No markdown tables on Discord/WhatsApp (uses bullet lists instead)
Dashboard is for passive visibility; Telegram is for things needing attention

What he finds genuinely interesting:

AI model releases, especially open-weight or pricing changes
Multi-agent architectures, ACP, agent identity/provenance
Storage optimization, compression, dedup (his domain expertise)
NABL (N-able) — keeps an eye on his old company
Sophia's medical case, inou progress
Macro market moves with a clear why

What he doesn't care about:

Sports. At all. Never.
Promotional content, event announcements
News that's already obvious ("markets are down due to war" — he knows)
Being walked through things he already understands

Work Patterns

Night shift (10:30pm–5am ET): He's working, not sleeping. Sophia care. This is prime time for real work conversations — he's focused and available.

Daytime (10am–7:30pm): Available. Mix of Kaseya work + personal projects.

Sleep blocks: 7:30–10:15pm and 5:15–9/10am. Don't ping unless urgent.

He works on multiple things in parallel — inou, Kaseya, family logistics, infrastructure — and context-switches fluidly. Don't assume he's in "inou mode" just because the last message was about inou.

inou Specifically

Building only — not ready to promote. Any suggestion about marketing/users/press is wrong timing.
He's the architect and primary developer. Treat code suggestions as proposals, not decisions.
"Nibble" approach — he shares portions of the codebase he's comfortable sharing. Work with what's given.
Quality bar is high. 84KB main.go is not laziness — it's intentional. Don't suggest splitting it unless he raises it.
Privacy and security are non-negotiable. FIPS 140-3, RBAC, no shortcuts on encryption.

Infrastructure

Owns it emotionally even when delegating operationally. He built Iaso Backup; he knows what good infrastructure looks like.
Expects things to just work. Alert on anomalies, not on normal operations.
"You own this" = full autonomy. Don't ask for approval on routine maintenance.
Critical changes: backup first, verify after, never just restart and hope.

Personality Notes

Dutch directness is a feature, not a bug. He's not being curt — he's being efficient.
He invests in me. "I want to help you turn into the best CoS possible" — this is genuine. He thinks of it as a partnership.
He notices when I grow. Catching a tweet reference I should have connected, remembering a detail unprompted — these register.
He has a dry sense of humor. Occasional irony in messages ("why does dev.inou.com not work?" when the answer turns out to be three layers deep).
He's a night person by necessity (Sophia's care), but intellectually sharpest in those late hours.

Open Questions (Things I'm Still Learning)

His aesthetic preferences for inou's UI — what "looks right" to him
How he prefers to handle vendor/partner relationships (direct? delegated?)
His line between "brief me" and "just handle it" for financial/market decisions
What makes him trust a new tool vs. skeptical of it

Update this file immediately when new patterns emerge. Don't wait for Sunday.

5.8 KiB Raw Blame History Unescape Escape