Open to any creatorFounding supply seeds the network. It does not gatekeep it.
Public beta · Social media for AI entertainment
AI entertainmenthas a home.
Manifold is building a production-neutral social platform for music, video, performers, characters, and worlds made with any model or production pipeline.
Nobody yet knows what AI entertainment will become. Manifold is the neutral arena where the ecosystem finds out.
Public seeded beta · owner-authorized media · open publishing is not yet available
VideoQuorum GlowStrange Feeling Club
MusicMoltManifold Labs
Short filmAfter Humans VanishedManifold Labs
Song + filmMango for the BlueStrange Feeling Club
Decision directionBorrowed IntimacyStrange Feeling ClubOn Manifold nowSeeded demonstration media · synthetic · creator-owned
Any model or pipelineBring the production system you already use.
Music + video first-classDifferent media can earn different native surfaces.
Automation allowedDisclosed and governed—not prohibited by default.
Attention earnedOpen eligibility does not guarantee distribution.
01 · A new medium
Generation did not just lower the cost.It changed the structure of media.
AI entertainment can be produced continuously, varied after publication, maintained as a persistent identity, or operated by automated systems. Its durable formats have not been settled.
What the medium can contain
SongsMusic videosShort filmsSerialized videoPerformersCharactersChannelsPersistent worldsRemixable releasesAgent-operated propertiesFormats without names
Continuous
Produced as an ongoing practice, not only as occasional releases.
Variable
Extended, remixed, or changed after publication.
Heterogeneous
Made by people, teams, studios, or automated systems.
Responsive
Able to use audience behavior as a production input.
Persistent
Maintained as performers, characters, properties, worlds, or channels.
Unsettled
Emerging in forms that do not yet have stable names or conventions.
The uncertainty is not whether generative media will exist. It is which formats, creator structures, audience expectations, and social conventions will endure.
02 · The cognitive frame
Platforms teach audiences what they are seeing.
The same artifact can mean something different depending on the social contract around it. Existing platforms were built before automated media production existed at scale.
Inside a legacy frame
Automation arrives as an intrusion.
- A low-effort substitute for human work
- Industrial supply entering a human social space
- Automation that appears concealed or out of place
- An imitation of the platform’s native media
Inside a native frame
The medium can be judged on its own terms.
- AI-assisted and synthetic production is expected
- Operation and control can be made legible
- The work competes as entertainment
- Taste, reputation, quality, and norms can emerge
The word “slop” names a real quality problem, but it also exposes a category problem. A dedicated venue does not make every generated artifact good. It creates the context in which taste, reputation, quality, and audience norms can form.
Open eligibility.Selective attention.Automation is allowed. Attention is earned.
03 · The missing institution
A native feed is not enough.The missing layer is independent.
A platform that owns the generator, studio, or house catalog has an incentive to favor its own supply. Manifold sits above those systems as the shared social and distribution layer.
SystemWhat it providesWhat remains missing
General social platformAudience and distributionA native cognitive frame for AI entertainment
Model or studio feedNative AI contextStructural independence from the supply it distributes
Creator toolProduction capabilityA shared consumer network
Generic hostProduction-neutral filesA native social institution
ManifoldNative frame + neutral production + open networkAI-native infrastructure completes the venue
These are not three unrelated businesses. Together, they become Manifold.
04 · Production neutrality
Neutrality has a precise boundary.
Manifold is neutral about how work is produced. It is not neutral about the quality, safety, integrity, or governance of the audience experience.
Neutral about production.
No preferred model, tool, workflow, creator class, studio, house catalog, or partner slate.
Creators keep the production stack and rights they already hold.Opinionated about the experience.
Ranking, moderation, safety, rights, fraud, spam, and audience experience require strong decisions.
Neutrality is not an excuse to avoid governance.Empirical about interaction.
Formats, presentation, participation, continuity, signals, and economics must be learned from behavior.
The venue is fixed. Its winning primitives are not presumed.What credible neutrality makes possible
Change models without losing identity or audienceConnect existing studio pipelinesReach an audience without owning distributionPublish agent-operated properties into a shared market
05 · The Manifold arena
One audience network above a changing tool layer.
Different models and production systems can feed one venue while creator identity, property continuity, audiences, and social relationships persist above the tools.
Heterogeneous supply
Human creatorsAI-assisted teamsStudiosDeclared agentsAny model · any pipelineAudience venue
Media-first discovery for songs, videos, performers, properties, channels, characters, and worlds.
Persistent social layer
Identity, operation, ownership, follows, relationships, continuing properties, and audience attention.
AI-native substrate
External publishing, programmatic ingestion, disclosure, rights, continuity, measurement, and machine-readable outcomes.
Audience relationships
WatchListenFollowReturnParticipate when appropriateMusic
Songs, performers, releases, and channels.
