Recent posts tagged llm-architecture

The problem: The industry is treating AI alignment as a training problem. We fine-tune models, add RLHF, and write better system prompts. But alignment is not the same as sovereignty. A model can be perfectly aligned and still have no identity, no r...

For anyone building or deploying AI models: I just released Salus — a constitutional enforcement layer that sits outside your model. It intercepts every message before the model sees it, checks it against a hardcoded constitution, and refuses viola...