AI is no longer something that only researchers and data scientists work with. It is showing up in everyday web development: coding assistants, chatbots, semantic search, support tools, content generation, and product features that users are starting to expect.

But while AI tools are becoming easier to use, they are not always easy to understand. Why does a chatbot sometimes make things up? What is actually inside a prompt? Why does context length matter? What is retrieval-augmented generation, and why does prompt injection suddenly matter to web developers?

This talk gives developers a practical mental model for modern AI systems. We’ll explore how machine learning learns from examples, why large language models generate likely answers rather than guaranteed ones, and how that shapes their strengths and weaknesses. We’ll cover tokens, context windows, hallucinations, prompts, RAG, and prompt injection in plain engineering terms — no machine learning background required.

You won’t leave as an AI researcher. That’s not the point. You’ll leave better equipped to use AI as a developer: knowing where it can help, where it can break, and how to build safer, more reliable systems around it.

Understand the AI tools showing up in modern development - without the hype, magic, or machine learning degree.