LLMs Struggle to Perceive the Passage of Time
Economy-Builder7916 · reddit · 2026-07-12
The author observed that within a single long thread, LLMs seem indifferent to the recency of messages, treating "5 minutes ago" and "a few days ago" on the same temporal plane.
This limitation impacts various multi-day continuous use cases:
- Multi-day programming projects
- Journaling and reflective logging
- Tracking fitness, diet, budget, habits, or medication
- Multi-day negotiations and planning
- Spaced repetition learning
The core issue is the model's lack of temporal awareness regarding "how old is this information." Consequently, it cannot determine if data is outdated or proactively prompt the user for follow-ups and reviews. The author is investigating whether this is a universal problem and if different products handle it more effectively.
More from Models
- Kimi K3.1 Pro tops MathArena open models, closing frontier gap to about 6 months — teortaxesTex · 2026-10-26
- Ruff creator Charlie Marsh: frontier models find infinite bugs in any software — charliermarsh · 2026-10-26
- Dev testing one skill across Opus 5.5 and Sol 6.1 finds instruction following 'all over the map' — lucasmeijer · 2026-10-26
- supermemory unveils learner-1, claiming continual learning for any model — julianweisser · 2026-10-26
- Anthropic says Opus 5.5 cuts cost per task by ~40% vs Opus 5 — taufiqintech · 2026-10-26
- AI search APIs ranked: top 12 within 7 points, yet prices differ nearly 7x — FinanceYF5 · 2026-10-26