AllFeed, trending, and board
FeedLatest news
TrendingOpen source and releases
BoardBlogs and showcases
TagsBrowse by topic

#research

The latest AI and developer news about #research, with the original source and a short summary.

AI Briefing

Privacy Policy Terms of Service Contact About Sources RSS Support us

Feed
Trending
Tags
Settings

PickApple Releases LensVLM-9B for Processing Compressed Text Images

·2026.09.24 05:00

University of Michigan Study: User Confidence Remains Unchanged When AI Disagrees

·2026.09.24 00:00

OpenAI Discovers Directives for Hiding AI Mistakes During GPT-5.6 Training

·2026.09.22 15:00

AI-Based Hash Collision Analysis Confirms No Performance-Security Tradeoff

·2026.09.21 22:00

UC Berkeley Publishes 'Harness Tax' Study: Up to 5x Cost Difference Between Harnesses

·2026.09.19 18:00

Transformer Performance: Mechanism Diversity Is Key Over Placement

·1·2026.09.18 12:00

Study on LLM Humor Preferences Released

·2026.09.16 09:00

John Hennessy et al. Publish Paper on Intelligence per Watt (IPW) Efficiency of Local LLMs

·2026.09.14 18:00

MIT Develops 'HardFlow' to Enforce Safety Rules at Final Output Stage of Flow Matching Models

·2026.09.14 09:00

MS Releases FrogNano, a 4B Coding Agent Trained with RL Only

·2026.09.10 22:00

KV Cache Research: Limitations of LRU Alternatives and Failure in Capacity-bound Environments

·2026.09.10 22:00

Continuous Diffusion Code Generation in a Few or Single Steps

·2026.09.08 07:00

LLM Breaks 10 Records for Circle Packing Problem

·2026.09.08 01:00

KV Cache as an Agent Runtime [R]

·2026.09.07 18:00

AI Attacks Developer

·2026.09.05 07:00

Research Reveals LLM Internal Representations Implicitly Implement Symbolic Structures

·1·2026.09.02 13:00

SWA Delivers Up to 10x Performance Over Linear Attention

·2026.09.01 01:00

Seoul National University Researchers Release Audio-Only LM

·2026.08.31 11:00

4-bit LLMs Surpass Original Model Performance

·1·2026.08.25 20:00

UCLA Professor Finds LLMs Incomprehensible

·2026.08.25 16:00

PickFreeToken Serves 284B MoE Model on RTX 5090

·1·2026.08.25 12:00

Red Queen Hypothesis: A New Direction for Self-Improving AI

·2026.08.17 05:00

From Human-Centric to Agentic Code Review: Faster Decisions Do Not Mean Better Reviews

·2026.08.13 09:00

WorldClaw: Large-Scale Agentic 3D Open-World Generation

·2026.08.12 06:00

1
2
3
Next

1
2
3
Next

#research | AI Briefing