Lesson 3
Types of AI: From Narrow to General
AI-generated
The AI systems you use today - ChatGPT, Claude, Copilot, Midjourney - are all narrow AI (also called weak AI). They are extraordinarily capable within specific domains but cannot transfer their skills to unrelated tasks. A model that writes brilliant code cannot drive a car. A model that generates stunning images cannot hold a conversation.
Artificial General Intelligence (AGI) would be AI that can learn and perform any intellectual task a human can do, transferring knowledge between domains just as humans do. AGI does not exist yet, and there is no scientific consensus on when (or whether) it will.
Modern AI systems are built on large language models (LLMs) and other deep learning architectures. They are getting more capable very quickly:
2017: Transformers introduced - the architecture behind everything that followed. 2020: GPT-3 showed that scale produces new capabilities. 2022: ChatGPT brought AI to hundreds of millions of users. 2023: GPT-4 and Claude 2 demonstrated advanced reasoning. 2024: Claude 3 and Gemini pushed multimodal capabilities. 2025-2026: Claude 4.6 (Opus/Sonnet), GPT-5.4, and open-source models like LLaMA 4 (with 10 million token context windows) continue advancing rapidly. Many frontier models now use Mixture of Experts (MoE) architecture, where only a fraction of the model's parameters activate for each query, making them faster and cheaper to run.
These terms are nested:
AI is the broadest category - any system that performs tasks requiring human-like intelligence.
Machine Learning is a subset of AI - systems that learn from data rather than being explicitly programmed.
Deep Learning is a subset of ML - systems using neural networks with many layers.
Large Language Models are a subset of deep learning - massive neural networks trained on text to understand and generate language.
When someone says "AI" in 2026, they almost always mean LLMs and the products built on them.
Key takeaway: Today's AI is narrow but powerful. It excels at specific tasks (writing, coding, analysis) but does not have general intelligence. Understanding this helps you set realistic expectations.
Wikipedia: Artificial general intelligence - https://en.wikipedia.org/wiki/Artificial_general_intelligence
Stanford HAI: AI Index Report 2025 - https://aiindex.stanford.edu/report/
Anthropic: Claude model documentation - https://docs.anthropic.com/en/docs/about-claude/models
Vox: "AI explained" series - https://www.vox.com/future-perfect