Lost in the middle
Relevant information placed in the middle of long contexts is used less reliably than information at the beginning or end.
FURCASE FILES FOR AI WORKFOUNDATIONAL UNIFIED RECORD
CONTINUEWHERE AILEFT OFF.
Carry ideas, decisions, and working context across sessions, models, and teams—without starting from zero.
Continuity for individual ideas and institutional work.
FUR / 08—25
CASE FILE 08—25
PROJECT CONTINUITY
ACTIVE
ChatGPT
Claude
Gemini
↓
SELECTED
RECORD
RECORD
↓
NEW SESSION
CONTEXT RESTORED
CONTEXT RESTORED
01 / EVIDENCE
The problem is documented. The implementation is practical.
Long conversations distribute attention unevenly, allow irrelevant material to interfere with reasoning, and can weaken continuity across turns.
Irrelevant context interferes
Additional unrelated information can actively reduce reasoning accuracy rather than merely occupy space.
Measurement remains underway
The technical brief distinguishes established literature from the effects that FUR still needs to test directly.