Codex vs Claude Code: My Real Division of Labor
How I use OpenAI Codex CLI and Claude Code together — the actual task split, review gates, what each one gets wrong, and why I stopped picking one.
Insights on finance, deal structures, systems, and building. 16 articles.
How I use OpenAI Codex CLI and Claude Code together — the actual task split, review gates, what each one gets wrong, and why I stopped picking one.
What Claude Code's --dangerously-skip-permissions flag disables, what auto mode's classifier catches instead, and the gates I run unattended agents behind.
Zach Vorsteg on AI systems architecture as a solo technical founder in Palm Beach — tools, costs, and what the job market data misses.
Most founder dashboards are noise. Here are the specific numbers I check every day across five ventures—and what I stopped tracking entirely.
A March 2026 snapshot of my $257/month development stack: coding tools, infrastructure, what I dropped and how I evaluate the cost.
A practical guide to agentic engineering: break work into tasks, set review gates and decide when AI-generated code needs human intervention.
How I configure agent instructions, tool routing, memory and permission hooks for production workflows across my ventures.
How I use AI agents across real estate, trading, hardware and apps: tasks I delegate, limits I work around, and lessons for a one-person company.
From ESP32 prototype to field deployment: lessons from building an IoT thermal detection system with infrared sensors and cellular connectivity.
Where AI-assisted coding helps, where it creates risk, and how I choose between rapid prototypes and engineering for production systems.
What breaks when a solo founder automates trading, hardware and apps: cascading failures, monitoring blind spots and what still needs human judgment.
I run 54 trading bots across 5 markets. Here's what 3 years taught me about strategy decay, slippage, correlated drawdowns, and what actually breaks.
How I share progress across software, trading and hardware: what to publish, what to keep private, and a practical framework for building in public.
I run 54 automated bots on a single Mac using Python and launchctl — for $8/month. Full architecture, real failures, and cost analysis vs. cloud.
How coding helps me analyze deals and serve commercial real estate clients, and why I combine software building with work as a licensed sales associate.
DCF, LBO, and merger models explained — the three financial models that drive every M&A valuation, with synergy frameworks and sensitivity analysis.