Part 2/12:
Shapiro explains his foundational knowledge, highlighting his work on prompt engineering, instruction tuning, and fine-tuning models like ChatGPT. The latter, he notes, is essentially a fine-tuned variant designed to function as a conversational agent. He critiques it as a "wall of text generator"—a system that responds to prompts without asking questions or understanding context beyond the immediate input. While GPT models are powerful, Shapiro remarks that their capabilities are often overstated, particularly when considered as revolutionary disruptors.