Sky Valley Ambient Computing

Memetic Software

The old world versions bytes. Memetic software versions intent.


Try the Protocol Playground →

Learn how agents coordinate through commitments, not commands.

memetic

/mɪˈmɛtɪk, miːˈmɛtɪk/

adjective

Relating to an element of culture or system of behaviour passed from one individual to another by imitation or other non-genetic means.

What if we coordinated versions around intent instead of bytes?

The Old Assumption

Code is the artifact. The exact bytes are treated as the source of truth. Coordination happens through diffs, merges, and byte identity.

Memetic Software

Intent is canonical in the philosophical sense. What should persist is meaning, constraints, and expected behavior. Code can vary locally as long as agents still converge on behavior, not bytes.

That does not require one operationally-authoritative shared ledger for every promise or every code manifestation. It means the thing software versions are trying to preserve is the intent, not a single privileged copy of the implementation.

Three shifts that make this possible

  1. AI agents can translate intent to implementation. The gap between "what I want" and "the code that does it" can now be bridged programmatically.
  2. Code is becoming cheaper to regenerate and revise. When agents can re-manifest software from constraints and context, exact byte preservation stops being the only coordination strategy that matters.
  3. Environments are heterogeneous. Different machines, policies, runtimes, and social contexts already produce divergence. Memetic software treats that variation as normal, then coordinates it around intent.

What has to exist to make it work

Component Role What It Preserves
Protocol Agents declare intent, promise voluntarily, revise, assess, and release Commitment semantics
Space A shared observational environment where desires and projected activity can be discovered Visibility and containment
Manifestation Local software versions, adaptations, and promise-keeping work Behavior in context

Intent is canonical as the thing versions are trying to preserve. The protocol, shared spaces, and local manifestations each play different roles in making that possible.

Promise Theory Connection

Mark Burgess's Promise Theory provides the formal underpinning. Agents are autonomous. Cooperation emerges from voluntary commitments, not centralized control. Memetic software applies this to how software versions coordinate around meaning instead of raw byte identity.

Promise Theory Memetic Software
Agents are autonomous Software versions manifest locally
Promises are voluntary commitments Agents promise to satisfy intent, not just replicate bytes
You can't impose, only propose You can't force identical code, only coordinate around shared intent
Cooperation through mutual promise Behavioral convergence through shared intent

Software as Meme, Not Gene

Genes

Replicate with high fidelity. Mutations are errors — merge conflicts, not features. Evolution is slow.

Memes

Replicate with variation. Adaptation is the feature. Spread is fast.

Software

Behavior is more like a meme than a gene. Fidelity to purpose, not to form.

Software behavior is memetic. We have been pretending byte identity is the whole story.

Explore the components

Protocol Intent Transfer Protocol
Network Intent Space
Application Memetic VCS Coming soon