Habr AI
AI news source. Articles are auto-selected and adapted by Hamidun News editors.
Latest publications
Sber, Yandex, and red_mad_robot showed how AI is changing the developer's role
At a meeting involving Sber, Yandex, T-Technologies, and red_mad_robot, participants discussed AI PDLC: AI is already writing most of the code, and the developer is shifting into the role of architect, editor, and agent
Nvidia, Google and Anthropic: Nemotron, Firefox bugs and worrying signals for the AI market
Nvidia released the open Nemotron-3-Super-120B, Anthropic found dozens of flaws in Firefox, and Google and ByteDance found themselves at the center of new disputes over the limits and risks of AI.
Claude Cowork and Chrome: 10 use cases for delegating repetitive browser work to an AI agent
Habr has published a practical breakdown of how Claude Cowork, paired with a Chrome extension, handles data lookup, form filling, price monitoring, QA, and other repetitive tasks in the browser.
Anthropic explained how businesses can implement agentic systems without unnecessary complexity
Anthropic showed that businesses are better off starting not with autonomous agentic systems, but with simple scenarios: they are cheaper, more reliable, and often more useful for marketplaces.
In five days, Yoyo grew from 200 lines into an autonomous agent that writes code on its own
An experiment with the Yoyo agent showed that 200 lines of Rust were enough to start a self-acceleration loop: the bot reads its own code, writes fixes, files tickets, and postpones the hard parts for later.
GLiNER 2 showed how compact encoders are catching up with LLMs in NER and classification
A new wave of zero-shot encoders led by GLiNER 2 shows that NER and classification tasks can be solved without heavy LLMs — faster, cheaper, and with a clear quality trade-off.
Ollama Cloud compared in a code review: DeepSeek v3.1 proved stronger than Qwen and GPT-OSS
The author tested three models in Ollama Cloud on real Python Pull Requests from a legacy project: DeepSeek v3.1 showed the best balance of depth, accuracy, and useful fixes.
CodeSpeak by Kotlin creator Andrey Breslav shifts development from code to specifications
Andrey Breslav's team released the public alpha of CodeSpeak — a platform where a developer describes a system in English and LLM generate, test, and refactor code, including in Python projects.
OpenAI, SoftBank and Oracle Argue Over Stargate as Mega Data Center Project Stalls
Stargate, a project worth up to $500 billion conceived as OpenAI’s computing backbone, is slowing as the partners argue over structure, funding, control and construction pace.
Claude as Part of Thinking: Why Losing an AI Tool Disrupts Not Just Access, but Work Itself
For active users, losing a Claude account turned out to be more than a service disruption: along with access, context, accumulated work, and part of the usual workflow logic disappear.
Seldon Vault turned Asimov's psychohistory into a multi-agent AI forecasting service
Seldon Vault turns Asimov's idea of psychohistory into a real AI service: seven agents, a skeptic, and an arbiter build daily forecasts of global events and assess their accuracy.
Anthropic and ETH Zurich: a long CLAUDE.md worsens agent performance and raises costs
A study by ETH Zurich across 138 repositories found that long CLAUDE.md and AGENTS.md files often reduce coding-agent success rates while also increasing token costs.
Gemini 3.1 Pro outperformed ChatGPT 5.4 and Claude Opus 4.6 in a text generation test
An author-led comparison of three top models across four literary tasks found that Gemini 3.1 Pro handled genre, emotion and brevity better, while ChatGPT and Claude trailed by only a small margin.
Anthropic, OpenAI, and Cursor: Eight Levels of Agent Engineering Maturity
Habr AI broke down eight levels of agent engineering — from tab-complete and context engineering to background agents and autonomous teams that are already changing software development.
AMD RX580 ran an LLM locally: how to tame ROCm, Ollama, and get GPU inference
An engineer showed how to make an old AMD RX580 run an LLM reliably with ROCm and Ollama, breaking down false signs of GPU activity, hipMemGetInfo failures, and Vulkan's role in diagnostics.
Why Claude 4.6 Isn't Enough Without Context: The Main Blind Spot in LLM Development
Even a strong model like Claude 4.6 loses effectiveness without systematically gathered context: a knowledge base, links between services, and environment rules matter more than a one-off prompt.
Developer builds AI-powered news aggregator with MCP, DeepSeek, and Telegram bot
After the drone attack over Dubai, a developer built a multilingual news aggregator: 80+ sources, an MCP server for AI agents, an AI chat, and a Telegram bot with quick digests.
BorisovAI tested MoE on an RTX 4090 and showed why perplexity breaks LLM evaluation
BorisovAI tested MoE with plug-in experts on a single RTX 4090 and found that an impressive perplexity score does not guarantee real quality gains on reasoning tasks.
LLM experiment showed how a model’s “personality” emerges in latent space
An experiment with a modular LLM showed that a separate latent vector can store not only text style, but also stable behavioral traits resembling the scaffold of a model’s personality.
Nvidia hints at an optical chip that could reshape AI data centers ahead of GTC 2026
Ahead of its March 16, 2026 presentation, Nvidia raised expectations with a promise of “a chip that will shake the world,” and silicon photonics is seen as the leading candidate for the surprise.
Why DeepMind's AGI advances do not answer the key question of machine consciousness
The author explains why growing computing power and DeepMind's AGI advances do not amount to the emergence of consciousness: intelligence can be imitated, while subjective experience remains tied only to living systems f
Habr AI on the future of work: how AI and robots could return society to a new antiquity
Habr AI argues that the combination of AI, robotics, and neural implants could do more than reshape the labor market — it could divide society into a narrow class of system operators and everyone else.
Unity showed how to build voice NPCs with memory and world context
A step-by-step guide shows how to build voice NPCs in Unity with a local model, dialog memory, knowledge of the game world, and spoken responses instead of standard scripted lines.
Harvard: AI is cutting junior hiring, and in three years that could affect the entire industry
Harvard reports a drop in junior hiring after AI adoption, while METR points to growing dependence of experienced developers on AI assistants, threatening talent shortages and tech debt.
DeepSeek and GLM-5 beat Yandex in a test of 34 AI models for managers without VPN
The authors of a large test of management scenarios found that DeepSeek V3.2 and GLM-5, available in Russia without VPN, are markedly stronger than Yandex's solutions and close to the global top tier.
Google released Gemini Embedding 2 for multimodal RAG with video, audio, and PDF
Google released Gemini Embedding 2, a model that embeds text, images, video, audio, and PDF into a single space and simplifies building multimodal RAG.
Bitrix24 listed eight common mistakes in developing MCP servers for LLMs
A Bitrix24 developer explained why MCP servers fail on authorization, call chains, poor tool descriptions, testing, security, and context overflow.
Why Yann LeCun's world model idea does not solve the main crisis in LLM development
After Yann LeCun's departure from Meta, his world model concept is once again being discussed as a path beyond LLMs, but critics argue that the physics of the world is too limited for true understanding.
A Physical AI pipeline for SO-101 was assembled on top of ROS2 and LeRobot for 30,000 rubles
An open-source stack based on ROS2 and LeRobot makes it possible to build a full Physical AI pipeline on the low-cost SO-101: teleoperation, episode recording, policy training, and deployment on hardware.
SimpleOne launched SimpleGen — an AI tool for development and deployment on the platform
SimpleOne introduced SimpleGen — an AI tool for generating solutions on its platform: developers only need to prepare a repository, access to an environment, and follow the workflow in Cursor.