Blog

Project leaders share hard-won lessons on embracing uncertainty, owning delivery, rescuing failing projects, and why the technology is the easy part.

How a self-assessed, evidence-backed quality program across dozens of teams became the foundation for safe AI-assisted development.

Heavy review packets, phase-gate approvals, and exception theater look like control—but they are waterfall habits wearing enterprise branding.

What research shows about AI-assisted development: the METR slowdown, code churn, security vulnerabilities, and mounting maintainability concerns.

How five AI coding agents coordinated 40 merged pull requests in one afternoon using nothing but GitHub Issues, labels, and the gh CLI.

Many teams invest heavily in AI coding tools expecting faster delivery but without the right processes, these tools can fail to deliver. Here's how to fix it.

How a local Qwen3.8 27b agent named Echo reverse-engineered two abandoned 2013 Android games — Delver and Flappy Bird — into signed, modern APKs.

AI-generated JIRA tickets and PRs have become word walls. Here's why your automation needs both a human summary and a machine-readable payload.

Practical techniques for giving AI agents the context they need: AGENTS.md files, documentation strategies, and plan mode workflows.

Gitflow is now a legacy workflow that bakes unplanned merge work into your sprints. Here's why to drop it for trunk-based development.