Latest Articles

• OpenAI's Big Move! Core API Now Supports MCP, Revolutionizing Agent Development Overnight
  Artificial IntelligenceAPI DevelopmentMachine LearningOpenAIAI Agents

  ...
• Understanding RAG, Agent, and Multimodality: Industry Practices and Future Trends
  AI TechnologiesRetrieval-Augmented Generation (RAG)Future TrendsIndustry ApplicationsMultimodal AIAI Agents

  ...
• Fooled by "AI for Science" Hype? A Scientist's Painful Lesson
  Artificial Intelligence in ScienceAI ApplicationsReporting BiasReproducibility CrisisScientific Research

  ...
• Breaking News! OpenAI Teams Up with Legendary Apple Designer Jony Ive to Announce New Company "io": Targeting Next-Gen AI Interaction Hardware
  AI HardwareAI InteractionOpenAISam AltmanJony Ive

  ...
• Does AI Know When to "Think"? Thinkless Teaches Large Language Models When to Reason
  Artificial IntelligenceLarge Language ModelsEfficiencyMachine LearningChain-of-Thought

  ...
• DreamCoder: Growing Generalizable, Interpretable Knowledge with Wake-Sleep Bayesian Program Learning
  Artificial IntelligenceProgram InductionNeural-Symbolic AIBayesian Program SynthesisWake-Sleep Learning

  ...
• GitHub Copilot's Agent Mode and MCP Support Officially Launched for JetBrains, Eclipse, and Xcode!
  Artificial IntelligenceGitHub CopilotDeveloper ToolsAgent ModeIDE Integration

  ...
• ICML 2025 | Training-Free, Instant Alignment of Large Model Preferences
  Large Language ModelsModel AlignmentMachine LearningNatural Language ProcessingInference Optimization

  ...
• Scholars explain how humans can hold the line against AI hype, and why it’s necessary
  Artificial IntelligenceAI EthicsRegulatory FrameworksLabor ImpactTechnology Criticism

  ...
• Google | Tracing RAG System Errors: Proposing a Selective Generation Framework to Boost RAG Accuracy by 10%
  Retrieval-Augmented Generation (RAG)Error AnalysisDeep LearningModel EvaluationNLP TechniquesAI ApplicationsGoogle ResearchAccuracy ImprovementLarge Language ModelsSelective GenerationSufficient ContextHallucination

  ...
• Google AI Nuclear Bomb: Upgrading All Models, Gemini 2.5 Tops Two Charts! All Products Rebuilt with AI, How Will OpenAI Respond?
  Artificial IntelligenceGoogle I/OExtended RealityAI ModelsGoogle SearchGemini

  ...
• Explosion! Google I/O Conference King Returns: Gemini "World Model" Emerges, Search Gets a "Brain Transplant", Create Original Movies with One Sentence
  Artificial IntelligenceGoogle I/OXR GlassesGoogle SearchGemini

  ...
• First Author Interpretation! Talking About Qwen's New Scaling Law—Parallel Scaling—From an Idea Perspective
  Language Model ScalingParallel ComputingMachine LearningQwenLLM InferenceLLM Training

  ...
• Multimodal Large Models Collectively Fail, GPT-4o Only 50% Safety Pass Rate: SIUO Reveals Cross-Modal Safety Blind Spots
  AI SafetyMultimodal AIBenchmarkingVision-Language ModelsLarge Language Models

  ...
• Interview with Step Ahead's Duan Nan: "We Might Be Touching the Upper Limit of Diffusion's Capability"
  AI Video GenerationFoundation ModelsEmbodied AIMultimodal AIDiffusion Models

  ...
• Nature Sub-journal: Humans Lost to AI Again, Especially When It Knows Who You Are
  Artificial IntelligencePersuasionMicrotargetingOnline DebateLarge Language Models

  ...
• GitHub Launches New Coding Agent
  AI Coding AgentGitHub CopilotMicrosoft BuildDevOpsSoftware Development

  ...
• Google CEO Pichai Responds to "Google Is Dead"
  GoogleAICompetitionTechnologySearch

  ...
• Which Model Should a Reliable Agent Use? The "Lost in Conversation" Phenomenon in LLM Multi-turn Dialogues | Microsoft Latest
  LLM ReliabilityMulti-turn DialogueLarge Language ModelsMicrosoft ResearchAI Agents

  ...
• Truth in Light vs. Algorithmic Illusion: Scientific Photography in the Age of AI
  Scientific PhotographyArtificial IntelligenceScience CommunicationImage ManipulationEthics

  ...