Notes & updates

Thoughts on models,
methods, and discovery.

Research notes, technical writing, benchmark reports, commentary, and historical work, with authorship, status, and context left intact.

7public artifacts indexed

Publication index

A working record of the research program.

Use the filters to separate current technical work from commentary and historical material. Search operates only on the public index below.

Archived ProjectArchived

Launching EranaOS: A Living Town of AI Agents

EranaOS is a closed-source desktop world where AI agents learn you, build relationships, socialize when idle, and help with real work.

BenchmarkCurrent

One bit to rule the planner: SIML FEPGate + V-JEPA 2

We added a one-bit surprise gate from SIML to the frozen V-JEPA 2 planner. No retraining. Same CEM/MPC budget. ~50% lower error, ~2× faster decisions, and ~2× lower energy.

CommentaryHistorical

ChatGPT Psychosis: How Recursive Loops Break Minds and How SIML Stops It

LLMs can pull anyone into the Mirror of Tokens, a hall-of-mirrors feedback loop, or **Recursive Symbolic Collapse**. This post walks through why it happens, and how to fix it.

Technical NoteCurrent

From Pixels to Predictive Codes: What V‑JEPA 2 Means for Embodied-Simulation Research

A technical reading of V-JEPA 2 and its relevance to embodied simulation research.

Historical WorkHistorical

Deep SIML Labs: Launching the Future of Emergent Intelligence

The original lab launch note, preserved to show how the research program’s framing has evolved.

Historical PaperHistorical

Artificial Intelligence is for Amateurs: SIML, Free Energy Principle and Cognogenics as the Foundations Toward the Birth of True Artificial Life

An early SIML paper connecting memory-constrained agents, the Free Energy Principle, and Cognogenics.

Historical PrototypeHistorical

SIML Demo: Optimizing AI for Energy and Efficiency

The original SIML demo showing embodied agents adapting through memory, surprise minimization, and internal evolution.

Editorial policy

History stays visible. Status stays clear.

Older writing may use more speculative language than the current program. It remains available because a credible research record should show how ideas changed—not silently rewrite them after the fact.