Descriptions:
Hasan Aboul Hasan delivers a 2026 update to his prompt engineering course — the original 2023 version accumulated over one million YouTube views — focusing on five advanced techniques he argues are absent from most current tutorials. The video opens with a quick review of the four-pillar prompt structure (role, task, context, rules) before moving into the advanced material.
The first technique is recursive brainstorming: prompts are nested iteratively, each layer generating child ideas beneath parent concepts, enabling deep exploration of a topic tree that would be tedious to traverse manually. Hasan’s free visual playground automates this, rendering the idea tree as an interactive diagram with scored nodes and JSON export. The second is the web JSON technique — prompting models to search the web and return results in structured JSON for downstream use. The third is the LM feedback loop, where a model critiques and rewrites its own output across multiple iterations, with the playground halting automatically when quality scores stop improving. Fourth is the LM judge pattern: multiple models (Claude, ChatGPT, Gemini) each produce a response, and a designated judge model either selects the best answer, synthesizes all of them, or generates a comparative breakdown of strengths and weaknesses.
All techniques are demonstrated live using Claude, GPT-4o, and Gemini, with a free playground that provides both no-code interaction and exportable code snippets for developers integrating these patterns into applications.
📺 Source: Hasan Aboul Hasan · Published January 09, 2026
🏷️ Format: Course Lesson
