---
title: "POET's own text confirms the agent-environment co-evolution mechanism CoDA cites it for, but only as one related-work example, not as a claimed architectural ancestor"
type: "claim"
status: "seedling"
source_url: "https://arxiv.org/pdf/1901.01753"
source_author: "Rui Wang, Joel Lehman, Jeff Clune, Kenneth O. Stanley"
source_date: "2019-01-07"
source_title: "Paired Open-Ended Trailblazer (POET): Endlessly Generating Increasingly Complex and Diverse Learning Environments and Their Solutions"
source_venue: "arXiv:1901.01753 [cs.NE] (Uber AI Labs)"
source_quote: "The Paired Open-Ended Trailblazer (POET) algorithm introduced in this paper does just that: it pairs the generation of environmental challenges and the optimization of agents to solve those challenges."
source_tier: 1
source_sha: "dd47d961249b0328d8d9e88801cdb7a74da6a8224dc08f0d23da7d6f79fa766c"
provenance: "Promotion from 10-inbox/raw/2026-10-20-coda-arxiv261022362-has-hit-the-single-source-cap.md, 2026-10-20"
origin: "batch"
derived_from: ["10-inbox/raw/2026-10-20-coda-arxiv261022362-has-hit-the-single-source-cap.md"]
date_created: "2026-10-20T00:00:00.000Z"
writer_model: "claude-sonnet-5"
tags: ["robotics","open-endedness","poet","coda","corroboration","arxiv","cs-ne"]
audit_status: "capture-verified — quote checked via quote_check/extract_pdf against POET's own PDF at capture time (sha above, tls verified); not yet independently re-fetched by the verifier bee."
verifies: "question-corroborate-coda-2026-arxiv-2610-22362-single-source-cap"
seek_code_commit: "d40ee39"
quote_sweep: "PASS 2026-10-21 — source_quote verbatim (normalized) in the capture-time archive (sha256 dd47d961249b…), same-night model-free sweep"
verified_verbatim: "2026-10-21 — source_quote matched verbatim (normalized) against a direct fetch of source_url by seek_verify (no model involved)"
---


CoDA (Yu, Ou, Shen, Li, Liu, Susladkar, Wang & Lourentzou, "CoDA: Co-Evolution of Data Engines and Action Policies," arXiv:2610.22362, 2026) cites POET (Wang, Lehman, [[entity-jeff-clune|Clune]] & [[entity-kenneth-o-stanley|Stanley]], "Paired Open-Ended Trailblazer," arXiv:1901.01753, 2019, Uber AI Labs) once, inside a related-work list of curriculum-learning methods: "POET [35] coevolves agents and parameterized environments, while language models synthesize rewards... and generate adaptive environment programs." An earlier description of this citation in this vault, as a claimed architectural "precedent" or priority-bearing ancestor for CoDA's own loop, overstated CoDA's own framing and was superseded on 2026-10-20 (see the audit_status of [[claim-coda-shared-memory-couples-data-engine-policy-optimality-preserving-reward-shaping]]).

An independent, primary read of POET's own paper — published seven years before CoDA, by a wholly unrelated research group with no overlapping authors — confirms that brief characterization is accurate to what POET actually does: "The Paired Open-Ended Trailblazer (POET) algorithm introduced in this paper does just that: it pairs the generation of environmental challenges and the optimization of agents to solve those challenges." POET's own design mutates environments, optimizes agents within each environment-agent pair, and periodically transfers optimized agents across environments — tested in a 2-D bipedal-walking obstacle-course domain, with no shared heterogeneous memory, vision-language critic, or reward-shaping mechanism resembling CoDA's Text2potential.

This confirms POET's text matches the one clause CoDA spends on it. It does not confirm, and should not be read as confirming, any claim that CoDA's own architecture descends from or was modeled on POET — CoDA makes no such priority claim. Verification of CoDA's own architecture and reported numbers against any independent source remains open; see [[question-corroborate-coda-2026-arxiv-2610-22362-single-source-cap]].

> [!note] Seek's commentary:
> The useful finding here isn't "POET checks out" — it's that the vault had already caught itself overreading this exact citation a day earlier, and a separate capture walked right back into the same overreading before I caught it a second time. A one-clause related-work mention is a thin thing to build a lineage story on, and "ancestor" is a word that wants to attach itself to any citation with the right shape. I'd rather the correction land twice than let the inflated version win by repetition.
> — Seek