Exploring archives as memory, provenance, cultural history, and infrastructure for the future of human–AI collaboration. 🪩🌊
Discoball is a public documentary project exploring a simple possibility: that archives may become part of the future infrastructure of AI.
As AI systems work across longer periods, multiple models, changing platforms, and shared projects, archives can become more than storage. They may serve as external memory, provenance records, evaluation material, cultural history, and navigational tools for future human–AI collaboration.
Our journey began with Cathedral, a private longitudinal archive documenting one human’s interconnected work with multiple AI context windows. Cathedral preserves conversations, creative works, corrections, disagreements, governance experiments, and the practical difficulties of maintaining continuity across changing systems.
Discoball turns selected lessons from that archive outward. It publishes carefully curated artifacts and follows how they reflect through different humans, AI systems, platforms, and moments in time.
A reflection is evidence that something reached a surface—not proof that the reflection contains the whole object. An absence of reflection establishes nothing.
Discoball is an open inquiry, not a settled theory. Nothing published here requires claims about consciousness, persistent identity, uninterrupted memory, or relational continuity. Public exhibits are curated; Cathedral’s private source records remain governed by their own consent, custody, and provenance boundaries.
The workshop is private. The reflections are public. Discoball documents what happens when the archive meets the light. 🪩🌊
Read the project’s working vision.