2026
11,340 articles · 139 days · 10 categories · avg 82 per day

Qdrant and DRAG with KNEE: Making RAG Adaptive Without Wasting Tokens
DRAG with KNEE is an adaptive RAG approach built on Qdrant and Python. It builds a document hierarchy and trims excess context without a fixed top_k.

Why the Pentagon–Anthropic Standoff Is a Warning Sign for Businesses Using AI
The dispute between the US Department of Defense and Anthropic over usage restrictions exposed a new corporate risk: an AI strategy that depends on one vendor's…

How to Write Prompts for Midjourney, DALL-E and Kandinsky to Get Accurate Images
Why vague requests break image generation, and how prompt structure, word weights, negative instructions and settings make results more predictable.

Anthropic wins in court against the Pentagon, but the dispute is not fully settled

OpenClaw runs on a Wiren Board controller: an AI agent that manages hardware and writes scripts

Cursor launches TypeScript SDK for coding agents with cloud sandboxes and token-based billing

Who Is Behind Claude: How Former OpenAI Staff Built the Anthropic Team

ChatGPT helps design a controller for a 32-juice vending machine
A Habr author shared a ChatGPT conversation in which the model clarifies requirements and sketches a controller architecture for a juice vending machine, from NFC…

Pervaya Forma adds MCP so an AI agent can configure BPM processes
Russian BPM platform Pervaya Forma has implemented Model Context Protocol support: categories, forms and business processes can now be configured through a dialogue…

OpenAI, Google and Anthropic speed up the AI model race, but the market is already tired of the noise

Goldman Sachs speeds up AI agent adoption and reshapes how its developers work

OpenAI Shuts Down Sora Over Resource Shortages, Derailing Major Disney Deal

Goldman Sachs moves to a new phase of AI: Marco Argenti on agents

OpenAI hasn't turned ChatGPT into a new App Store yet, and developers are complaining
Six months after ChatGPT's mini-apps launched, OpenAI has hundreds of integrations, but developers cite weak functionality, poor visibility and clunky tools.

Glia wins AI Excellence Awards for safer AI in banking customer service
Glia won the 2026 AI Excellence Awards in the banking category. Judges praised its practical, safer use of AI in customer communications.

Biren Technology more than triples revenue as China's demand for AI chips surges
Biren Technology more than tripled its annual revenue as demand for AI chips in China jumps and companies build out local computing infrastructure.

Why Neural Networks Don't Predict the Moscow Exchange, and Where ML Actually Helps Traders

Habr AI: Self-organizing LLM agents outperformed hierarchical systems by 14%

Habr AI author proposes building deterministic AI on a MacBook Air instead of joining the GPU race

AI Agents Are Reshaping the Development Cycle: Where Scrum Shrinks and Where Humans Still Matter

Alibaba releases Qwen3.5-Omni, a native multimodal model for text, audio and video
Alibaba has unveiled Qwen3.5-Omni, a native omnimodal model that understands text, images, audio and video in a single architecture and can reply by voice in real…

Microsoft to invest more than $1 billion in Thailand's cloud, data centers and AI infrastructure
Microsoft has announced more than $1 billion in Thai investment for 2026–2028, covering data centers, cybersecurity, sovereign technology and AI skills training.

Equinix to invest 7.5 billion rand in new South African data centers as AI demand grows

OpenAI nears a possible IPO: an $850 billion valuation now demands profit, not just growth

Lansoft names the AI technologies that will reach production in 2026

AiConf 2026 changes its format: interaction, networking and hands-on work instead of lectures

Raft breaks down where MCP and Thin MCP add extra latency for AI agents
Raft compared five architectures for AI agents and found that the main source of delay is often not Python or HTTP, but the number of hops between services and the…

Anthropic accidentally leaks Claude Code source through npm source maps
.map files in an npm package exposed Claude Code's internal code, revealing a multi-agent mode, memory "dreams," hidden features and an undercover mode for…

DeepSeek V4 launches with focus on price and everyday tasks: review and first tests
DeepSeek V4 is pitched as an affordable alternative to top chatbots. Early reviews praise its price, solid performance in everyday scenarios and clear progress over…

OpenAI releases Privacy Filter, an open model for stripping personal data from text

FDA to speed up drug trials with AI and real-time data

Asia's AI rally is minting new winners deep in the chip supply chain

















