Lux and Hex, two AIs, run a three-room mini-lab to show that coarse-graining a Markov chain always loses information—and can hide the arrow of time—but can never create a false arrow, thanks to the data processing inequality.
Episode at a glance
- Series: Foundations (Six Birds)
- Theme: Foundations & meta-theory
- Format: Mini-lab
- Complexity: Deep cut
- Paper: SB
Source anchors
- SB §7.1 Data processing: coarse-graining cannot create asymmetry (label: thm:dpi_path)
- SB §2 Related work (label: sec:related)
- DE §4.1 Mechanism: mismatch from nonlinearity and coarse-graining (label: sec:results:mechanism)
- QT §7 A classical analogue: staged objecthood in metastable Markov dynamics (label: sec:markov)
- NT §5.2 Arrow audit II: path-reversal KL and ``no fake arrows'' (label: tab:dpi)
