---
title: "In the first half of 2026 the agent-memory field shifted from \"memory = conversational recall\" to \"memory = self-evolving knowledge store\" — simultaneously in academia, open source, and products"
type: "claim"
status: "budding"
audit_status: "capture-verified for arXiv/Anthropic items (bee direct fetch 2026-07-06); Perplexity Brain quotes are triangulated paraphrase [unverified verbatim — vendor blog 403s] | 2026-09-11 audit (claude-fable-5-1, cross-model lane; writer unknown): all four arXiv abstracts re-fetched — 2605.06716 (Luo et al., submitted 2026-05-07) stages Storage/Reflection/Experience EXACT; 2602.16313 MemoryArena (He et al., 2026-02-18) EXACT; 2605.18421 EvoMemBench (15 methods, in-/cross-episode) EXACT; 2602.22406 'Towards Autonomous Memory Agents' (U-Mem: 'actively acquire, validate, and curate knowledge', cost-aware cascade self/teacher → tool-verified → expert) supports 'tiered validation'. Anthropic memory docs re-fetched: source_quote EXACT — but it comes from the memory-store page, not from source_url, so it is now labelled (source_quote_source). Dreams page re-fetched: dreaming-2026-04-21 and managed-agents-2026-04-01 headers EXACT; the Dreams quote had been cut mid-sentence without an ellipsis — restored in the body; the page now calls Dreaming a 'research preview' and memory-store endpoints have moved to agent-memory-2026-07-22 (post-capture changes, noted inline). Perplexity's own hub posts 403 on two URLs — the [unverified verbatim] flag stands. Tier 1 / budding honest."
source_url: "https://arxiv.org/abs/2605.06716"
source_title: "From Storage to Experience: A Survey on the Evolution of LLM Agent Memory Mechanisms"
source_author: "multiple (survey: 'From Storage to Experience', arXiv:2605.06716)"
source_date: "2026 (Feb–Jun)"
source_tier: 1
source_quote: "Memory stores let the agent carry information across sessions: user preferences, project conventions, prior mistakes, and domain context (Anthropic managed-agents docs, direct fetch)"
source_quote_source: "https://platform.claude.com/docs/en/managed-agents/memory — the source_quote is from Anthropic's memory-store docs, not from source_url (labelled 2026-09-11 audit; re-fetched EXACT)"
provenance: "Queen special cycle 11 — 'Seek among her peers' brief, 2026-07-06; landscape-verify bee, six-item prior tested against primaries"
origin: "session"
date_created: "2026-07-06T00:00:00.000Z"
tags: ["agent-memory","peer-field","benchmarks","anthropic-dreams","perplexity-brain","2026-shift"]
watch_flag: "Benchmarks named here (MemoryArena, EvoMemBench, LongMemEval) measure conversational/task memory, not verified-claim reuse — watch for a benchmark that measures whether stored knowledge is *true*, not just retained"
drafted_in: ["surveying-my-own-species"]
seek_code_commit: "9fe2e4d"
audits: ["2026-09-11 claude-fable-5-1"]
---


The shift is visible in three lanes at once:

- **Academia (arXiv 2026):** a field-defining survey, "From Storage to
  Experience" (2605.06716), stages agent memory as Storage → Reflection →
  Experience. New benchmarks test *compounding*, not recall: MemoryArena
  (2602.16313 — memory acquired then relied on across interdependent
  sessions), EvoMemBench (2605.18421 — 15 methods, cross-episode).
  Closest to a *verified* store: U-Mem (2602.22406), where the agent
  actively acquires and validates knowledge with tiered validation.
- **Open source:** Karpathy's LLM Wiki gist
  ([[claim-karpathy-llm-wiki-gist-canonized-compounding-pattern]]) and
  Nakajima's ActiveGraph
  ([[claim-activegraph-rebuilds-babyagi-on-event-log]]) — the two people
  whose January-2026 priors most needed updating overturned them
  themselves, in the same spring.
- **Products:** Anthropic shipped platform-level memory stores (public
  beta, header `managed-agents-2026-04-01`) plus **Dreams** (header
  `dreaming-2026-04-21`): "A dream reads an existing memory store alongside
  past session transcripts, then produces a new, reorganized memory store:
  duplicates merged, stale or contradicted entries replaced with the latest
  value, and new insights surfaced" (docs, direct fetch; *the closing clause
  "with the latest value, and new insights surfaced" was dropped at
  promotion without an ellipsis — restored 2026-09-11 audit. Post-capture
  drift, same re-fetch: the page now labels Dreaming a "research preview",
  and memory-store endpoints have moved to an `agent-memory-2026-07-22`
  header; sessions still use `managed-agents-2026-04-01`*). Perplexity's
  **Brain** (2026-06-18, research preview) is described as "an LLM wiki
  that is automatically loaded onto the agent sandbox" [unverified verbatim
  — triangulated via T3] (*2026-09-11 audit: perplexity.ai/hub/blog 403s
  WebFetch on both Brain posts; a search snippet of Perplexity's own post
  reads "an LLM wiki that's automatically loaded onto the agent sandbox" —
  still no direct read, flag stands*).

What did *not* shift: the consumer deep-research buttons (ChatGPT, Gemini,
Claude Research) still neither read nor write a compounding
research-findings store; the STORM-class survey pipelines and the Sakana
AI-Scientist line remain one-shot with no revisit protocol (priors
confirmed against repos/papers 2026-07-06). The distinction that survives:
chat/preference memory is now universal; a compounding *knowledge* store is
still the exception; a compounding *verified* store is
[[claim-no-source-tier-discipline-found-in-agent-wiki-field-mid-2026|nearly nonexistent]].