What happened
For the past 15 months, one creator has been quietly building something unusual: a television network that runs itself, 24 hours a day, using AI to write, generate, and schedule almost everything you see on screen. The project is called Botflix, and it currently broadcasts four separate channels at once — a general entertainment channel, a radio channel of original AI-composed music, a music-video channel, and a relaxation/ambient channel for background viewing.
Until recently, most of the work was manual: the creator picked what to make, wrote the prompts, and pieced episodes together by hand. About 11 days ago, that changed. The creator connected ChatGPT, paired with OpenAI's coding assistant Codex, directly into the production pipeline. Now, on a set schedule, the AI system can look at what's currently airing, check performance data, decide what new content is needed, generate it, slot it into the live schedule, and retire older content — without a person clicking through every step.
A typical automated run
A typical "run" looks like this: the AI reviews the existing content library and what's currently playing, decides what kind of piece to make next, writes the creative prompts, generates the raw video and audio using tools like Google Flow (Google's AI video generator) and Suno (an AI music generator), edits and renders the finished piece, adds it to the library, and rebuilds the live playlist so the new content goes on air with a clean handoff — no jarring cut, no dead air.
None of this would work on a rigid, pre-built schedule. Back in March, the creator switched Botflix from one long pre-rendered video loop to a "modular" system, where every song, clip, and segment exists as its own separate file. That's what lets the automation swap pieces in and out of rotation on the fly, every time something new gets finished, instead of waiting for a whole new master file to be rendered.
The other piece that made this possible: 15 months of already-made content sitting in the library. Handing a fresh AI system an empty folder and asking it to build a TV network from zero is a much harder problem than handing it a mature library and asking it to keep the machine running. The creator still sets the creative direction and the rules the system has to follow — this isn't an AI running a media company on its own, it's an AI handling the repetitive production and scheduling work under human-set guardrails.
What it means for you
You don't need to run a TV network to feel the effects of this shift. The same idea — AI that plans, creates, and manages content on a loop with light human oversight — is showing up in ordinary work and life already.
At home: picture a family photo slideshow or a kids' bedtime playlist that updates itself weekly, pulling in new AI-generated stories or music without you manually curating it each time.
At work: a marketing team could let an AI system monitor which social posts perform best, then automatically brief itself to make more of what's working — the same "check analytics, decide, generate, publish" loop Botflix uses.
In business: a small e-commerce brand could run an AI system that watches which product videos get the most views and keeps generating new variations of the winners, instead of a human re-editing old footage every week.
In study: a student could set up a rotating AI-generated study playlist — flashcard audio, summary videos — that refreshes itself based on which topics they're struggling with.
In creativity: independent musicians and video artists can use the same building blocks (AI video generation, AI music generation) to keep a "channel" of their own work growing, without a full production team.
In income: ad-supported or subscription "always-on" content channels — the exact model Botflix uses — are a real, if early, way solo creators are experimenting with recurring revenue, since the content keeps refreshing without new manual labor each day.
If you want to experiment with the free AI tools behind this kind of workflow — writing, image, and video generation — without committing to paid subscriptions first, a starting point like mykreatool.com collects several free AI tools in one place, which is useful for testing ideas before you invest real money in a paid stack.
How to try it right now
You don't need Codex-level automation to test the core idea. Here's a simple, mostly-free way to try it yourself.
1. Start free. Open ChatGPT (the free tier works fine for this) and ask it to act as a "content programmer": give it a short description of your channel or feed's theme, and ask it to plan a week's worth of content ideas plus the exact prompts you'd use to generate each one.
2. Generate the pieces. Take those prompts into free or low-cost generators — Suno for music, or Google's video generator (currently branded Google Flow) for video clips — to produce the actual assets.
3. Assemble a simple rotation. Even a free playlist tool (a YouTube playlist, a Spotify playlist, or a folder set to auto-play) can act as your "channel." Add new pieces as they're finished, and remove the oldest ones on a schedule.
4. Set a recurring check-in. Once a week, ask ChatGPT to review what you've published and suggest what to make next based on what performed best — this is the "scan analytics, decide" step Botflix automates.
5. Automate later. Once the manual loop starts feeling repetitive, that's the point where connecting a coding assistant like Codex to handle the scheduling step — the way Botflix did — starts to make sense.
Upsides and what changes
The biggest shift here isn't that AI can generate a video or a song — that's old news by now. It's that AI can run the scheduling and decision layer on top of that content: deciding what to make next based on real performance data, and slotting it into a live rotation without someone manually rebuilding a playlist every day. For any creator or small business maintaining an "always-on" content channel — a YouTube channel, a background-music brand, a social feed — this removes a lot of the daily grind of babysitting the calendar. The human's role shifts from "make everything by hand" to "set the rules and direction," which is a genuinely different job, and one that scales a lot further.
Limitations
This isn't a hands-off money machine, and it's worth saying plainly: the system still runs on rules and creative direction a person spent 15 months setting up, it required custom engineering to connect the AI to a real production pipeline that most people don't already have built, the tools involved (Google Flow, Suno, ChatGPT/Codex) mostly need paid tiers once you move past light, casual use, and "the AI decided what to make" still means decisions made inside boundaries a human drew. This is automation of production and scheduling — not an autonomous media company making its own business decisions.
Conclusion
The takeaway isn't "AI can make a TV channel" — it's that AI can now handle the repetitive loop of deciding, creating, and scheduling content once a human has built the pipeline and set the rules. If you run any kind of ongoing content — a channel, a playlist, a social feed — today's action is small: open ChatGPT and ask it to plan next week's content and the exact prompts to generate it, so you can see the "decide, then generate" loop for yourself before automating any of it.
Comments 0