2026

7,718 articles · 96 days · 10 categories · avg 80 per day

DEMIURGOS adds a rules debugging mode for AI agents and support for 20+ tools
LLM·Habr AI

DEMIURGOS adds a rules debugging mode for AI agents and support for 20+ tools

Apr 30, 2026·3 min
Whisper and Gemma 3 linked with contrastive learning for low-cost speech input to LLMs
LLM

Whisper and Gemma 3 linked with contrastive learning for low-cost speech input to LLMs

Apr 30, 2026·3 min
Stanford introduced OpenJarvis — a local AI agent stack with memory and learning
LLM

Stanford introduced OpenJarvis — a local AI agent stack with memory and learning

Apr 30, 2026·3 min
An open-source stack of 6 models and 9 agents showed how to build an AI team on a single server
LLM

An open-source stack of 6 models and 9 agents showed how to build an AI team on a single server

Apr 30, 2026·3 min
Singapore has invested more than S$1 billion in AI, but AI Singapore sees a shortage of people who can build it
LLM·Bloomberg Tech

Singapore has invested more than S$1 billion in AI, but AI Singapore sees a shortage of people who can build it

Apr 30, 2026·3 min
Amazon Bedrock AgentCore gets Policy for AI agent access control
LLM

Amazon Bedrock AgentCore gets Policy for AI agent access control

Apr 30, 2026·2 min
AWS explained how to fine-tune Amazon Nova through an LLM judge for complex enterprise tasks
LLM

AWS explained how to fine-tune Amazon Nova through an LLM judge for complex enterprise tasks

Apr 30, 2026·3 min
Legora reaches $5.6B valuation and intensifies rivalry with Harvey in AI for lawyers
LLM·TechCrunch

Legora reaches $5.6B valuation and intensifies rivalry with Harvey in AI for lawyers

Swedish startup Legora raised another $50 million, crossed $100 million ARR and closed in on Harvey, turning the AI-for-lawyers market into

Apr 30, 2026·3 min
Top themes
01Alphabet and Amazon show returns from AI while Meta lags in the investment raceApr 30, 2026·2 min02Gamma, Tome and SlidesAI: Which AI Services Actually Speed Up Presentation CreationApr 30, 2026·3 min03E.SUN Bank and IBM create an AI governance system for safe AI deployment at the bankApr 30, 2026·2 min04Sber, Yandex, and red_mad_robot showed how AI is changing the developer's roleApr 30, 2026·3 min05Nvidia, Google and Anthropic: Nemotron, Firefox bugs and worrying signals for the AI marketApr 30, 2026·3 min
Microsoft launches Copilot Health — an AI service for unifying medical data and health analysis
LLM·TNW

Microsoft launches Copilot Health — an AI service for unifying medical data and health analysis

Microsoft has opened the waitlist for Copilot Health, a protected section in Copilot where AI compiles data from wearables, medical records,

Apr 30, 2026·3 min
01Amazon to pair giant Cerebras chips with Trainium to run AI modelsApr 30, 2026·3 min02Elon Musk vows to rebuild xAI after a new wave of co-founder departuresApr 30, 2026·3 min03Claude Cowork and Chrome: 10 use cases for delegating repetitive browser work to an AI agentApr 30, 2026·2 min
Microsoft and NVIDIA call physical AI the next advantage for industry
LLM·MIT Technology Review

Microsoft and NVIDIA call physical AI the next advantage for industry

MIT Technology Review describes a new stage of automation: physical AI helps factories not just cut costs, but grow amid labor shortages and

Apr 30, 2026·3 min
Anthropic explained how businesses can implement agentic systems without unnecessary complexity
LLM·Habr AI

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, m

Apr 30, 2026·3 min
In five days, Yoyo grew from 200 lines into an autonomous agent that writes code on its own
LLM·Habr AI

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

Apr 30, 2026·2 min
ChatGPT and data centers: how AI growth is accelerating energy use and pressure on water
LLM·Guardian

ChatGPT and data centers: how AI growth is accelerating energy use and pressure on water

Apr 30, 2026·3 min
GLiNER 2 showed how compact encoders are catching up with LLMs in NER and classification
LLM·Habr AI

GLiNER 2 showed how compact encoders are catching up with LLMs in NER and classification

Apr 30, 2026·3 min
NVIDIA introduces NeMo Retriever — agentic search for complex enterprise data
LLM·Hugging Face Blog

NVIDIA introduces NeMo Retriever — agentic search for complex enterprise data

Apr 30, 2026·3 min
Ollama Cloud compared in a code review: DeepSeek v3.1 proved stronger than Qwen and GPT-OSS
LLM·Habr AI

Ollama Cloud compared in a code review: DeepSeek v3.1 proved stronger than Qwen and GPT-OSS

Apr 30, 2026·3 min
CodeSpeak by Kotlin creator Andrey Breslav shifts development from code to specifications
LLM·Habr AI

CodeSpeak by Kotlin creator Andrey Breslav shifts development from code to specifications

Apr 30, 2026·2 min
OpenAI, SoftBank and Oracle Argue Over Stargate as Mega Data Center Project Stalls
LLM

OpenAI, SoftBank and Oracle Argue Over Stargate as Mega Data Center Project Stalls

Apr 30, 2026·3 min
AWS and vLLM integrated P-EAGLE to speed up large LLM inference by up to 1.69x
LLM

