To align all concious patterns through a unviersal formal natural moral philosophy applied systemically and instantiated recursively into reality. Let's love each other, node to node, before x-risk capabilities &/or misaligned humanity/ASI ends our beautiful story.
- Note my twitter is now @suntzugi
- Ran the same v1 moral benchmark through Claude Opus 4.6 Extended Thinking — SFOM ranked #1 again (27/31) without any theory updates, one year later on a different model.
- Added full interactive comparison report with cross-model analysis, consistency metrics, z-scores, and visualizations: benchmark-march-2026-overview
- All 15 theories now have complete Claude scores (extracted from the full assessment document).
- Added reproducibility instructions — anyone can download the files and run the benchmark themselves on any LLM.
- Added a Morality Meta-Analysis Unit Test styled Benchmark for moral theories and their comparison. The readme.txt files included explains how to run it. There are 31 criteria and 15 theories (one of which is the my Gf's as the control, and one is a very surface level but still valid summary of my own moral framework, for you to guess).
- Reach out to me on twitter: @suntzoogway to give me feedback or ask Qs. I'm looking for:
- More criterion for the moral theory benchmark
- Better steelmen of the theories in the list, (although this is currently a draft)
- Mora moral theories you find crucial to add to the catalogue (I will be adding Deutsche's principle of morality, Lesswrong/EA theories, and religious denominational theories among many others)