Do you recall the first time you posed a question to ChatGPT and received an answer in clear, understandable English? For many of us it was an eye opener because this language ability was far above anything that had been possible before. This breakthrough in language understanding propelled ChatGPT to an unprecedented 100 million users in just two months, marking the fastest adoption of an application in history. Only a couple of years later, this capability — the Large Language Model (LLM) — is widely available in many forms and is powering more tools and apps every day. This is leading to a fundamental shift in human-computer interaction and is why Microsoft CEO Satya Nadella says “Human language is the new UI layer.”

But how is it this possible? This talk takes a pragmatic look at how LLMs process human language starting with how they break it down and represent it. Our focus will be developing an intuitive understanding that will include a gentle introduction to tokens and embedding models. We will look at illustrative examples in multiple languages and audience participation here will be welcome. There will be no math. 🤓

We’ll close by briefly connecting the dots to other popular AI and AI-adjacent technologies like semantic search, vector databases, retrieval augmented generation (RAG), coding assistants, and even the humble chatbot. Don’t worry if any of these terms are unfamiliar. We won’t go deep, but rather we’ll zoom in on how the language understanding piece is relevant. This can help you appreciate when these other technologies might be useful.

The target audience for this talk is business leaders, developers, product managers, and anyone curious about how AI really works behind the scenes. If you’ve used ChatGPT (or similar), you’re ready for this talk; no other technical background required.

You’ll walk away with an intuitive understanding of how LLMs process language that you can apply when evaluating AI tools, building AI features, or just demystifying the tech for your team.