Latest Articles
Why Brownfield Codebases Still Need Human Review
AI tools write clean local code, but real production context remains scattered across team boundaries and legacy contracts.
Paying for Your Own Tokens: Slashing 45M to 12M in a Multimodal Critic Loop
Corporate accounts hide token waste behind generous buffers. Building an automated comic generation pipeline exposed severe image and context bloat in the art review loop: and the specific gates that dropped token burn from 45M to 12M.
A Couple of Conversations Away: Turning Pushback into Company-Wide Leverage
How an engineering review panel on a ticket campaign led to a mindset shift—moving from defending legacy limitations to aligning incentives and solving systemic platform challenges.
Don't Just Ask 'How', Ask 'What If': Letting AI Rethink Your Workflows
How brainstorming with an LLM completely transformed my comic generation workflow from a clunky CLI to a state-managed web app.
Earning the Bright Future: Why We Can't Leave Tomorrow to Chance
Reflecting on Dr. Fei-Fei Li's words on why technological progress requires active risk mitigation rather than passive optimism.
Why Pushback is a Gift (And How to Use It)
A story about how unexpected pushback on a change management project led to questioning assumptions, breaking muscle memory, and finding a better way.
