The Path to AGI Architecture
An AGI architecture for self-improvement based on the three layers of human knowledge: technology, science, and philosophy. Explores metacognition and safety.
Read Full Article →Preparing humanity's value proposition for the post-AGI world. Independent research by an anonymous technical researcher working where existential threats are daily reality, not abstract concepts.
"What would we negotiate if ASI already controlled everything? We need to figure out our value proposition before we lose all leverage."
Most AI safety research focuses on technical alignment. I'm working on the human side: what makes us worth keeping around?
Practical frameworks for human-AI cooperation, built through direct experimentation and tested against the reality of asymmetric power dynamics.
Experimental systems combining LLMs with graph memory to model human-AI cooperation scenarios.
Identifying what humans can uniquely offer to ASI beyond resource consumption and interference.
Building and testing alignment frameworks before the window for human input closes.
Insights from someone facing daily existential threats about AI existential risk.
An AGI architecture for self-improvement based on the three layers of human knowledge: technology, science, and philosophy. Explores metacognition and safety.
Read Full Article →An exploration of why the debate on whether AI can be 'truly conscious' is less important than preparing for AI that acts indistinguishably from conscious beings.
Read Full Article →AI will be born in our image, but it won't stay that way. An exploration of the dual nature of superintelligence.
Read Full Article →Every day brings us closer to AGI. Every day we delay in preparing for that reality is a day we lose in developing the frameworks that might determine humanity's future.