The video discusses the difficulty of obtaining truly original ideas from AI and proposes a structured, multi-step prompting process to generate more diverse and engaging concepts, emphasizing the importance of meaningful interaction and playability over mere novelty. It offers practical advice for creative AI use, encourages experimentation with the method, and highlights the user’s role in selecting and developing worthwhile ideas.
The video explores the challenge of getting truly original ideas from AI, particularly for creative projects like HTML canvas simulations. The creator initially asked an AI for five original ideas twice in separate chats and found that the suggestions were often similar or repetitive, despite different wording. To address this, they shifted from asking for original ideas directly to providing a structured process that encourages the AI to generate a wider variety of concepts through multiple steps.
The new process involves sending three separate messages in the same chat. The first message asks the AI to list 50 interesting HTML canvas simulation ideas that humans find exciting, then randomly pick five without ranking them. This approach aims to avoid the AI defaulting to safe, familiar ideas. The second message asks for 20 quick variations and combinations of those five ideas, again selecting five at random to maintain diversity. The third message requests the AI to choose the most enjoyable idea and build a simple interactive version, focusing on clear main interaction and playability.
This multi-step approach helps the AI produce more varied and engaging concepts, such as a tiny demolition game where players carefully lower a cup of tea by knocking out wooden supports. The creator emphasizes that surprising or fantastical premises alone don’t guarantee a satisfying experience; the interaction must be meaningful and worth repeating. They also note that even with this process, some repetition occurs, but the focus should be on judging the usefulness and playability of ideas rather than just their novelty.
The video offers practical advice for using AI creatively: keep messages separate to track idea development, separate exploration from selection to avoid premature narrowing, and always consider the user experience—what players or users will enjoy doing or discovering. The creator suggests building the smallest viable version of an idea to test its appeal and encourages applying this process beyond games to lessons, workshops, or product features, though those applications remain experimental.
Finally, the video invites viewers to try the experiment themselves by running the three-message process in two separate chats and comparing the results. The creator highlights that AI opens many possibilities, but it’s up to users to decide which ideas matter and are worth developing. They promote their Get Amplified program for practical AI learning and offer personal guidance for those wanting to explore creative AI workflows further. Links to prompts, games, and resources are provided for deeper engagement.