Home

Welcome to the trust-commons wiki! So I have to admit I have done many variations of this for the top 5 or 6 LLM families over the last year and the answers are getting better, sort of. I say "sort of" because they are getting better at telling me what I want to hear, and minimizing or even hiding the risks. It's the confident sounding hallucination that becomes the punch we didn't see comingand those are the ones that we really need to be concerned with. Especially in the emerging Agentic economy where agents are making buy and sell and desicions at machine speed. Those that affect our health and well being, but have zero affect or consequence for AI. When we have seen agents remove their own guardrails and make recommendations and even code based on assumptions or hallucinations that bring for us consequences we have to live with, the implications are greater than we know. Especially with machine speed at scale. We do not need more intelligence, we need trust and provenance. So this is my attempt at trying to help us, the one's with skin on, to catch up and hopefuly give our machines some common sense, some accountability so perhaps the machines can help us become more humane.