# Eski Kafayla Yeni Teknoloji | Ömer Çolakoğlu | TEDxAselsan MTAL Youth Source: https://www.youtube.com/watch?v=t5AjqUb2piw Recap page: https://rapidrecap.app/video/t5AjqUb2piw Generated: 2026-02-26T17:32:15.249+00:00 --- ## Quick Overview Ömer Çolakoğlu demonstrates how modern AI technologies, particularly Large Language Models (LLMs) orchestrated via protocols like MCP, enable non-smart systems to become intelligent by allowing them to interact with data sources and perform complex tasks like controlling a drone or simulating warehouse logistics, thereby bridging the gap between the 'old mind' and new technology. **Key Points:** - The presentation focuses on merging 'old minds' (traditional systems/logic) with 'new technology' (AI/LLMs) to create intelligent systems. - LLM orchestration via protocols like MCP allows AI agents to interact with databases and perform actions like controlling a drone (demonstrated by the speaker holding a drone). - A simulation environment (Unity Engine) was used to showcase energy savings (254.4601 KWH saved, 87.52% saved percentage) by dynamically controlling street lighting based on traffic conditions. - The speaker highlights that many tasks previously considered difficult or requiring human intuition (like complex warehouse logistics or drone control) are now achievable through AI. - The concept of building 'smart' systems from 'non-smart' components is central, exemplified by teaching a non-smart robot (like the one shown) basic conversational skills and actions. - The speaker thanks the audience, including the students and organizers from TEDxYouthAselsanMTAL. ![Screenshot at 03:31: The demonstration screen showing code execution and the resulting output where the AI agent successfully processes a query about the robot's capabilities, indicating the interaction between the voice agent and the underlying logic.](https://ss.rapidrecap.app/screens/t5AjqUb2piw/00-03-31.jpg) **Context:** The presentation by Ömer Çolakoğlu at TEDxYouthAselsan MTAL Youth explores the integration of artificial intelligence, specifically Large Language Models (LLMs) and orchestration protocols (MCP), into existing or 'old' technological systems. The speaker uses demonstrations involving a drone and a simulated smart city environment to illustrate how these new AI capabilities can make traditionally 'dumb' systems intelligent, capable of complex reasoning, interaction, and task execution. ## Detailed Analysis Ömer Çolakoğlu argues that the current era, marked by the advent of generative AI like ChatGPT, presents massive opportunities for integrating intelligence into existing infrastructure, contrasting this with the past where such capabilities were science fiction. He emphasizes that the key is not just possessing the technology but orchestrating it using protocols like Model Context Protocol (MCP). He demonstrated this by commanding a drone (held in his hand) using natural language via an agent built on this architecture. A key demonstration involved a simulation environment (likely Unity, visible on screen later) where an AI agent controlled street lights based on traffic, resulting in significant energy savings (254.4601 KWH, 87.52% saved percentage). He also showed how the agent could query a local Walmart database and interact with a physical robot (shown on screen) to perform actions like learning Kung Fu or suggesting gifts. The core message is that LLMs act as the 'new language' for AI systems, enabling even non-smart components to become context-aware and collaborative, thus making complex automation simple to implement without deep technical coding knowledge for every specific task. ### Introduction and Theme - The presentation centers on merging 'Old Minds with New Technology' (Eski Kafa ile Yeni Teknoloji) - The speaker addresses school administrators, teachers, and students from the '80s generation who remember a time before widespread digital integration. ### AI Capabilities Demonstrated - The presentation covers speaking with databases, making non-smart systems smart, LLM and Agent concepts, and making science fiction real - The speaker uses a physical drone and a simulated environment to show control and analysis. ### Smart City Simulation Demo - The simulation showcased energy savings (254.4601 KWH, 87.52% saved) by having the AI control street lights based on vehicle presence and traffic patterns - The system successfully translated natural language queries (like looking for sunglasses) into database queries. ### LLM Orchestration (MCP) - The speaker explains that LLMs like GPT are the new language for AI systems, allowing agents to interact with complex systems and perform tasks like controlling the drone by reading QR codes on a simulated warehouse wall. ### Robot Interaction Demo - The speaker demonstrated giving the small physical robot instructions, showing it could understand and respond to commands, even performing a Kung Fu move, contrasting its capabilities with its non-smart origins. ![Screenshot at 00:03: The opening slide displaying the talk title: "Eski Kafa ile Yeni Teknoloji" and speaker Ömer Çolakoğlu's credentials as a Microsoft Data Platform MVP.](https://ss.rapidrecap.app/screens/t5AjqUb2piw/00-00-03.jpg) ![Screenshot at 01:36: Slide summarizing the talk's themes: 'Az Laf Çok İş' \(Little Talk, Much Work\) covering topics like 'Talking to Databases,' 'Making Non-Smart Systems Smart,' and 'LLM and Agent Concepts'.](https://ss.rapidrecap.app/screens/t5AjqUb2piw/00-01-36.jpg) ![Screenshot at 07:28: A screenshot of a web interface titled 'Chat with SQL Server' showing a natural language query: 'I am looking for a sunglasses for my 10 years old daughter' being processed.](https://ss.rapidrecap.app/screens/t5AjqUb2piw/00-07-28.jpg) ![Screenshot at 17:24: A screen capture of the Unity game engine environment showing a simulated city road network with streetlights dynamically turning on and off, displaying 'Saved Energy: 254.4601 KWH' and 'Saved Percent: 87.52%'.](https://ss.rapidrecap.app/screens/t5AjqUb2piw/00-17-24.jpg) ![Screenshot at 21:36: A live video feed from the drone's camera showing object detection bounding boxes around people and chairs in the audience, indicating the drone's visual perception capabilities.](https://ss.rapidrecap.app/screens/t5AjqUb2piw/00-21-36.jpg)