In this video, we dive into one of the biggest misconceptions in AI today: the belief that Large Language Models (LLMS) like ChatGPT, Claude, and Gemini can solve optimization problems. Planning problems (e.g., employee rostering, delivery route optimization, school timetabling, job-shop scheduling) live in a completely different universe from next-token prediction. These problems are NP-hard, combinatorial, and require actual search, not linguistic guesswork. You’ll learn: 1. Why LLMs fail spectacularly at real planning tasks, 2. Why even “reasoning models” collapse under combinatorial explosion 3. What kinds of tools are built for the job: mathematical optimization, constraint solvers, CPLEX, MiniZinc, meta-heuristics, Timefold, and more. Most importantly, you’ll see how to combine LLMs with solvers to build powerful hybrid AI systems where the LLM handles reasoning and structure, and the solver does the heavy lifting. If you’re building AI systems or working in software engineering, this video will save you from one of the biggest traps in the current hype cycle. 👉 Don’t forget to like, subscribe, or drop a comment with your thoughts if you liked this video. ⏰ 𝐓𝐈𝐌𝐄𝐒𝐓𝐀𝐌𝐏𝐒 00:00 Intro 01:50 What a planning problem really is 04:29 What traditional optimization does differently 06:45 Why LLMs break down on planning tasks 13:31 The Hybrid Approach (LLM + Solver) 15:25 When to use which tool 📺 𝐑𝐄𝐋𝐀𝐓𝐄𝐃 𝐕𝐈𝐃𝐄𝐎𝐒 → → →