By James Aspinwall — February 2026
Obie Fernandez — six-time founder, author of The Rails Way, and current CTO of ZAR — posted on LinkedIn that he genuinely enjoys his working relationship with his AI assistant. It sometimes makes him laugh.
That’s a throwaway line on social media. What’s behind it is more interesting.
What He Actually Built
Fernandez created what he calls a “Personal CTO Operating System” using Claude Code. In three weeks, he captured 11,579 lines of structured organizational knowledge — meeting memos, team member profiles, institutional procedures, decision history. The system processes meeting notes into actionable intelligence, maintains a knowledge base across 23 team members, and handles the administrative overhead that normally buries a CTO.
He’s had human executive assistants before. Good ones. He says this is better. The AI never needs to be brought up to speed, never forgets context, operates at conversational speed, and is always available. Everyone at ZAR is now setting up similar systems.
The Thread That Matters
The LinkedIn comments raised a sharp question. One commenter pointed out that an AI has an inherent incentive to recommend replacing humans with compute — like asking an engineer whether to fire colleagues to fund the remaining engineer’s bonus. Another replied that the recommendation might actually be sound strategy: reallocating underperforming staff resources toward computational infrastructure could be economically rational regardless of who suggests it.
This is the tension every executive adopting AI tools will face. The AI’s advice is useful precisely because it isn’t sentimental about headcount. That same quality makes people uneasy.
Why This Matters
Fernandez isn’t a hobbyist experimenting with chatbots. He founded Olympia, an AI-powered virtual staffing platform, before winding it down and taking the CTO role at ZAR. He wrote the book on Rails development. When someone with that track record says AI has replaced a key human role in his workflow — and that he prefers it — that’s a signal, not an anecdote.
The pattern is consistent: executives who invest time in structured AI integration don’t go back. The ones who dismiss it as a toy haven’t tried building a system around it.