How to Perform Effective Project Management with AI

Project management for software engineers has been fundamentally altered by the rise of Large Language Models (LLMs). The distribution of time has shifted from writing code to prompting AI agents, with a significant increase in time spent on testing. To adapt, engineers should map out work more thoroughly beforehand to enable agents to work autonomously for longer periods. Specific techniques like using a `/goal` command can force agents to complete all aspects of a task, while automating testing with tools like Playwright can reduce new bottlenecks. This new workflow requires abandoning old project management heuristics to fully leverage the productivity gains offered by AI.