From Autocomplete to Autopilot: A Practical Guide to AI-Assisted Coding

​AI coding tools have evolved fast — from inline suggestions to full-blown agents that can scaffold modules, write tests, and refactor code across files. But what does that actually mean for your day-to-day work, and where should a seasoned developer start?

​This session led by Scott Falconer of the Drupal AI Initiative, breaks down the current landscape of AI-assisted coding into clear, practical tiers: inline completions (think autocomplete on steroids), chat-in-your-IDE copilot workflows, and the newer "agentic" coding loops where AI plans and executes multi-step tasks with your oversight. We'll look at what each style is good at, where it falls down, and — critically — how much control you keep at each level.

​You'll come away understanding what tools like GitHub Copilot, Cursor, Claude Code, and others actually do under the hood, how to evaluate which approach fits your workflow, and how to move up the ladder at your own pace without feeling like you're handing the keys to a junior dev who doesn't know what <?php print 'hello world'; ?> does.

​No hype, no "AI will replace you" nonsense — just a clear-eyed look at what's useful now and how to adopt it without abandoning the engineering discipline that got you here.