The contemporary landscape of AI development is witnessing a proliferation of programs and services dedicated to the creation of cultural media, often relying on web-based interfaces for their generative processes. Among these, conversational interfaces and visual programming have emerged as prominent interactive methods catering to diverse user preferences. While general users are enthralled by the possibilities of text-to-image generation exemplified by Dall-e and Midjourney, specialized users gravitate towards platforms like Comuify and bespoke interfaces for fine-tuning model parameters to their specific needs. This discourse aims to delve into the concept of WYSIWYP (what you see is what you prompt), mirroring a prevalent paradigm in web design, and its relevance within the context of AI-driven media production.Furthermore, the discussion will explore the role of visual diagrams in deciphering latent images, shedding light on how visual programming provides insights into the factors shaping image variations before generation. Although these paradigms may appear distinct, they are not mutually exclusive; rather, they represent a transitional phase towards more symmetrical models for engaging with generative content. Ultimately, they converge to enrich user experiences and foster deeper interaction with AI-generated media across a spectrum of applications and user types. By embracing these diverse approaches, the field is poised to unlock new realms of creativity and innovation, empowering users to navigate and harness the potential of AI in shaping cultural media in ways previously unimagined.