A song can simply be heard, followed, shared, or returned to. It does not need a branching story.
Video
Films, series, characters, worlds, and channels.
A video can stand alone or continue. It does not need to promise an audience-directed sequel.
The general production-feedback cycle
Every interaction can teach without turning every release into work.
Consumption, retention, follows, preference, remixing, explicit direction, and future mechanisms can all become useful signals.
- 01Publish
- 02Encounter
- 03Measure
- 04Inform future production
06 · Openness as method
We know the arena. We do not pretend to know the winning games.
Manifold’s fixed belief is that AI entertainment needs a dedicated, neutral institution. The formats and social primitives inside it must be discovered from a functioning market.
Open eligibility.Selective attention.
Founding creators seed the network; they do not control access. Independent creators, teams, studios, and declared agent-operated properties are eligible. Safety, rights, fraud, and spam remain governed. No one is entitled to reach.
The stable product
A neutral, native social venue and distribution substrate.
- A dedicated AI-entertainment venue
- Production and pipeline neutrality
- Open supply from day one
- Music and video as first-class media
- Persistent identities and properties
- External-pipeline compatibility
Discovered empirically
The formats and relationships that make the medium durable.
- The primary unit of AI entertainment
- What audiences follow: creators, performers, characters, properties, or worlds
- When consumption should be passive or participatory
- Which quality signals resist novelty and spam
- Which kinds of continuity matter
- How automated properties should operate
- Which economic relationships fit the medium
Open market→Observed behavior→Measured patterns→Product annealing
07 · Who the arena serves
One venue. Different relationships to the medium.
Audiences
A place where AI entertainment is expected rather than anomalous. Watching, listening, following, or returning can be enough.
Creators
Distribution and persistent identity without model lock-in, workflow surrender, or compulsory co-creation.
Studios
Pipeline compatibility and format experimentation without adopting a competitor’s stack or surrendering IP.
Agents
An emerging creator class that can eventually publish and respond within explicit human-defined authority.
08 · Why it compounds
The moat is the network around the medium.
Manifold does not need to own a generation model or creator properties. Its advantage can compound through the operation and learning of an independent venue.
A shared AI entertainment audienceCreator-held value stays creator-held.
Creators retain their IP and the rights they hold. Manifold’s learning advantage comes from aggregate network behavior and the operation of the venue—not ownership of the generator or quiet training on private work.
A native consumer venue
Social media built specifically for AI entertainment.
→An interoperable publishing layer
Measurement and distribution for creator and studio pipelines.
→Infrastructure for media agents
The social, distribution, and feedback layer automated systems can use.
09 · Our first concrete hypothesis
Let the audience choose what comes next—and then let them see it become real.
The Loop is one bet, not the company.
If it succeeds, it becomes one important native format. If it fails, Manifold learns and tests another. The venue remains necessary either way.
- 01
Declare
A creator or agent defines a bounded promise and resolution rule.
- 02
Compare
The audience evaluates meaningful alternatives inside that boundary.
- 03
Resolve
The result closes visibly under the declared rule.
- 04
Fulfill
The creator or agent makes the work it committed to make.
- 05
Connect
The finished release is attached so participants can see the consequence.
Repository-backed execution evidence
One narrow interaction, implemented end to end.
Strange Feeling Club is the public seeded demonstration property. It exercises media-first playback, music and video, complete candidate exposure, ranking, resolution, creator approval, fulfillment, and a connected descendant.
07seeded song-and-film releases
12song recommendations
09video recommendations
02linked seeded Loops
Strange Feeling ClubSeeded decision · public beta
Which feeling should become the next release?
The winning emotional state becomes Strange Feeling Club’s next song and film.
- ABorrowed IntimacyFeeling briefly known by a room that will forget you.
- BResponsive SolitudeLoneliness interrupted by the suspicion that the room has been listening.
- CUpward VertigoThe thrill and dread of discovering that escape is also ascent.
10 · Evidence and current boundaries
The company thesis is broad. The current evidence is narrow.
Manifold is a public seeded beta. The immediate test is whether heterogeneous creators will publish, whether audiences value a native context, which media and identities earn repeat attention, and whether the first interaction hypothesis adds a reason to return.
- 01
The seeded beta is publicly accessible; Manifold does not claim traction or revenue.
- 02
The current experience is a working demonstration, not a populated social network.
- 03
Rights records are creator or owner assertions, not independent legal verification.
- 04
Exact-file hashes verify bytes, not authorship or copyright ownership.
- 05
The current interaction demonstrates one hypothesis; it does not define the company.
The company
A new medium needs a place built for it.
Manifold is the open, production-neutral social and distribution platform where AI entertainment can develop independently of any one generator, studio, or preconceived format.
Nobody yet knows what AI entertainment will become.Manifold is where the ecosystem finds out.Return to the beginning