Modern Applications of Generative AI

From Control to Emergent Intelligence focuses on helping learners understand how generative AI behavior is shaped, guided, and extended, moving from surface-level interaction to a systems-level perspective. The course begins with how humans control models at inference time through prompting strategies and sampling parameters, then steps back to examine how models are shaped during training through reinforcement learning, fine-tuning, and feedback. Learners develop a clear mental distinction between intelligence that is baked into a model during training and intelligence that emerges at inference time through structure, reasoning, tools, and memory. This framing allows learners to see modern generative AI not as a static tool, but as a dynamic system whose behavior depends on both how it was trained and how it is used. As the course progresses, learners move beyond single prompts to structured reasoning, model comparison, and evaluation across different architectures and ecosystems, including open-source and mixture-of-experts models. They then explore how tools, memory, and context persistence allow AI systems to operate across time, enabling action-oriented workflows rather than isolated responses. The course concludes with real-world applications across domains such as coding, business, accessibility, and creative work, paired with individual-level ethical reflection on what it means to work alongside AI systems. By the end of Course 2, learners understand not only how to use generative AI effectively today, but how the combination of control, feedback, reasoning, evaluation, and external capabilities gives rise to more autonomous behavior, setting the foundation for agents and more advanced systems explored in Course 3.