Agenda

• 17:30 Registration

• 18:00 Introduction

• 18:30 Main session - What's inside the AI?

• 20:00 Networking

What's inside the AI?

Today, AI is usually associated with Claude, ChatGPT, Gemini, Grok, or some other sleek chat interface answering our questions. As tech professionals, we know that underneath it all lies an LLM wrapped in a nice UI. But is that really all there is to Artificial Intelligence? Are LLMs truly the revolution, or just a piece of it? Is an LLM still as impressive once stripped of its exoskeleton? What is the role of
harness?

Welcome to a short journey into what lies under the hood of AI - what it is, what it isn’t, what it excels at, and where it struggles. At a sprinting pace, we will trace the foundations of AI from classical approaches (sometimes called GOFAI), through analytical methods, the earliest neural networks, and AI winters, all the way to deep learning and transformers. We will take a close look at what actually makes an LLM useful, discuss scientific research on when using them genuinely makes sense, and explore the landscape of various tools for developers looking to integrate LLMs into their applications.

We will have a closer look at open LLMs that are available for developers, especially focusing at possibilities offered by Gemma 4. We will try to understand how to fit it into the app seamlessly.

If time permits, I will also take you into the darker corners of computer science to examine whether LLMs are truly secure, look at potential attack vectors, and perhaps even try our hand at a live demo.

We will conclude our AI adventure with a quick glance at what is currently happening on the scientific frontier and what researchers are focused on next.

Prepare for a dynamic walkthrough of the most fascinating AI concepts that led to the creation of Large Language Models. By the end, you will be able to fully answer the question: what does AI really have inside?

Detailed Plan

• Introduction: Stripping the Exoskeleton

• The Foundations: Classical AI (GOFAI) & Early Approaches

• The Shift to Learning: From Analytical to Early Neural Networks

• Deep Learning & Architectural Evolution

• The Transformer Revolution (Core Section)

• LLMs on the Operating Table: Open Models & Gemma 4 (Core Section)

• LLMs as the Heart of Applications: Building the Harness (Core Section)

• The Darker Corners: Security & Vulnerabilities

• Conclusion: The Scientific Frontier