AI Article Tags

Browse articles by professionally maintained tags. Every tag opens a focused collection with its own editorial introduction and closely related published articles.

Doubao Doubao spans everyday chat, search, multimodal creation, desktop assistance, and system-level access, so product updates directly reshape practical work and daily-use scenarios. Apple M6 Starting from local compute, see how Apple M6 and M5 Ultra rewrite on-device large-model work through unified memory and bandwidth. Runway Put generation inside the edit timeline, then see how Runway's plug-in, Solaris and Gen-4 consistency fit post-production. Claude Opus Tracing the Opus flagship line: from 4.5's coding and agent focus, through 4.6's million-token context, to 4.8's long-horizon agents. Gemini 3 Pro Search AI Mode, the Deep Research agent and Antigravity all run on the same engine: Gemini 3 Pro. DeepSeek V4.1 Flash With a 1M context and 384K of output, DeepSeek V4.1 Flash now absorbs traffic once handled by V4 Pro. iPhone 18 Pro Analyze iPhone 18 Pro through the A20 Pro, on-device AI, camera system, content-authenticity tools, and professional workflows. Apple Watch Series 12 Understand Apple Watch Series 12 through health sensing, workout recovery, audio intelligence, privacy controls, and battery performance. iPhone Duo Examine iPhone Duo through its foldable construction, dual-screen interaction, on-device AI, multitasking, durability, and app adaptation. iPhone 18 Follow the standard iPhone 18 model through official announcements, feature updates, hands-on findings, generation comparisons, and accessory support. Transformer Architecture Why the Transformer became the default backbone of large models, what attention actually costs, and what newer routes like diffusion language models and single-stream designs try to change. Tongyi Qianwen Track the open-source progress of Alibaba's Tongyi Qianwen series and examine trade-offs between language, multimodal and programming models in parameter scale, context length, and deployment cost. Sora Follow OpenAI's Sora video generation model and product iterations to understand the real progress of text-to-video in physical consistency, controllability, and social sharing. RAG Explain how retrieval-augmented generation works, and the roles Embedding, vector databases, and resorting play in enterprise knowledge base Q&A. Perplexity Track Perplexity's evolution as an AI answer engine and see how verifiable sources, browsing safety, and deep search are changing search and research methods. OCR Text Recognition Characters on a scanned page have to be read correctly before any model can reason about them. Layout analysis, multilingual recognition, structured PDF output, and where PaddleOCR and HunyuanOCR part ways are the core concerns. NotebookLM Track the evolution of Google NotebookLM's features, see how notes, data organization, and AI-generated content are combined, and understand the differences in capabilities brought by subscription levels. Model Compression and Quantization The routes to smaller, faster models: how distillation moves capability into a smaller network, what low-precision quantization actually buys, and where small language models stop being enough. Midjourney Track Midjourney's version updates and image creation practices to see what advantages it still retains in style control, community ecosystem, and cover image scenarios. MCP Explain what the MCP model context protocol is, how APIs differ from plugin calls, and why it has become the universal interface for the Agent platform. Kimi Track the progress of Kimi's models and products, and observe the trade-offs between the K2 and K3 series regarding long context, open-source weighting, and programmable agents. Humanoid Robots Walking on two legs sounds like a stunt, but the reason is practical: doors, stairs and tool handles were all built around human proportions. Degrees of freedom, payload and realistic timelines differ machine by machine. Grok Follow xAI's Grok model and products to see how real-time retrieval, multimodal reasoning, and chatbots come together, and how it differs from ChatGPT. GPT-6 Track the progress and direction of OpenAI's next-generation flagship model GPT-6, and see how long-term memory, personalization, and continuous self-learning will change the boundaries of large model capabilities.