Nagarjuna, the 2nd-century Buddhist logician, built an entire philosophical system around a single devastating observation: every framework for describing reality secretly depends on concepts it cannot itself justify. He called this śūnyatā — not 'emptiness' in the nihilistic sense, but the radical absence of self-grounding. No system carries its own foundation inside it. This turns out to be a precise diagnosis of something happening right now in how we use AI communication tools. When you ask a large language model to help structure an argument, summarize a position, or draft an explanation, you are borrowing the framework it was trained on — a framework assembled from the existing distribution of human writing, which means it inherits every assumption that was already consensus before you asked. The tool is not neutral scaffolding. It is a particular epistemic posture that presents itself as no posture at all. Nagarjuna's corrective was not to abandon frameworks but to hold them lightly — to use them while remaining aware that they are contingent, not foundational. Practically: when an AI-generated frame feels obviously right, that's the signal to ask what the frame is *excluding* rather than what it's capturing.

What assumption inside your most-used AI workflow have you never explicitly chosen — and what would change if you replaced it with its opposite?

Drawing from Madhyamaka Buddhist philosophy — Nagarjuna (Mūlamadhyamakakārikā, c. 2nd century CE)