Habr AI
AI news source. Articles are auto-selected and adapted by Hamidun News editors.
Latest publications
OpenAI, Google, and Anthropic Accelerate AI Model Race, but the Market Is Already Tired of the Noise
In February 2026, leading AI labs released dozens of models, but the real story isn't record-breaking benchmarks—it's falling prices, growing open source, and weak ROI for businesses.
Resilience: AI Accelerated Cyberattacks, Response Window for Defenders Narrowed to Minutes
Reports from Resilience, IBM, and ReliaQuest show: generative AI hasn't invented new attack types, but has dramatically accelerated phishing, breaches, and fraud, narrowing defense response time to minutes.
Why Neural Networks Fail to Predict Moscow Exchange Markets—and Where ML Actually Helps Traders
Analysis of ML in trading reveals why neural networks struggle with direct price forecasting, yet help identify market regimes, filter trades, and calculate transaction costs.
Habr AI: Self-Organizing LLM Agents Outperformed Hierarchical Systems by 14%
An experiment on 25,000 tasks showed that LLM agents perform better without imposed roles: self-organization, specialization choice, and the right to opt out improve quality.
Habr AI Author Proposes Deterministic AI on MacBook Air Over GPU Race
A GPU race manifesto has appeared on Habr: the author argues that a deterministic AI kernel in Rust with O(1) logic can run locally on MacBook Air and reduce hallucinations.
AI agents reshape the development cycle: where Scrum contracts, where humans remain essential
AI agents do accelerate development, but unevenly: in new projects, control shifts to observability; in legacy code, validation becomes more critical; and regulators don't forgive mistakes.
FSBio Describes Metabolic AI Runtime — AI Architecture with "Homeostasis" Instead of Empathic Prompts
FSBio Claims That True Machine Empathy Is Impossible in Standard Transformer Models, Proposing Metabolic AI Runtime with Internal State, Memory, and "Will"
Lansoft names AI technologies that will reach production in 2026
In Lansoft's analysis, optimized transformers are named as the main candidates for production in 2026, while quantum ML and mass-market BCI remain outside practice for now.
AiConf 2026 shifts format: interactive sessions, networking, and hands-on work replace lectures
AiConf 2026 moves away from traditional AI conference format, pivoting toward interactive sessions, collaborative work, and networking so participants leave with solutions, not notes.
How ecom.tech sees the union of AI, analytics, and 1C: what already works and where the risks lie
ecom.tech demonstrated how AI is integrating into the 1C ecosystem: from semantic search and code review to video support bots, but without an analyst's oversight, such tools create a dangerous illusion of accuracy.
Raft Analyzed Where MCP and Thin MCP Add Latency to AI Agents
Raft compared five architecture options for AI agents and showed that the main cause of delays is often not Python or HTTP, but the number of inter-service transitions and the interaction model.
Anthropic Accidentally Leaked Claude Code Source Code via npm Source Map
Internal Claude Code source code was exposed through .map files in an npm package, revealing a multi-agent mode, memory 'dreams' system, hidden features, and a cover mode for open-source work.
DeepSeek V4 Launches With Focus on Price and Everyday Tasks: Review and First Tests
DeepSeek V4 is positioned as an affordable alternative to leading chatbots: early reviews praise its competitive pricing, reliable performance in everyday scenarios, and notable model improvements.
EVRAZ Implements Neural Network for Steel Heating and Gas Consumption Reduction in Rolling Mill
EVRAZ deployed a recommendation system with digital twins of furnaces in NTMK hot rolling: it suggests heating modes for blanks and helps reduce gas consumption.
Turnitin and OpenAI Are Losing: Why AI Detectors Can No Longer Distinguish Humans From Models
Language models have learned to mimic human style so well that AI detectors work almost like a coin flip: they make mistakes, catch the innocent, and increasingly fail to protect education and the online environment.
OpenAI and LinkedIn: Why Prompt Writing Has Become Critical for Career
OpenAI, Microsoft, and LinkedIn reveal one trend: the ability to write precise prompts is becoming a work skill, and vague requests now harm both answer quality and hiring prospects.
Sber Develops GigaChain Without Forking LangChain, Betting on a Compatibility Package
Sber explained how it develops GigaChain around GigaChat: the team rejected a heavy LangChain fork, built a compatibility package, and is creating an open AI agent ecosystem.
Timeweb Cloud explains how edge detection filters work: from Roberts to Canny detector
Timeweb Cloud released a detailed guide on edge detection in images: it covers derivatives, Roberts, Prewitt, Sobel, Laplacian, Canny operators, and the Wallace filter.
Peter Publishing House released a book on conversational AI and chatbots that work
Peter Publishing House presented a book on effective conversational AI: it focuses on LLM-based chatbot creation, dialogue design, and real user experience.
SpeShu.AI explained how to find work in AI amid growing job vacancies in Russia
SpeShu.AI highlighted a market paradox: the number of AI-skill vacancies in Russia reached 16.5k per quarter, but high competition and hundreds of applications complicate hiring.
Anthropic's Opus 4.7 consumes more tokens: hidden inflation at unchanged prices
In Anthropic's Opus 4.7, the same text may require significantly more tokens than in Opus 4.6 — with prices unchanged, this is a subtle but real increase in API costs.
Alfa-Bank Showed How Skeptics Deployed an Internal HR Service to Production with GLM-5
Alfa-Bank described an experiment with GLM-5: a team of team lead, analyst, and product manager shipped an HR service to production in just a few weeks, but the main bottleneck turned out not to be AI, but Git.
How a YADRO Developer Configures Claude Code with Spec-Driven Development
A YADRO architect has been managing AI agents using SDD methodology for nine months and explains how to configure Claude Code to switch models between projects without reloading.
How Just AI Helped a Bank Break Through Automation Ceiling with LLM-Agents
Hundreds of NLU-scenario branches — yet the automation rate still stalled. Just AI migrated the bank's support to LLM-agents and added a judge-agent against hallucinations.
10 RAG approaches that actually work in production: from basic to GraphRAG
A practical breakdown of 10 RAG techniques — from naive RAG and hybrid search to GraphRAG and Self-RAG — with guidance on when to apply each in real-world projects.
SENSE: why the LLM market is shifting models from service to paid access to thinking
The SENSE author proposes viewing LLMs not as another SaaS offering, but as infrastructure: tokens become the unit of access to thinking, price and security emerge as the primary competitive arena.
Sber: Yago knowledge graph barely helped search, while LightRAG added 12 percentage points of accuracy
Sber team compared two approaches to improving search: the ready-made Yago graph barely changed results, while LightRAG built on their own documents delivered 12 percentage points of accuracy gains.
Hugging Face and ModelAudit: research revealed the limits of built-in ML model protection
An experiment on Hugging Face models showed that ModelAudit detects additional dangerous signals, but often generates noise, and the accuracy of findings strongly depends on rule versions and manual triage.
Local Vision for z.ai GLM-5.1: 8B Model Closes 70% of the Gap to the Frontier
An author demonstrated how to add local vision to a low-cost coding agent based on z.ai GLM-5.1 via an MCP server and qwen3-vl:8b, enabling it to verify the interface from screenshots.
Federal Treasury Launches AI Monitoring of National Project Construction Sites Using QMonitoring
Federal Treasury has integrated QMonitoring for 24/7 monitoring of national project construction sites: AI-powered video analysis detects risks, failures, and schedule delays.