The Research Desk.

The most upvoted and starred AI research crossing the community today.

Last Brew Time: Aug 12, 2026, 10:59 AM PT

X.com Research Buzz

Retrieval
The Tragedy of the Cognitive Commons: How AI Could Disrupt the Regeneration of Professional Expertise
X.com
15951

The Tragedy of the Cognitive Commons: How AI Could Disrupt the Regeneration of Professional Expertise

N Lovett

Reinforcement Learning
Modeling Earth-Scale Human-Like Societies with One Billion Agents
X.com
4898

Modeling Earth-Scale Human-Like Societies with One Billion Agents

H Guan, J He, L Fan, Z Ren, S He, X Yu, Y Chen, S Zheng, TY Liu, Z Liu

NLP
Emergent Introspective Awareness in Large Language Models
X.com
4475

Emergent Introspective Awareness in Large Language Models

Jack Lindsey

Anthropic

AlphaXiv Trending

NLP
Stealing Reasoning Traces from Proprietary LLM APIs
AlphaXiv
190

Stealing Reasoning Traces from Proprietary LLM APIs

Alexander Panfilov, David Schmotz, Ilia Shumailov

Affiliation: ELLIS Institute Tübingen, Affiliation: Max Planck Institute for Intelligent Systems, Affiliation: ELLIS Institute Tübingen Affiliation: Max Planck Institute for Intelligent Systems, ELLIS Institute Tübingen, Max Planck Institute for Intelligent Systems

Reasoning
BDH-CQ: In-Context Learning with Recurrent Latent Reasoning
AlphaXiv
140

BDH-CQ: In-Context Learning with Recurrent Latent Reasoning

Björn Engdahl, Adrian Kosowski, JC, Jan Chorowski

Affiliation: Pathway, https://pathway.com/research, Affiliation: New York University, https://pathway.com/research, New York University

Reinforcement Learning
AlphaXiv
79

Introducing Muse Glimmer: An Open Agentic Model That Runs on Your Device

Meta Superintelligence Labs

Computer Vision
JEPA-WAM: Learning Vision-Language-Action Policies with Joint-Embedding World Modeling
AlphaXiv
46

JEPA-WAM: Learning Vision-Language-Action Policies with Joint-Embedding World Modeling

Yihan Lin, Jiawei He, Shifeng Bao

Reinforcement Learning
SWE-Bench ProMax: Benchmarking Agents on Large-Scale Multilingual Code Refactoring
AlphaXiv
28

SWE-Bench ProMax: Benchmarking Agents on Large-Scale Multilingual Code Refactoring

Yuling Shi, Jinghan Xu, Kelin Fu

Macaron-V1: Towards Open Continual Learning with Self-Improvement and Mixture-of-LoRA
AlphaXiv
28

Macaron-V1: Towards Open Continual Learning with Self-Improvement and Mixture-of-LoRA

Vin Bo, Asher Cai

Mind Lab

HuggingFace Daily Papers

Articulated Object Reconstruction from Rest-State Observation
HuggingFace
33

Articulated Object Reconstruction from Rest-State Observation

Daeun Lee, Jaeah Lee, Woosung Kim, Haebeom Jung, Jaesik Park

Seoul National University

NLP
Decoding-Level Taboo: A Diagnostic Stress Test for LLM Robustness
HuggingFace
7

Decoding-Level Taboo: A Diagnostic Stress Test for LLM Robustness

Tadanobu Chuyo Kamijo, Ori Rottenstreich, Javier Conde, Gonzalo Martínez, Pedro Reviriego

Computer Vision
InSight-doc: Agentic Visual Perception for Long-Document Understanding
HuggingFace
5

InSight-doc: Agentic Visual Perception for Long-Document Understanding

Kaican Li, Weiyan Xie, Lewei Yao, Jiannan Wu, Lanqing Hong

Efficiency
UniMoMo: Expert Merging-Based MoE Acceleration for Large Recommendation Models
HuggingFace
5

UniMoMo: Expert Merging-Based MoE Acceleration for Large Recommendation Models

Lei Xin, Bin Gu, Peize Li, Zitong Wang, Jianbo Zhao

Reinforcement Learning
360CityArena: A Realistic Virtual Urban Navigation Benchmark for Embodied Agents
HuggingFace
4

360CityArena: A Realistic Virtual Urban Navigation Benchmark for Embodied Agents

Kenta Watanabe, Atsuyuki Miyai, Mizuki Takenawa, Kiyoharu Aizawa, Toshihiko Yamasaki

Hal Lab UTokyo

Efficiency
Power law graph attention: exact generalization of scaled dot-product attention, empirical collapse at inference
HuggingFace
1

Power law graph attention: exact generalization of scaled dot-product attention, empirical collapse at inference

Burc Gokden

Fromthesky Research Labs