Metamodern Meaning The Physics of Meaning w/ Artemy Kolchinsky
10 snips
Jul 15, 2026 Artemy Kolchinsky, a researcher in informatics and complex systems who studies semantic information and thermodynamics. He explains how meaning can be formalized as correlations that support a system’s viability. They cover informational knockouts, forager simulations showing embodiment shapes meaning, tradeoffs between information cost and benefit, and implications for evolution and AI alignment.
AI Snips
Chapters
Transcript
Episode notes
Meaning As Viability-Preserving Information
- Meaning can be formalized as a subset of correlations that causally help a system maintain its existence (viability).
- Kolchinsky ties this to autonomy and non-equilibrium thermodynamics: systems must acquire free energy to stay organized, creating intrinsic norms of good/bad interactions.
Knockout Tests Define Meaningful Information
- Semantic information value is measured by a knockout: scramble correlations and see how the system's viability changes.
- The 'viability function' (e.g., low entropy or available free energy) quantifies how much information causally contributes to survival.
Embodiment Determines What Information Matters
- Embodiment shapes which correlations are meaningful: an agent's size and action capabilities set resolution limits on useful information.
- The forager model showed extra precision (beyond bodily resolution) yields no added viability benefit, so those bits are meaningless.