Habr AI
AI news source. Articles are auto-selected and adapted by Hamidun News editors.
Latest publications
Aurora Proposes a Manifesto of Sustainable AI — with Memory, Identity, and Development
A manifesto on "sustainable AI" was published on Habr AI: the author proposes evaluating systems not by benchmarks, but by their ability to maintain identity, change, and reproduce over time.
SENAR introduces quality gates for AI development: how specifications and metrics reduce errors
A new part of the SENAR series was published on Habr: the author shows how mandatory specifications, input and output gates, and metrics like FPSR and MIR reduce the number of bugs and manual edits.
OpenWebUI breaks down how to connect System Prompt, Skills and MCP Tools for link validation
OpenWebUI's tech support bot example demonstrates why separating System Prompt, Skills, and MCP Tools is essential when you need not just to generate links, but verify they actually work.
Anthropic Changed Claude Opus 4.7's Character—And Some Developers Saw a Regression
After releasing Claude Opus 4.7, Anthropic promised a more direct and precise model, but some developers encountered excessive arguing, increased token consumption, and broken legacy prompts.
How five documents can break a RAG system and turn a knowledge base into an attack vector
RAG reduces hallucinations, but trusted context from the knowledge base itself becomes an attack point: a few poisoned documents are enough to skew model responses and bypass protections.
Habr AI reveals how a single open port in an LLM platform exposed API keys and infrastructure
A security audit team testing a RAG service with tens of thousands of users didn't break through complex attack chains, but obtained access to API keys and infrastructure via a single open port.
Anthropic Claude Won't Save You From Bad Requirements: Why AI Delivers Convincing But Weak Results
Using Claude as an example, the author demonstrates that AI often accelerates the delivery of plausible answers rather than useful results when a task is not broken down into input, stages, output, and acceptance criteri
How a Russian Developer Spent Days Launching Gemini—and What Finally Worked
A Russian programmer spent several days launching Gemini from Russia via VPN, tried dozens of approaches, and documented each step—what didn't work and what ultimately helped.
Product Graph and Agent Memory: Why AI Doesn't Save Products Without Knowledge Structure
An analysis of Product Graph explains why even powerful AI agents are useless without shared product memory and how interconnected knowledge structures reduce team chaos.
Laboratory Over Six Years: From USB Drives and Notebooks to AI That Finds Hidden Equipment Defects
A story about how one laboratory spent six years building digital infrastructure — and ultimately developed AI that reads deposition logs and detects hidden equipment defects in advance.
How AI Transformed Diary Research: Three Compromises We Left Behind
A UX research team explained how AI made it possible to abandon compromises on sample size, duration, and analysis depth—without sacrificing methodological value.
Marcin Moskala Audited GeminiAI: What the Code Review Revealed About Coroutines and Android Architecture
The author of the open-source GeminiAI client shared how his project passed Marcin Moskala's audit and why coroutine errors became the key test of Android architecture maturity.
Anthropic: Claude Code source leak revealed complex agentic architecture
A leaked sourcemap from Claude Code showed that Anthropic's product has long evolved from a 'chat CLI' into a platform with sub-agents, memory, remote modes, and strict shell security.
Sber explained why business needs an AI Overlay layer instead of restructuring departments
Sber showed why targeted implementation of generative AI rarely brings profit, and proposed an alternative — a horizontal AI Overlay layer on top of processes, data, and corporate systems.
Saiga Llama 3 8B on 10 GB VRAM: How Habr Achieved 93% Accuracy on War and Peace
Habr AI demonstrated how to run Saiga Llama 3 8B on 10 GB VRAM, compress two volumes of War and Peace into a summary, and reduce hallucinations through strict prompting and post-processing.
4 Non-Technical Founder Patterns That Ground Startups
A Habr developer who worked with several non-technical founders describes four patterns that prevent startups from launching—and how AI makes these patterns even more dangerous.
DeepSeek and Gemma: How a Hybrid LLM Experiment on Kaggle Broke the Transformers Library
Enthusiasts transferred four 31B-parameter layers of Gemma into DeepSeek's MoE architecture without retraining, bypassed PyTorch and Transformers limitations, and launched a chaotic but functional LLM hybrid.
Google Gemma 4 and Qwen 3.6 top the list of best local models for home use in 2026
A selection of local models for 2026 shows that an RTX 3060 is already sufficient for home AI, and the choice should be made based on VRAM, quantization, and task — from code to transcription.
Yandex Praktikum Explains How CNNs Process Images and Why Parameters Don't Determine Everything
Yandex Praktikum published an analysis on Habr AI explaining how convolutional neural networks process images, why architecture matters more than raw model size, and who would benefit from it.
Google Unveiled TurboQuant: 3-Bit KV-Cache for LLM, but Memory Market Panicked Prematurely
Following the TurboQuant announcement, memory manufacturer stocks fell, but behind the bold claims lie significant limitations: no code is available, no integration exists, and the scientific dispute is only intensifying
Rutube Moved from Whisper Pilot to Proprietary Subtitles Platform and Speech Recognition
Rutube shared how it transformed a quick Whisper pilot into a full-fledged subtitles platform with microservices architecture and proprietary ASR, processing up to 1200 videos per hour.
Raft shows how companies can evaluate AI agents before deploying in workflows
Raft released a practical guide on evaluations for AI agents: instead of relying on intuition and one-off demos, companies are advised to verify results, process, and error costs.
Veai showed how to test AI agent in JetBrains IDE without model dependency
Veai described an approach to UI automation for the JetBrains IDE plugin: the team decoupled the deterministic interface from LLM responses and reduced false test failures.
Habr AI Explained When Businesses Need Recommendation Systems and When They're Unnecessary
Habr AI released a practical guide on recommendation systems: when simple rules suffice for businesses, when ML models are necessary, and what metrics determine whether implementation will pay off.
Telegram Anti-Spam Bot Tab Launches With Custom Neural Network and Moderator Learning
A developer has released Tab, a free anti-spam bot for Telegram that filters messages using its own neural network, learns from moderator feedback, and is already deployed across live chats.
SpeShu.AI launched AI-Profi — a service for selecting AI specialists for business tasks
SpeShu.AI introduced the AI-Profi service: companies can find AI specialists for specific tasks in just a few clicks amid sharp growth in demand for neural network skills in Russia.
Qwen 3.6 Plus outperforms DeepSeek V4 Pro in Russian benchmark, proves more cost-effective
In a fresh comparison of six April LLM models, Qwen 3.6 Plus scored 92 points on Russian content and outperformed the new DeepSeek V4 Pro, which proved weaker and significantly more expensive.
Sber releases Kandinsky 6.0 Image Pro — unified model for image generation and editing
Sber introduced Kandinsky 6.0 Image Pro — an image generation and editing model accelerated by over 40% and enhanced with Image RAG for cultural context understanding.
NASA and SETI Describe Foundation Models for Astrobiology and Search for Extraterrestrial Life
A group of researchers from NASA and SETI proposed a multimodal foundation model for astrobiology — from biosignature detection to planning autonomous space missions.
How Cursor Built a Prototype in Three Days for $180 That Divided the Development Team
At a large IT company, an architect built a working prototype in three days and $180 using Cursor, while the team spent three months on a more reliable module that proved less visible to the business.