Decentralised storage answers who holds the bytes. A creator's business depends on three other questions: who made it, who may watch it and who gets paid.
Dennismary Chidiebere8 October 20255 min read
Operator Notes
Get the next essay in your inbox.
Field notes from building, deciding and carrying the cost of clarity.
Every creator has heard the promise. Own your content. Own your audience. Never again wake up to find a platform has buried your work, changed its rules or switched you off.
This year, with messaging settled, we turned the same question we had asked of chat on Circo's real product: video. If a peer-to-peer design could keep a conversation in the hands of the people having it, could a decentralised design put a creator's videos in the hands of the creator?
It was a serious question, and we gave it a serious look. The answer we arrived at is more useful than yes or no. "Owning a video" turns out to be four different things, and decentralised infrastructure is very good at one of them.
What creators actually mean by "own"
Ask a creator what owning their work means and you hear a cluster of fears, not a definition. That a platform can delete a video. That an account can be closed. That someone can re-upload their work as their own. That the money can stop.
Pull those apart and they sit at different points along the path a video travels: from the creator, as a file, through distribution, to viewers.
The creator wants to be recognised as the author, provably, wherever the video ends up.
The video has to exist somewhere, intact, and stay there for as long as the creator wants it to.
Distribution decides whether the video can be reached, by whom, and whether anyone can quietly take it down.
Viewers watch, and on a paid platform, a viewer watching is supposed to mean a creator getting paid.
Centralised and decentralised designs answer each of those differently. That is the comparison worth making.
Creator → Video → Distribution → Viewers
01 · Creator
Who can prove they made it?
The platform
Its account records say who uploaded the video.
02 · Video
Who holds the file?
The platform
It stores the master copy on its own infrastructure.
03 · Distribution
Who can take it down?
The platform
It can remove the video on its own decision.
04 · Viewers
Who decides who watches, and who is paid?
The platform
Its rights and payment systems decide every view.
Where control sits along a video’s path: Creator, Video, Distribution, Viewers.
Stage
Question
Centralised platform
Decentralised storage
Creator
Who can prove they made it?
The platform. Its account records say who uploaded the video.
The key holder. A signed log proves which key published it. It does not prove who made it.
Video
Who holds the file?
The platform. It stores the master copy on its own infrastructure.
Creator and peers. Any peer can hold a copy and check every block against the creator's signature.
Distribution
Who can take it down?
The platform. It can remove the video on its own decision.
Nobody alone. Every peer that holds the blocks keeps them. Removal becomes a request.
Viewers
Who decides who watches, and who is paid?
The platform. Its rights and payment systems decide every view.
Still a gatekeeper. Paid video in a public swarm must be encrypted, and someone has to issue the keys.
Decentralisation moves custody and removal away from the platform; the decision about who watches still needs a gatekeeper.
Fig. 1Switch the model. Decentralisation moves custody and removal away from the platform; the decision about who watches still needs a gatekeeper.
What decentralised infrastructure does well
The building blocks we explored are elegant, and they solve real problems.
In the Holepunch stack, a creator's files can live in a Hyperdrive, a distributed file system built on signed, append-only logs. The drive is identified by the creator's public key. Every block is signed and checked against a Merkle tree, so a viewer can confirm that what they downloaded is exactly what the creator published, even when it came from a stranger. Readers can fetch only the ranges they need, which matters for video, where nobody wants the whole file before the first frame.
Distribution happens over a swarm. Anyone who has the video can serve it, so it does not depend on one company's servers staying up, or staying willing. A creator who keeps a copy and a few seeds online keeps the video reachable without asking anyone's permission.
That is genuine control over two things: integrity and custody. Nobody can alter the video undetected, and nobody can single-handedly make it disappear.
Where the promise runs out
Then we followed a paid video through the same design, because on Circo most of the value is in paid video: subscriptions, pay-per-view, gifts.
A swarm is public by nature. Anyone with the address can join and fetch the blocks. For a paid video, that means the blocks must be encrypted, and the question moves to the key. Who holds it? Who decides which viewer receives it, and when? Who checks that the viewer has paid?
Whatever answers those questions is a gatekeeper. It has to be online, it has to know about payments and it has to be trusted to say no. The video's bytes can be spread across a thousand peers; the decision to let someone watch cannot. That decision is a business rule, and a business rule needs somewhere to run.
The same is true of removal. A creator may want a video gone, and a platform may need to act on a video that breaks its rules. In a swarm, every peer that holds the blocks keeps them. Decentralisation makes a video hard to take down, which protects the creator against a hostile platform and also protects an abuser against a creator.
And authorship is weaker than it looks. A signed log proves that a key published a file. It does not prove that the person behind the key made the video, and nothing stops someone from downloading it and publishing it again under their own key.
Four kinds of control
By the end of the exploration we had stopped asking "can a creator own a video" and started asking four sharper questions.
Who holds the bytes? Decentralised storage answers this well. Custody can be spread across peers the creator chooses, rather than held by one company.
Who made it? This is provenance: a signed, verifiable record of where content came from and what happened to it. A signed log is a start. A provenance standard that travels with the file is better.
Who may watch? This is access control. On a paid platform it needs a gate that is always on and enforces the rule at playback, not in the app.
Who gets paid? This is commerce, and it is the reason the access gate exists.
Seen that way, decentralised storage is not a complete answer to creator ownership. It is a strong answer to the first question and a starting point for the second. The third and fourth need something accountable in the loop.
What we chose
Circo's video runs on a conventional pipeline: transcoding into adaptive streams, delivery from a content network and multi-DRM protection that is now live in production. Every viewer can fetch the same encrypted segments. The licence that unlocks them is issued only when the platform's rights service confirms the viewer is entitled. That is the gate, and it is where payment turns into playback.
The exploration still changed what we built. It made us separate access from custody, so the encrypted video is the same bytes for everyone and only the licence decides who watches. And it left the provenance question open on purpose, as its own problem, rather than something storage would solve by accident.
That question comes back later in these diaries. First, What Happens When a P2P Video Has 100,000 Viewers? takes the decentralised design at its word, and asks what would happen if one video found a very large audience.
You can own a copy of a video. You can prove you made it. What you cannot do is decentralise the decision about who gets to watch.