A developer describes automating a 24/7 television network called Botflix using ChatGPT/Codex for content generation and playlist management.
For the past 15 months, I’ve been building a 24/7 AI-native television network called Botflix. What started as a mostly manual production process has now turned into a system where ChatGPT/Codex handles a large part of the ongoing production and programming. The network currently runs four continuous channels: a general TV channel, an original-music radio channel, a music-video channel, and a relaxation/ambient channel. About 11 days ago, I connected ChatGPT/Codex directly to the production pipeline. On a scheduled run, it can now inspect the current programming, scan analytics, decide what to make, generate finished pieces, add them to the live playlist with a clean switchover, and cycle older material out. A typical run looks something like this: -Review the existing library and current playlists -Decide what kind of material to make -Write the prompts. -Generate assets in Google Flow, Suno, etc. -Assemble and render the finished piece -Add it to the programming library -Rebuild the relevant playlist and update the live rotation One of the most important evolutions that made this possible was switching to modular programming back in March. The network isn’t one giant pre-rendered loop, although it did start that way. Every piece exists as individual assets. The automation changes the playlist every time something new is added. Getting a system to run the network more like a real programming/production operation was only made possible by adding GPT-6 Astra to the workflow. And by having 15 months of accumulated content already in the library, handing it off to the agent was likely much easier than trying to build something like this from scratch. I still manage overall creative direction and the rules the system operates under. So I wouldn’t describe this as an AI independently creating and running a TV network entirely. It’s more like a human-designed production environment where more and more of the creative and operational work is being done by an agent. I’d be especially interested in hearing from people working on similar automated media pipelines. Botflix TV: https://www.youtube.com/live/6xuQnM810Zw?is=jxJq68QcICSJZMa3
OpenAI launches Codex, a cloud-based AI software engineering agent powered by codex-1 (optimized o3), capable of writing features, fixing bugs, and proposing pull requests with parallel task execution. Available to ChatGPT Pro, Business, Enterprise users today, with Plus and Edu support coming soon.
OpenAI engineers showcase Codex as an agent harness for software engineering, capable of reviewing code, splitting work across sub-agents, and running workflows autonomously, effectively turning one person into a full engineering team.
An individual open-sourced a system of 10 AI agents that automate YouTube Shorts creation from podcast clips, including transcription, editing, captioning, and scheduling, with a quality-checking agent.
OpenAI introduces Codex as a distinct AI agent designed to execute tasks across files and tools, complementing ChatGPT's conversational capabilities by handling workflows like file updates and automation.