Worldview Diversification

Worldview diversification is Holden Karnofsky's answer to a problem that is as much philosophical as practical: how to allocate enormous resources when the most basic questions about what matters most remain genuinely unresolved. The conventional move is to find the single best answer and pursue it wholeheartedly. Karnofsky saw that this fails when the disagreement runs all the way down—when which cause does the most good depends on assumptions that cannot be settled by evidence alone. He uses the term worldview for a set of highly debatable, perhaps impossible-to-evaluate beliefs that together favor a particular kind of giving, and his insight is that the uncertainty separating worldviews is a tangle of several kinds—empirical, normative, and methodological—that no amount of research can fully resolve. His response is to put significant resources behind each worldview one finds highly plausible. In the cycle that began with [YOU] on AI, the method models a mature relationship with the deep uncertainty that surrounds transformative AI.