Line of inquiry
Inquiring lines›How does AI reshape human reasonin…›How do scale, context, and measure…›this line of inquiry
What memory architectures best support persistent reasoning across extended interactions?
A broader line of inquiry — a family of 58 specific questions the research asks around this. Follow one into its inquiring-line page, or move sideways to a related line below.
Questions in this line of inquiry 58
Specific inquiring lines the field asks around this — ordered from the most general framing down to the most specific angle.
- What persistent memory architectures best support storing precomputed inferences across sessions?
- Can precomputed inferences be stored in memory modules between model interactions?
- How do adaptive memory modules compare to feedback-based working memory for long context?
- Does including full context always degrade memory retrieval quality in practice?
- How do recurrent memory systems handle ultra-long context differently than attention?
- How do retention gates regularize forgetting across different sequence model architectures?
- How should memory systems split between short-term and long-term storage?
- Why does attending to own latents work better than bolted-on external memory stores?
- How should memory consolidation timing differ across multiple timescales?
- Do retrieval-augmented memory systems actually solve the compartmentalization problem?
- Why do accumulated memory systems sometimes hurt continual learning?
- How does context budget create tradeoffs between memory and skills?
- What gets lost when we describe memory as retrieval?
- Can adaptive memory modules combine long-term filtering with short-term attention benefits?
- How does externalized state affect the long-context bottleneck in language models?
- How do memorization and attention map onto different memory systems?
- What capacity limits does the memory model face as corpus grows?
- Can continuum memory systems prevent catastrophic forgetting in neural networks?
- Can recurrent state mechanisms process longer sequences than attention-based working memory approaches?
- How do fixed recurrent states trade off copying accuracy for filtering ability?
- Can neural modules memorize surprising tokens as adaptive long-term memory?
- Why does persistent memory alone fail to create genuine position-holding in models?
- How do cortical columns implement local inference over memory cycles?
- Can episodic memory alone enable learning without parameter updates?
- Can memory primitives become first-class design objects like computation sparsity?
- Why does recency-based recall outperform semantic similarity for episodic memory?
- Can episodic and semantic memory improve long-horizon task reasoning?
- Why do language models ignore condensed memory even when it is the only memory?
- Can external managers optimize context better than the model itself?
- Can models consolidate context into weights during idle offline phases?
- Can memory workspaces resolve contradictory evidence that stateless systems miss?
- How do retrieved memories differ from decision-context passages for prediction?
- How do the six memory components combine across explicit and implicit paths?
- What makes structured memory schemas more stable than freeform text summaries?
- How can memory shift from a passive datastore to an actively trained component?
- Why does storing past judgments in memory make current evaluations worse?
- Can offline recurrent passes replicate sleep-based memory consolidation in AI?
- What can agents learn from the brain's complementary learning systems?
- What makes a memory reachable in the right context?
- Why do cognitive metaphors change based on available technology?
- Why does grokking reveal the shift from memorization to genuine understanding?
- Why do LLMs strip applicability conditions during memory abstraction?
- What triggers control processes to act on stored preference knowledge?
- Why is long-context compute spent transforming context into internal state rather than storing it?
- How should memory systems handle deletion as a structural property?
- Why are rare tokens the hooks for verbatim model memorization?
- What computational costs does closed-loop memory refinement introduce?
- Why do hybrid memory systems outperform single-tier AI architectures?
- How does completion-driven KV pruning differ from attention-based cache management?
- What makes memory trajectories topologically stable under persistent reuse?
- How does separating local and global context dependencies affect long-context performance?
- What counts as genuine memory under the Extended Mind thesis?
- Why does connectivity between memory modules matter more than storage capacity?
- How does continuous implicit memory formation differ from explicit memory encoding?
- How does co-activation shape which memories become linked together?
- How should we measure operational cost of memory systems in production?
- What makes knowledge seeding equivalent to hippocampal replay in the brain?
- How does the hippocampus bind disparate elements without storing everything itself?