OpenAI’s MentalHealthBench is a serious attempt to address a serious problem: Millions of people are already using generative AI for emotionally sensitive conversations, while the models themselves remain probabilistic systems capable of misunderstanding the person on the other side. The instinct behind the work is good. OpenAI has increasingly brought...
AI Hallucinations Aren’t Going Away. Their Authority Can.
One of the strangest things about modern AI is how convincingly it can be wrong. An answer can arrive beautifully written, logically constructed and delivered with the same confidence as something completely accurate. We call these errors “hallucinations,” but the name can make them sound like an unusual malfunction rather...
Where Jev Fits With VERN
Jev is interesting because it separates a decision from the language model generating the conversation. That’s very close to how we’ve been thinking about VERN OS. We don’t expect one AI model to reason, converse, interpret signals, make decisions, govern itself, and execute actions safely. As AI systems become more...