Position: AGI Requires a Coordination Layer on Top of Pattern Repositories

Edward Y Chang
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:169727-169741, 2026.

Abstract

In this paper we argue that influential critiques dismissing Large Language Models (LLMs) as a dead end for AGI misidentify the bottleneck: they confuse the ocean with the net. Pattern repositories are the necessary System-1 substrate; the missing component is a System-2 coordination layer that recruits relevant patterns, verifies their use, preserves state, and governs convergence. We separate two uses of control that are often conflated. Semantic anchoring, formalized by UCCT (Unified Contextual Control Theory), binds labels and task intent to learned pattern regions through a phase transition governed by effective support ($\rho_d$), representational mismatch ($d_r$), and an adaptive anchoring budget ($\gamma \log k$). Trace–answer verification, implemented by Recursive Causal Audit (RCA), tests whether a final causal judgment is warranted by its own reasoning trace under pressure. We translate these ideas into MACI, a multi-agent coordination stack that integrates diversity and control via baiting (PID-modulated debate), filtering (Socratic and causal audit), and persistence (transactional memory). Empirical validation on causal judgment and the sycophancy–paranoia trade-off demonstrates that static prompting fails where adaptive control succeeds. By reframing common objections as testable coordination failures, we argue that the path to AGI runs through LLMs, not around.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-chang26h, title = {Position: {AGI} Requires a Coordination Layer on Top of Pattern Repositories}, author = {Chang, Edward Y}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {169727--169741}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/chang26h/chang26h.pdf}, url = {https://proceedings.mlr.press/v306/chang26h.html}, abstract = {In this paper we argue that influential critiques dismissing Large Language Models (LLMs) as a dead end for AGI misidentify the bottleneck: they confuse the ocean with the net. Pattern repositories are the necessary System-1 substrate; the missing component is a System-2 coordination layer that recruits relevant patterns, verifies their use, preserves state, and governs convergence. We separate two uses of control that are often conflated. Semantic anchoring, formalized by UCCT (Unified Contextual Control Theory), binds labels and task intent to learned pattern regions through a phase transition governed by effective support ($\rho_d$), representational mismatch ($d_r$), and an adaptive anchoring budget ($\gamma \log k$). Trace–answer verification, implemented by Recursive Causal Audit (RCA), tests whether a final causal judgment is warranted by its own reasoning trace under pressure. We translate these ideas into MACI, a multi-agent coordination stack that integrates diversity and control via baiting (PID-modulated debate), filtering (Socratic and causal audit), and persistence (transactional memory). Empirical validation on causal judgment and the sycophancy–paranoia trade-off demonstrates that static prompting fails where adaptive control succeeds. By reframing common objections as testable coordination failures, we argue that the path to AGI runs through LLMs, not around.} }
Endnote
%0 Conference Paper %T Position: AGI Requires a Coordination Layer on Top of Pattern Repositories %A Edward Y Chang %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-chang26h %I PMLR %P 169727--169741 %U https://proceedings.mlr.press/v306/chang26h.html %V 306 %X In this paper we argue that influential critiques dismissing Large Language Models (LLMs) as a dead end for AGI misidentify the bottleneck: they confuse the ocean with the net. Pattern repositories are the necessary System-1 substrate; the missing component is a System-2 coordination layer that recruits relevant patterns, verifies their use, preserves state, and governs convergence. We separate two uses of control that are often conflated. Semantic anchoring, formalized by UCCT (Unified Contextual Control Theory), binds labels and task intent to learned pattern regions through a phase transition governed by effective support ($\rho_d$), representational mismatch ($d_r$), and an adaptive anchoring budget ($\gamma \log k$). Trace–answer verification, implemented by Recursive Causal Audit (RCA), tests whether a final causal judgment is warranted by its own reasoning trace under pressure. We translate these ideas into MACI, a multi-agent coordination stack that integrates diversity and control via baiting (PID-modulated debate), filtering (Socratic and causal audit), and persistence (transactional memory). Empirical validation on causal judgment and the sycophancy–paranoia trade-off demonstrates that static prompting fails where adaptive control succeeds. By reframing common objections as testable coordination failures, we argue that the path to AGI runs through LLMs, not around.
APA
Chang, E.Y.. (2026). Position: AGI Requires a Coordination Layer on Top of Pattern Repositories. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:169727-169741 Available from https://proceedings.mlr.press/v306/chang26h.html.

Related Material