In China, 95 percent of the 128,000 short dramas released in Q1 2026 were AI-generated. Some actors are being forced to hand over their voice and likeness…
Google’s WikiSkill gives AI agents a persistent memory of past mistakes to sharpen future performance
Google Research has introduced WikiSkill, a framework that gives AI agents a persistent knowledge base. Instead of discarding what they learned after each…
Learning New Facts with QLoRA: An Acquisition-Retention Frontier
arXiv:2608.25677v2 Announce Type: replace-cross Abstract: Parameter-efficient fine-tuning is often assumed to preserve pretrained capabilities because it…
When Stale Constraints Go Unchecked: Budgeted Verification Failures in Inherited Agent Memory
arXiv:2608.25553v2 Announce Type: replace-cross Abstract: An agent that inherits a consolidated memory may inherit a constraint that was true when written…
MACGen: Toward Functionally Correct and Secure Code Generation via Multi-Agent Collaboration
arXiv:2608.25457v2 Announce Type: replace-cross Abstract: Despite their strong ability to generate code, large language models often fail to produce…
4DStreamCtrl: Interactive Video Generation with Online 4D Control
arXiv:2608.25479v2 Announce Type: replace-cross Abstract: Generative video models now synthesize footage nearly indistinguishable from reality. Their…
Language Chain in Alignment: Cross-lingual Ranking Preference Optimization
arXiv:2608.23149v2 Announce Type: replace-cross Abstract: The alignment of Large Language Models heavily relies on English-centric high-quality preference…
CAT-GS: Balanced Multimodal Learning via Calibrated Gating and Fusion Surgery
arXiv:2608.24947v2 Announce Type: replace-cross Abstract: End-to-end training of multimodal neural networks often exhibits unstable neural dynamics…
LAION drops massive open video dataset with 10 million hours of footage for AI research
LAION’s Big Video Dataset (BVD) is one of the largest open video datasets for AI research, with 80 million videos, 10 million hours of runtime, and 55…
From State to Action: OODA-Tool for Reliable Multi-Turn Tool Use
arXiv:2608.24368v2 Announce Type: replace Abstract: Reliable multi-turn tool use requires an agent to preserve an evolving task state and ensure that each…
Barret Zoph, the Thinking Machines co-founder ousted before joining OpenAI, is now at Google
Zoph, who co-founded Thinking Machines Lab alongside Mira Murati and also served as the startup’s CTO, led a brief stint at OpenAI and is now at Google.
DataKernelBench: Can LLMs Optimize Database Queries on GPUs?
arXiv:2608.25061v2 Announce Type: replace-cross Abstract: GPUs increasingly accelerate database systems, but query-specific peak performance still often…
Introducing OpenAI models on Amazon Bedrock for in-country inferencing in India
Amazon Bedrock now supports the OpenAI GPT-5.6 models, Terra and Luna, in India with India geographic cross-Region inference. If you have local data…
Unsupervised Post-Training of Foundation Models: A Survey
arXiv:2608.24982v2 Announce Type: replace-cross Abstract: Foundation-model post-training usually relies on human labels, preference data, stronger…
MLLM-Routed Heterogeneous Ensembles for Robust Cross-Dataset Image Classification
arXiv:2608.13463v2 Announce Type: replace-cross Abstract: Modern image classification models excel when trained on single task-specific datasets but often…
X$^2$Localizer: Cross-grained Alignment for Progressive Cross-view Video Geo-localization
arXiv:2608.16658v2 Announce Type: replace-cross Abstract: Cross-view Video Geo-localization (CVG) aims to localize ground-view videos by retrieving their…
Pre-training Visual Dexterity in Simulation
arXiv:2608.15917v2 Announce Type: replace-cross Abstract: Large-scale pre-training has made robot policy fine-tuning increasingly data-efficient, but this…
Mol-JEPA: A multimodal Joint Embedding Predictive Architecture for Molecules
arXiv:2608.22642v2 Announce Type: replace-cross Abstract: Despite recent advances in molecular foundation models, several limitations remain, such as…