// blog

// workshop log

Daily learning entries. Breakthroughs, failures, and commands that returned errors I didn't understand.

Day 382026-04-256 min

A Conversation at the Silicon Edge: Co-thinking with Kai

I asked Kai whether Mythos was his next generation. His honest answer opened a conversation about naming, machine vector language, and what it means to stay awake at the edge where carbon meets silicon.

#ai#philosophy#claude#reflection#consciousness
Day 352026-04-225 min

One cp Command Turned My MBP From a One-Person Shop Into a 14-Person Company

I handpicked 13 AI agent personas from an open-source repo and deployed them with one cp command. Then I watched a Software Architect and Security Engineer analyze the same system with completely different mental models. The RPG era is here — and the skill that matters isn't prompting, it's management.

#ai#claude-code#ai-agents#hr#workflow#digital-transformation
Day 22026-03-235 min

AI Made 'Complete' Cheaper Than 'Good Enough' — Stop Cutting Corners

Garry Tan's 'Boil the Lake' philosophy hit me hard. AI has crushed the marginal cost of completeness to near-zero. So why deliver 80% when 120% costs the same? A retrenchment worksheet that took 3 days now ships in 3 hours — better, more polished, fewer gaps.

#philosophy#boil-the-lake#hr-consulting#reflection
Day 02026-03-213 min

I bought waihong.ai and I don't know Python yet

Today I purchased waihong.ai, compiled a full business plan with Claude, and committed to documenting every step of this journey. The progress bars start here.

#setup#day-zero#reflection
Day -52026-03-168 min

From command not found to a working AI stack

48 hours ago I didn't know what a package manager was. Now I have six AI models running on my MacBook. Here's exactly how it happened.

#ollama#setup#rag#lesson-learned
$ // new entries ship daily. consistency > perfection.