Demystify artificial intelligence — from neural networks to ChatGPT
10 Episodes
Audio Lessons
237 Minutes
Total Learning
Beginner
Friendly
Artificial intelligence has moved from science fiction to daily life. ChatGPT passed the bar exam. AI generates art, writes code, and threatens to transform every industry.
But what actually is AI? How does GPT work? What can it do — and what can't it? This collection cuts through the hype to explain what's really happening.
Most modern AI is machine learning: systems that improve from experience. Instead of programming rules, we train models on data.
Inspired by brains, neural networks are layers of connected nodes. Deep learning uses many layers. This enables breakthrough capabilities in vision, language, and more.
GPT (Generative Pre-trained Transformer) predicts the next word in a sequence. Trained on internet-scale data, it develops remarkable capabilities — and limitations.
Where is AI heading? Multimodal models. AI agents. Safety concerns. Job displacement. The debate over consciousness.
10 audio lessons • 237 minutes total
Defining AI. History from Turing to today. Narrow vs general AI. The AI winter and revival. Why AI is everywhere now.
~25 min
How ML differs from traditional programming. Supervised, unsupervised, reinforcement learning. Training data and models. Real-world applications.
~25 min
What neural networks are. Neurons and weights. Forward propagation. Backpropagation. Why depth matters. Universal approximation.
~30 min
Generative Pre-trained Transformer explained. How GPT predicts next words. Pre-training on massive data. Fine-tuning. The evolution from GPT-1 to GPT-4.
~25 min

Under the hood of ChatGPT. RLHF: reinforcement learning from human feedback. Why it seems smart. What it's really doing. Limitations and capabilities.
Emergent abilities. Passing exams. Code generation. Reasoning tasks. Chain of thought. What's impressive and what's overhyped.
~25 min
Hallucinations. Reasoning limits. Knowledge cutoffs. Common sense gaps. Why AI isn't conscious. The alignment problem.
~25 min

Bias in AI systems. Misinformation concerns. Job displacement. AI safety research. Existential risk debates. Regulation approaches.
OpenAI, Anthropic, Google, Meta. Open source models. AI startups. Enterprise adoption. Investment landscape. Competition dynamics.
~25 min

Multimodal AI. AI agents. AGI predictions. Jobs of the future. Coexisting with AI. Preparing for an AI-transformed world.
Master the building blocks of programming — variables, loops, functions, and computational thinking
Learn the principles of user-centered design — from research to prototyping
Master the art of clear thinking — spot fallacies, evaluate evidence, and make better decisions
AI is impressive but not magical. Here's what current AI systems actually can't do — and why it matters.
ChatGPT took the world by storm. Here's how large language models actually work and what they can (and can't) do.
Transform your commute, workout, or downtime into learning time. Our AI-generated audio makes complex topics accessible and engaging.
Related topics: