home / issues
Issues
Long-form essays on failure modes in human, artificial, and hybrid intelligence. Fourteen categories. 4 of 14 written; the rest are planned.
- #001 Recall as reconstruction, not playback
- #002 Reasoning plannedWhen the conclusion arrives before the work that should support it
- #003 Confident answers that outrun actual accuracy
- #004 Confident interpretation of sensory input that diverges from reality
- #005 Generalization plannedKnowing a thing in one context but not in another
- #006 Patterned error that becomes institutional through deference
- #007 Introspection plannedWhen the agent's account of its own reasoning isn't actually what happened
- #008 Suggestibility plannedWhen external input rewrites internal state more than it should
- #009 Attention plannedFinite capacity to attend, and the failures that follow from spending it badly
- #010 Theory of mind plannedModeling the minds of others — and being modeled by them
- #011 Temporal reasoning plannedTracking time, sequence, and the difference between "now" and "then"
- #012 Numeric reasoning plannedQuantities, magnitudes, and the failures specific to handling them
- #013 Causal reasoning plannedDistinguishing what causes what from what merely co-occurs
- #014 Language plannedParsing, producing, and being shaped by language