Tech for product managers in 2026: a four-level path from Lovable to agent systems, and what a Chief AI Officer says about PMs pushing code.
Tech for product managers in 2026: a four-level path from Lovable to agent systems, and what a Chief AI Officer says about PMs pushing code.
How /loop, /goal, scheduled tasks and a Ralph shell loop differ in Claude Code, and what starts and stops each one.
How to use OpenClaw: the current install commands from its README, what onboarding does, the security steps to take first and how a CPO runs it in production.
What Claude Cowork is, which plans include it, the risk Anthropic flags, and three workflows from my February 2026 guide.
What Claude Design is, who can use it, how it compares with Figma and AI prototyping tools, and how I run a first prototype and deck in it.
What Claude Code skills are, where they live, a real SKILL.md in full, and three lessons from my tests on descriptions, templates and skipped steps.
Claude Code MCP: how to add servers, how scopes and output limits work per Anthropic’s docs, and what running them in a real content repo taught me.
What Claude Code subagents are per Anthropic’s docs, the three I keep in my repo, and seven rules from real runs: budgets, models, verification.
How Claude Code plan mode works per Anthropic’s docs, how to turn it on, where my repo requires it for newsletters and scripts, and when to skip it.
How Claude Code memory works per Anthropic’s docs: CLAUDE.md, auto memory, the 200-line and 25KB limit, and the numbers behind the setup I run.