Latest developments in artificial intelligence, including large language models (LLMs), generative AI, multimodal AI, AI agents, robotics, and AI regulation and policy.
A practical guide to fine-tuning large language models using LoRA and QLoRA, covering theory, implementation steps, performance comparisons, and real-world pitfalls, with the latest trends from 2026.
Anthropic suspends Claude Fable/Mythos 5 worldwide after US export order over jailbreak flaws. Company calls it overreach. May spur AI supply chain diversification.
Amazon CEO Andy Jassy reportedly conveyed security concerns about Anthropic's AI models to the U.S. government, which then imposed export restrictions. Anthropic subsequently blocked global access to two models, highlighting the complex Amazon-Anthropic relationship.
The openJiuwen community has released the Physical AI architecture "Jiuwen Symbiosis" as open source. This article explains the design philosophy that brings real-world perception and action to AI agents.
Google's Gemini Omni allows simultaneous input of text, images, and video, enabling video generation and editing through natural language commands. Its output understands physics and culture, setting it apart from conventional AI video generation.
Ukraine confirmed it deployed 10 fully autonomous AI-controlled drones to the front lines in 2024, which killed humans. This is the first instance of operating in so-called "Terminator mode."
RAG (Retrieval-Augmented Generation) and fine-tuning are the two main approaches for customizing LLMs. This article compares their mechanisms, strengths, and limitations, and provides criteria for choosing the best approach in real-world deployments as of 2026.
China expands medical tourism with low-cost CAR-T therapy. Prices under half of US attract global cancer patients. Market to reach $3.4B by 2035.
OpenAI is considering significant token price cuts amid intensifying competition with Anthropic. Both companies have filed for IPOs, and the race for users is expected to heat up.
The image generation model HiDream-O1-Image-1.5, released by HiDream.ai, has secured second place globally on the Artificial Analysis Text to Image Leaderboard, trailing only OpenAI and surpassing models from Google and NVIDIA.
Visa integrates its payment network with OpenAI's ChatGPT, allowing AI agents to autonomously search for and complete purchases on behalf of users.
A management scholar presents a theoretical reconstruction against the argument that "AI makes organizations unnecessary," proposing a growth model based on uneven AI penetration across technical, managerial, and institutional layers and resource versatility.
The GGUF format is the standard file format for running local LLMs. This article provides a practical explanation of model selection criteria from the perspectives of quantization and compatibility.
Apple announced a Personal Context feature for Siri at WWDC 2026. The upgrade allows cross-app search across native apps like iMessage and Calendar, evolving Siri into a proactive personal AI assistant.
The latest 2026 guide to running local LLMs. Covers a comprehensive comparison of Ollama and llama.cpp, operating environments, model selection, and practical use cases.
Dr. David Sinclair of Harvard University is planning a human trial of an oral "reprogramming" drug, SL-100, within the framework of the XPrize competition. The attempt at whole-body rejuvenation sparks both hope and caution.
The latest UN "World Ocean Assessment" reveals the current state of ocean degradation: accelerating sea-level rise, expanding plastic pollution, and collapsing coral reefs. Analyzing environmental warnings from a technological viewpoint.
At WWDC 2026, Apple announced full-scale Apple Intelligence. iOS 27 makes Siri standalone app with Dynamic Island, reveals Google Gemini cooperation.
Moon's Dark Side launches Kimi Work. Up to 300 AI agents run in parallel on local PCs, automating file organization to deep research. Agent Swarm is moving toward practical use.
Anthropic begins offering its advanced vulnerability detection AI, "Claude Mythos," to select organizations, including the Australian government. Despite its potential to favor defenders, risks such as false positives and misuse are also highlighted.