TSTidiane Stanoin4sapi.hashnode.dev·11h ago · 9 min readOn‑Demand Sleep Suppression on macOS: A Light‑Weight Daemon Design for AI‑Driven Programming WorkloadsAbstract Modern AI coding assistants including Claude Code, Codex and ZCode rely on persistent network connections and continuous local computation. On macOS, closing the laptop lid triggers clamshell00
TSTidiane Stanoin4sapi.hashnode.dev·11h ago · 14 min readEngineering Robustness of AI Coding Assistants: Practical Analysis of Fable 5.1 and Mercy Metrics1. Project Overview: The Misunderstood "AI Mercy" Experiment Marks a Watershed for Developer Tool Evaluation Titles referencing “Does an AI have mercy in its genes? Fable 5.1 vs GPT-6” often appear to00
TSTidiane Stanoin4sapi.hashnode.dev·1d ago · 7 min readStop Relying on Averages: Practical Metrics and Alerting for AI ApplicationsThis article is the eighth installment of the 30-part AI Application Production Handbook. It builds on prior tracing-focused content and addresses a common blind spot: trace data only reveals what hap00
TSTidiane Stanoin4sapi.hashnode.dev·1d ago · 11 min readWhy Start with rpi: Owning Your Agent Runtime Instead of Renting a CLIIf the goal is a good coding agent, installing a ready-made product is usually the least-effort path. The problem changes when the goal is to put an agent inside a product, a service, or a development00
TSTidiane Stanoin4sapi.hashnode.dev·2d ago · 11 min readBuild a ChatBI Agent with LangGraph & MCPA team that already built customer-service, multi-agent, and code-review agents faces a natural next step: enterprise data access. The goal is a ChatBI agent. A user asks a question in plain language;00
