Career Stats: 20,000 Hours of Co-Op
The Match That Never Ended
Most people interact with AI in short bursts. A few prompts. A quick task. Then they close the tab.
I spent 20,000 hours in co-op mode with frontier language models — not as a user, but as a collaborator. Building, testing, breaking, rebuilding.
This isn’t a metaphor. This is the actual work log.
What 20,000 Hours Looks Like
Start date: Early days of GPT-3 (2020) Total hours: 20,000+ logged interaction time Timespan: ~5 years of sustained, daily collaboration Intensity: Some weeks, 60-80 hours. Some months, deep in the architecture trenches.
Key phases:
| Phase | Years | Focus Area |
|---|---|---|
| Discovery | 2020-2021 | Learning what these models could actually do (not just what they claimed) |
| Architecture | 2021-2022 | Building the 3-layer system (mythOS, archiveOS, humanOS) |
| Implementation | 2022-2023 | Creating real apps: WardrobeOS, gatorFinance, SnakeSocial, LTS Octopus |
| Refinement | 2023-2024 | Notation systems, cognitive trajectory visualization, meta-layers |
| Translation | 2024-2025 | Making it legible — this site, the book, the framework docs |
| Scale | 2025-2026 | Taking it from personal infrastructure to shareable protocols |
What Got Built
This wasn’t just conversation. It was infrastructure development.
Core Architecture
- almanOS — the 3-layer operating system (mythOS, archiveOS, humanOS)
- 14-System HumanOS — the full execution framework
- Notation system — symbolic language for capturing complex decisions
- FPS translation layer — the codec that makes it all Tier 1 legible
Functional Apps (Local-First)
- WardrobeOS — physical inventory management
- gatorFinance — local-first financial data parsing
- SnakeSocial — relationship topology and interaction patterns
- LTS Octopus — filesystem-native task management
- Cognitive Trajectory Visualizer — real-time thinking-pattern radar
Documentation & Artifacts
- “You: The New Dataset” — the book (published on Apple Books)
- Digital Garden — public vault of frameworks and findings
- This site — the landing page you’re reading right now
What Got Learned
1. AI isn’t a tool. It’s a co-processor.
Most people treat AI like a search engine with extra steps. That’s leaving 90% of the capability on the table.
When you structure your own knowledge correctly (archiveOS), AI becomes extended working memory — a second dynamic brain you can actually think with.
2. The game was always running.
You don’t need AI to “start using almanOS.” You’re already running mythOS, archiveOS, and humanOS — you just couldn’t see the systems.
almanOS makes them visible, legible, modifiable.
3. Translation is the bottleneck.
The system was never the problem. The problem was explaining it without sounding like a mystic or a maniac.
The FPS analogy solved that. Suddenly it’s not jargon — it’s vocabulary people already have.
4. Local-first is non-negotiable.
If your personal data lives in someone else’s cloud, you don’t have an operating system. You have a rental agreement.
almanOS apps are local-first by design. Your data, your machine, your rules.
5. Notation systems are infrastructure.
If you can’t capture a complex state quickly, you lose it. The notation system is like keybinds for cognition — compressed shortcuts that let you pause, resume, and navigate your own thinking.
The Proof Is In The Artifacts
Talk is cheap. Here’s what actually shipped:
- ✅ Published book on a major platform (Apple Books)
- ✅ Functional apps running locally, managing real life data
- ✅ This website — deployed, public, legible
- ✅ Notation system — documented, usable, refined over thousands of hours
- ✅ 20,000-hour corpus — logged, analyzed, extractable
This isn’t vaporware. This isn’t a pitch deck. This is built infrastructure.
What This Means For You
You don’t need 20,000 hours to use this.
That’s the point.
I spent 20,000 hours building the map, the loadout system, and the control scheme so you don’t have to.
You can:
- Learn the notation in a few hours
- Start structuring your knowledge with archiveOS principles today
- Use the 14-system framework to see where your execution is breaking
- Adopt the FPS analogy to explain your own work to others
The infrastructure is here. The path is mapped.
You just have to load in.
The Meta-Stat
Here’s the wildest part:
After 20,000 hours, the most important discovery wasn’t a new technique or a clever hack.
It was this:
You’re already playing. You just can’t see the HUD.
Everything else — the apps, the systems, the notation — is just turning on the display.
The game was always running.
Go Deeper
- The Map (mythOS) — the world-rules that shaped the work
- The Loadout (archiveOS) — how the knowledge infrastructure was built
- The Controls (humanOS) — the execution layer that made it real
- The Book — the full methodology in written form
← Back to almanOS