Generative AI for Developers: How To Use AI In Your Workday

You open your inbox and surprise! There’s a chaotic email from your boss trying to squeeze five vague requests into one sentence. This lesson shows how to use AI as your inbox assistant. You’ll watch as ChatGPT helps break down a stream-of-consciousness message into a task list, organizes missing information, and even drafts a response email. Instead of wasting brainpower early in the day, you’ll learn how to delegate the grunt work to AI and start your dev day focused and organized.

Developers write code, not polished emails. Someone still has to explain what just shipped. In this session, you’ll see how to take shorthand Jira notes or dev comments and turn them into clean updates that make sense to a product manager. You'll compare AI output from multiple tools and learn how to adjust tone, audience, and clarity without leaving your coding mindset.

You’ve inherited code older than your laptop, written in a language you barely remember. In this module, you’ll learn how to use AI tools to analyze legacy code, identify bugs, suggest security improvements, and even rewrite sections for clarity or best practices. You'll compare how ChatGPT, Claude, Gemini, and Copilot each handle the same task and see where they shine or fall short.

Regex is powerful but painful. Especially when you're stuck Googling patterns that almost work. In this session, you’ll see how to describe what you want in plain English and let AI generate the regex. You'll test real patterns, refine results with better prompts, and use tools like ChatGPT to debug overly greedy matches or subtle syntax issues. No more trial and error in the dark.

Data migrations don’t have to ruin your lunch break. In this lesson, you’ll feed AI a source and target schema, and let it handle the busywork: mapping fields, flagging risky transformations, and generating real INSERT INTO ... SELECT SQL statements. You’ll also see how AI helps avoid common traps like unencrypted passwords or bad type casting—and why it’s still your job to double-check everything.

Bloated tables and missing foreign keys slowing you down? This session shows how to use AI as your database design assistant. You’ll drop in schema definitions and let the AI flag issues like redundant fields, vague column names, or lack of normalization. Then it suggests improvements, new table structures, smarter types, indexes, and better naming, so your database works as hard as your code does.

Blurry whiteboard photos. Slack screenshots. Jira tickets with bad formatting. You’ve seen them all. In this lesson, you’ll learn how to use AI to extract checklists, requirements, and structured summaries from messy visual inputs. Whether it’s a password policy mockup or a hand-drawn schema, you’ll turn it into dev-ready output you can act on, or hand off, with confidence.

Time for something hands-on. In this lesson, you’ll prompt AI to build a playable tic-tac-toe game using HTML5 Canvas and JavaScript. You’ll run the game live, explore how readable the output is, and then refactor it using AI tools that explain their improvements step-by-step. You’ll also compare how different models write and revise the same game, and learn how to guide the AI when it gets stuck.

Vibe coding looks slick in a demo, but does it hold up in production? In this intro to the 4 PM hour, you’ll learn what vibe coding really means, how it fits into the AI landscape, and what it can and can't do for your codebase. You'll unpack the two types of code generation, one-way vs. round-trip, and get a grounded view of the real limitations of LLMs, from hallucinations and context limits to non-deterministic behavior. This talk reframes AI as a productivity booster, not a replacement for architectural thinking.

Now we roll up our sleeves and put vibe coding to the test. In this session, you’ll walk through how to take a loosely written set of app requirements and feed them to an AI coding assistant inside a real development environment (VS Code with agent support). You’ll learn how to give context, how the different modes: Ask, Edit, and Agent and change the behavior of the tool, and how AI responds to varying levels of specificity in your inputs. You’ll also see why vague requirements like “make the UI feel clean” are where AI gets weird—and how to steer it back on track.

In this hands-on demo, you’ll watch AI build a small Vue.js commodity trading app, complete with bugs, missed components, and a few moments of triumph. You’ll see how to run agent mode inside VS Code, give instructions with varying clarity, and keep the AI on track as it scaffolds components, builds logic, and refactors based on feedback. You’ll also compare how Claude and ChatGPT handle the same task and learn when to let the AI run and when to step in. By the end, you'll better understand what “working code” really means in an AI-assisted dev cycle.