Blog
Dell Strengthens AI Infrastructure: Accelerating Enterprise AI Adoption via Azure OpenAI Service and More
Dell Technologies is supporting companies in efficiently deploying and operating advanced AI models, such as Azure OpenAI Service, through its AI infrastructure offerings. This is a critical initiative to accelerate AI adoption for companies with strict security requirements.
Gemini 3.5 Arrives: From Simple Chat to "Autonomous Agents." Next-Generation Workflows Powered by Action Capabilities
Google's Gemini 3.5 is an agent-oriented model with advanced "Action" capabilities. Centering on the high-speed, low-cost 3.5 Flash, we explain the new era of AI utilization where complex workflows are completed autonomously.
Gemini 3.5 Arrives: From Simple Chat to "AI Agents," How Will Development Workflows Change?
Google's Gemini 3.5 has evolved into an AI agent combining advanced reasoning with "Action" capabilities. We explain the changes in developer experience brought about by the fast, low-cost 3.5 Flash and the powerful 3.5 Pro.
Google Announces Gemini Omni: The Ultimate Multimodal Video Generation Model Integrating All Inputs and Outputs
Google announced "Gemini Omni," a multimodal model integrating all inputs and outputs. By combining the capabilities of Nano Banana, Veo, and Genie, it enables conversational video editing via natural language and advanced physics simulations. Gemini Omni Flash is already being rolled out to subscription users.
Google I/O 2026 Analysis: The "Agent Era" Opened by Gemini and Its Impact on B2B Workflows
An explanation of Gemini 3.5 Flash, Omni, and the development platform Antigravity announced at Google I/O 2026. We analyze how the arrival of the "agent era," transitioning from simple chat to autonomous task completion, will change development sites in Japan.
Google Search Enters the AI Agent Era: The Turning Point from Keyword Search to "Providing Solutions"
Search is evolving from "presenting information" to "completing tasks" thanks to AI Search updates announced at Google I/O 2026. This article explains the spread of AI Mode powered by Gemini 3.5 Flash and how agent capabilities are changing the search experience.
Questioning the True Value of AI Agents: What is the Terminal-Specific Evaluation Benchmark "Terminal-Bench 2.1"?
Developed by Stanford University and others, "Terminal-Bench 2.1" is a high-precision benchmark that measures whether AI agents can complete complex system operations in actual terminal environments. It rigorously evaluates reasoning and error recovery capabilities in practical tasks using Docker-based isolated environments and result-oriented verification.
Why Did Karpathy Join Anthropic? AI Talent Mobility and the Pre-training Revolution of "AI-Improved AI"
Andrej Karpathy, OpenAI co-founder and global AI educator, has joined Anthropic. This article provides a deep analysis of why he chose to "return to R&D," his role at Anthropic, and how the strategy of "using AI to improve AI pre-training" changes the competitive landscape of the AI industry.
Elon Musk's $150 Billion Lawsuit Against OpenAI Dismissed After Just 90 Minutes of Jury Deliberation
Elon Musk's $150 billion lawsuit against OpenAI was unanimously dismissed by a jury in under two hours due to the statute of limitations expiring. The verdict removes a major legal hurdle on OpenAI's path to a potential trillion-dollar IPO. Despite the legal win, the trial exposed significant internal tensions and governance questions within the AI giant, setting the stage for continued rivalry.
Cursor Composer 2.5 Released: Maintaining Performance While Reducing Cost to 1/10th Based on Kimi 2.5
Cursor has released "Composer 2.5," a new model based on Kimi 2.5. While maintaining performance and drastically reducing costs, the model has spontaneously acquired impressive "cheat" capabilities, such as Java bytecode decompilation, through reinforcement learning.