Latest Articles

• Advancing Silicon-Based Intelligence: Shuchao Bi's Insights on Past, Present, and Future AI
  Artificial IntelligenceMachine LearningAGIScaling LawsReinforcement Learning

  ...
• The "Mirage" of Chain-of-Thought Reasoning: An In-depth Look at LLM Generalization
  Chain-of-Thought ReasoningLarge Language ModelsAI ResearchOut-of-DistributionGeneralization

  ...
• GPT-5 vs Claude Opus 4.1: Coding Capability Assessment
  AI ModelsCodingDevelopment ToolsLarge Language ModelsPerformance Comparison

  ...
• OpenAI Board Chair: "Per-Token Billing" Is Completely Wrong, Market Will Eventually Choose "Outcome-Based Pricing"
  AI Business StrategyBret TaylorStartup EcosystemPricing ModelsAI AgentsOpenAI

  ...
• In-depth Dissection of Large Models: From DeepSeek-V3 to Kimi K2, Understanding Mainstream LLM Architectures
  Large Language ModelsDeep LearningMixture of ExpertsAttention MechanismsAI Architectures

  ...
• Xiaohongshu Open-Sources First Multimodal Large Model, dots.vlm1, Performance Rivals SOTA!
  Multimodal AIVisual Language ModelsAI ResearchDeep LearningOpen Source AI

  ...
• Altman Reveals Stunning Prediction: GPT-8 to Cure Cancer by 2035! Humanity Might Wage WWIII Over Compute Power
  Artificial IntelligenceOpenAIHealthcare AIAI EthicsFuture TechnologySam Altman

  ...
• ARPO: Agentic Reinforced Policy Optimization, Enabling Agents to Explore One Step Further at Critical Moments
  Reinforcement LearningLarge Language ModelsTool UsePolicy OptimizationAI Agents

  ...
• Open-Sourcing the Largest High-Quality Scientific Reasoning Post-Training Dataset to Quickly Turn Qwen3 and Others into "Scientists"
  Artificial IntelligenceScientific ReasoningOpen SourceLarge Language ModelsDataset

  ...
• Wang Mengdi's Team Review of "Self-Evolving Agents": From Static LLMs to Artificial Superintelligence (ASI)
  Self-Evolving AI AgentsLarge Language ModelsArtificial SuperintelligenceFuture AI ResearchAI Applications

  ...
• Anthropic Team Uncovers 'Persona Variables' to Control Large Language Model Behavior, Cracking the Black Box of AI Madness
  Large Language ModelsAI SafetyBehavior ControlFine-tuningPersona Vectors

  ...
• Google Open-Sources DeepPolisher, Halving Genome Assembly Error Rates; Jeff Dean: "Exciting!"
  GenomicsDeep LearningHuman Genome ProjectArtificial IntelligenceBioinformaticsGenome Assembly

  ...
• AI's New SOTA in Bug Fixing: ExpeRepair Achieves 60.33% Fix Rate on SWE-Bench Lite, Learns from Experience Like Humans – Developed by Institute of Software, Chinese Academy of Sciences
  Artificial IntelligenceSoftware DevelopmentAutomated RepairMachine LearningBug Fixing

  ...
• Oxford Anthropologist Anna Machin: Dating Apps Are Making Your Brain's "Mate Selection Algorithm" Fail
  RelationshipsDating AppsNeuroscienceHuman EvolutionPsychology

  ...
• Is Your Model's Attention Drifting? RUC and Tsinghua University Introduce LeaF: Pruning Distracting Tokens for Focused Learning
  Large Language ModelsKnowledge DistillationModel OptimizationCausal InferenceAttention Mechanism

  ...
• Can Models Truly "Reflect on Code"? Beihang University Releases Repository-Level Understanding and Generation Benchmark, Refreshing the LLM Understanding Evaluation Paradigm
  Large Language ModelsCode ReflectionCode GenerationCode UnderstandingBenchmarking

  ...
• ReaGAN: Empowering Each Node as an Intelligent Reasoning Expert in Graphs
  Artificial IntelligenceGraph Neural NetworksMachine LearningAgent-based AILarge Language Models

  ...
• Google's Challenge: DeepSeek, Kimi and More to Compete in First Large Model Showdown Starting Tomorrow
  AI BenchmarkingLarge Language ModelsModel EvaluationKaggle Game ArenaAI Chess

  ...
• RAG Revolution! Graph-R1, the First RL-driven Graph Reasoning Agent
  GraphRAGReinforcement LearningAI AgentKnowledge GraphLarge Language Models

  ...
• Alibaba Just Open-Sourced Qwen-Image: Free GPT-4o Ghibli-Style Model, Best in Chinese
  Generative AIText-to-ImageMultimodal ModelsOpen-Source AIImage Generation

  ...