On information, intelligence, memory and the boundaries of systems — from AI agents to questions about the nature of reality.
It began with Devs, a question about free will and an unusual workflow with AI models. The decisive moments, however, were not the attractive answers but the times when the original idea broke.
It began with an audacious question: could time, space and stable reality share a deeper informational basis? The result is not an answer to everything, but a map of what holds, what fails, and why that journey produced ITHZ.
The .ithz format did not begin with a plan to compress files at any cost. It began with a question: can a compression decision remain stable, explainable and reproducible regardless of processing order?
How to separate superintelligence from power — and why we may not need to teach it to be “good” first.
OpenClaw and similar agents can use tools without owning the credentials. CCG (Constitutional Capability Governance) moves real authorization into a separate, auditable gateway.
A one-shot token is not enough when the gateway forgets previous steps. The article now includes generation 3: real isolated OpenClaw, an external broker, ITHZ state and 24 auditable runs.
Series If General ITHKOR were true 1/4
If General ITHKOR were true, an electron might not be a tiny thing stored in space. It could be a stable informational pattern that preserves its identity through change.
Series If General ITHKOR were true 2/4
If General ITHKOR were true, space and time might not be the stage on which reality unfolds. They could emerge from relationships among events and their local order.
Series If General ITHKOR were true 3/4
Could three spatial dimensions and dimensionless constants such as the fine-structure constant emerge from a smaller set of informational constraints?
Series If General ITHKOR were true 4/4
What would black holes, dark matter and dark energy look like if the deepest layer of reality were informational? A map of possibilities, boundaries and tests.
Special ITHKOR is no longer only a philosophical idea. It has small controlled models, frozen predictions, negative controls, results from quantum hardware and packages prepared for independent review.