Retro PC Games Podcast Series
A podcast experiment using AI-generated dialogue to explore classic PC games. Each episode dives into the history, mechanics, and cultural impact of retro titles, featuring engaging discussions between virtual hosts. This project blends AI-driven content creation with gaming nostalgia, pushing the boundaries of automated storytelling.
About The Project
This podcast series was a personal experiment in AI-generated media, blending a love of retro PC games with an interest in emerging generative tools. Built using Google’s NotebookML platform, the project tested how far structured prompts and curated reference material could be used to script full-length podcast episodes complete with characterful narration, structured storytelling, and factual depth.
The core idea was simple: pick a classic PC game, research its background, mechanics, and cultural context, then feed that data into the system to generate a two-host conversation. The hosts fully AI-generated would discuss the game as if recording a real podcast episode, covering its impact, quirks, and legacy. I then edited the responses, cleaned up any awkward phrasing, and layered in an intro jingle and basic audio treatment to give the episodes polish and personality.
Across the series, I produced around 15 episodes, each focusing on a different title from the DOS and early Windows era games that defined a generation of home computing. From tactical strategy to point-and-click adventures, each episode aimed to balance nostalgia with insight, offering listeners a mix of factual information and casual banter between the virtual hosts.
Ultimately, this project was a low-stakes creative sandbox an experiment in AI-assisted storytelling rather than a polished production. But it offered a unique way to explore the limitations and potential of large language models in long-form media generation, and it showed how even simple tools can be repurposed for playful, genre-specific content when given the right structure and intent.
Key Features
AI-Generated Podcast
Leveraged Google’s NotebookML to generate dialogue-driven episodes.
Retro Gaming Focus
Explored classic PC games, their legacy, and why they remain relevant today.
Data-Driven Content
Curated structured information sources to ensure informative discussions.
Technical & Creative Experimentation
Tested the potential of AI-generated media while fine-tuning output quality.
Process & Technical Insights
The project began as a way to explore AI-generated media and its capabilities for structured storytelling. Using NotebookML, I provided datasets and prompts that shaped the conversations between virtual hosts, ensuring the dialogue remained coherent and engaging. The AI-generated discussions covered game mechanics, industry trends, and historical significance, all framed in a casual, podcast-style format.
While the results were impressive, the process required substantial refinement—curating data sources, tweaking conversational flow, and managing AI quirks to maintain a natural feel. Despite these challenges, the project demonstrated the potential of AI-assisted content creation, particularly in niche topics like retro gaming.
Ultimately, the Retro PC Games Podcast served as both an exploration of new AI tools and a deep dive into gaming history, blending technical innovation with nostalgia.
Got questions?
Feel free to reach out