The Research Desk.

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

Last Brew Time: Aug 20, 2026, 10:59 AM PT

X.com Research Buzz

The AI Layoff Trap
X.com
20359

The AI Layoff Trap

Brett Hemenway Falk, Gerry Tsoukalas

University of Pennsylvania - The Wharton School, Boston University

NLP
Mind Viruses: Self-Propagating Ideas in Multi-Agent LLM Systems
X.com
7348

Mind Viruses: Self-Propagating Ideas in Multi-Agent LLM Systems

Vassilis Papadopoulos, McNair Shah, Sam Zimmerman, Jack Lindsey

Mathematics in the age of AI
X.com
2522

Mathematics in the age of AI

Terence Tao

AlphaXiv Trending

Mathematics in the age of AI
AlphaXiv
238

Mathematics in the age of AI

Terence Tao

Address: UCLA Department of Mathematics, Los Angeles, CA 90095-1555, UCLA Department of Mathematics, Los Angeles, CA 90095-1555

Robotics
Autonomous De Novo Protein Binder Design with Claude
AlphaXiv
96

Autonomous De Novo Protein Binder Design with Claude

Amir Shanehsazzadeh

Recirculation
AlphaXiv
85

Recirculation

Michael C. Mozer, Shoaib Ahmed Siddiqui, Danny Sawyer

Thanks: Work performed while a student researcher at DeepMind, Michael C. MozerGoogle DeepMind, Shoaib Ahmed SiddiquiGoogle DeepMind, Danny SawyerGoogle DeepMind, Rosanne LiuGoogle DeepMind

Computer Vision
HarnessEval-W: Agentifying the Evaluation of Visual Worlds
AlphaXiv
84

HarnessEval-W: Agentifying the Evaluation of Visual Worlds

Weiliang Chen, Haowen Sun, Jun Gao

https://github.com/mirros-lab/harnesseval-w, https://mirros-lab.github.io/HarnessEval-W

Machine Learning
Agentic ESOpt: Fine-Tuning Long-Horizon LLM Agents with Minimal GPU Requirements
AlphaXiv
67

Agentic ESOpt: Fine-Tuning Long-Horizon LLM Agents with Minimal GPU Requirements

Zhi Zheng, Rongsheng Chen, Yunpeng Ba

† † affiliation: National University of Singapore, † † affiliation: Southern University of Science and Technology, NUS — National University of Singapore

Hydra-0: Action Flow for Generalist World Modeling and Control
AlphaXiv
46

Hydra-0: Action Flow for Generalist World Modeling and Control

Hongyu Li, Bowen Wen, Xinghao Zhu

NVIDIA, Brown University, Columbia University, Harvard University

HuggingFace Daily Papers

NLP
The More Popular, The Harder to Forget: Adaptive Popularity for LLM Unlearning
HuggingFace
11

The More Popular, The Harder to Forget: Adaptive Popularity for LLM Unlearning

Anna Borisiuk, Andrey Savchenko, Alexander Panchenko, Elena Tutubalina

Safety Alignment
Temporal Multi-Signal Fusion for Token-Level Hallucination Detection
HuggingFace
2

Temporal Multi-Signal Fusion for Token-Level Hallucination Detection

Igor Itkin

NLP
LLMs Get Smarter from Targeted Synthetic Multilingual Data
HuggingFace
1

LLMs Get Smarter from Targeted Synthetic Multilingual Data

Ishika Agarwal, Arkajyoti Charaborty, Tanner Sorensen, Neha Gupta, Andreas Stolcke

Reinforcement Learning
Bounded Agents: Delegation Security for Multi-Agent AI Systems
HuggingFace
1

Bounded Agents: Delegation Security for Multi-Agent AI Systems

Xabier Muruaga

Machine Learning
SkillGate: Training In-Policy Skill Selection in Long-Horizon Agents
HuggingFace
1

SkillGate: Training In-Policy Skill Selection in Long-Horizon Agents

Qingyao Li, Wenxiang Jiao, Shuai Shao, Kangning Zhang, Yuan Lu

Shanghai Jiao Tong University

Computer Vision
SPK: Eliciting Structured Prior Knowledge for Interpretable Out-of-Distribution Detection in Real-Time Object Detection
HuggingFace
0

SPK: Eliciting Structured Prior Knowledge for Interpretable Out-of-Distribution Detection in Real-Time Object Detection

Changshun Wu, Weicheng He, Xiaowei Huang, Saddek Bensalem

University of Liverpool

AI Research Papers — August 20, 2026 Edition | Agentic Brew