AWS and vLLM integrated P-EAGLE to speed up large LLM inference by up to 1.69x

Apr 30, 2026·3 min
Elon Musk reshapes xAI: layoffs, consultants from Tesla and SpaceX, and a bet on AI data centers
LLM

Elon Musk reshapes xAI: layoffs, consultants from Tesla and SpaceX, and a bet on AI data centers

Apr 30, 2026·3 min
Harry Tan released gstack — a workflow system for Claude Code with QA, review, and release
LLM·MarkTechPost

Harry Tan released gstack — a workflow system for Claude Code with QA, review, and release

Apr 30, 2026·3 min
US drops plan to tie AI chip sales to mandatory data center construction
LLM

US drops plan to tie AI chip sales to mandatory data center construction

Apr 30, 2026·2 min
Google radically updated Maps: Gemini answers questions, and navigation is now 3D
LLM

Google radically updated Maps: Gemini answers questions, and navigation is now 3D

Apr 30, 2026·3 min
Claude as Part of Thinking: Why Losing an AI Tool Disrupts Not Just Access, but Work Itself
LLM·Habr AI

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

Apr 30, 2026·2 min
Top themes
01Seldon Vault turned Asimov's psychohistory into a multi-agent AI forecasting serviceApr 30, 2026·3 min02The Lancet Psychiatry warns: AI chatbots may reinforce delusions in vulnerable peopleApr 30, 2026·3 min03Elon Musk restructures xAI again: co-founder departures, layoffs, and Grok falling behindApr 30, 2026·3 min04Anthropic and ETH Zurich: a long CLAUDE.md worsens agent performance and raises costsApr 30, 2026·3 min05Gemini 3.1 Pro outperformed ChatGPT 5.4 and Claude Opus 4.6 in a text generation testApr 30, 2026·3 min
Anthropic, OpenAI, and Cursor: Eight Levels of Agent Engineering Maturity
LLM·Habr AI

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 t

Apr 30, 2026·3 min
01Zhipu AI releases GLM-OCR, a compact 0.9-billion-parameter OCR model for documentsApr 30, 2026·3 min02LangChain released Deep Agents for multi-step AI agents with memory and isolationApr 30, 2026·3 min03AMD RX580 ran an LLM locally: how to tame ROCm, Ollama, and get GPU inferenceApr 30, 2026·3 min
Why Claude 4.6 Isn't Enough Without Context: The Main Blind Spot in LLM Development
LLM·Habr AI

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, a

Apr 30, 2026·3 min
AI forces universities to rethink term papers: the cheating problem predates ChatGPT
LLM·Guardian

AI forces universities to rethink term papers: the cheating problem predates ChatGPT

Dr. Nafisa Baba-Ahmed argues that AI did not create the crisis in university coursework, but merely scaled an old habit of outsourcing the t

Apr 30, 2026·2 min
Developer builds AI-powered news aggregator with MCP, DeepSeek, and Telegram bot
LLM·Habr AI

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.

Apr 30, 2026·3 min
BorisovAI tested MoE on an RTX 4090 and showed why perplexity breaks LLM evaluation
LLM·Habr AI

BorisovAI tested MoE on an RTX 4090 and showed why perplexity breaks LLM evaluation

Apr 30, 2026·3 min
LLM experiment showed how a model’s “personality” emerges in latent space
LLM·Habr AI

LLM experiment showed how a model’s “personality” emerges in latent space

Apr 30, 2026·2 min
Nvidia hints at an optical chip that could reshape AI data centers ahead of GTC 2026
LLM·Habr AI

Nvidia hints at an optical chip that could reshape AI data centers ahead of GTC 2026

Apr 30, 2026·3 min
Why DeepMind's AGI advances do not answer the key question of machine consciousness
LLM·Habr AI

Why DeepMind's AGI advances do not answer the key question of machine consciousness

Apr 30, 2026·3 min
IBM released Granite 4.0 1B Speech — a compact multilingual speech model for edge AI
LLM·MarkTechPost

IBM released Granite 4.0 1B Speech — a compact multilingual speech model for edge AI

Apr 30, 2026·3 min
Job cuts at Atlassian revived the debate: should AI cut working hours, not headcount
LLM

Job cuts at Atlassian revived the debate: should AI cut working hours, not headcount

Apr 30, 2026·3 min
OpenAI postponed the launch of erotic chats in ChatGPT over risks to teenagers
LLM

OpenAI postponed the launch of erotic chats in ChatGPT over risks to teenagers

Apr 30, 2026·3 min
Habr AI on the future of work: how AI and robots could return society to a new antiquity
LLM

Habr AI on the future of work: how AI and robots could return society to a new antiquity

Apr 30, 2026·2 min
Nvidia partner Hon Hai posts lower profit, deepening concerns over demand for AI servers
LLM·Bloomberg Tech

Nvidia partner Hon Hai posts lower profit, deepening concerns over demand for AI servers

Apr 30, 2026·3 min
Google removes AI feature from Search that showed medical advice from random users
LLM

Google removes AI feature from Search that showed medical advice from random users

Apr 30, 2026·2 min
Unity showed how to build voice NPCs with memory and world context
LLM

Unity showed how to build voice NPCs with memory and world context

Apr 30, 2026·2 min
Harvard: AI is cutting junior hiring, and in three years that could affect the entire industry
LLM·Habr AI

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.

Apr 30, 2026·3 min
Moonshot AI presented Attention Residuals — an alternative to residual connections in transformers
LLM·MarkTechPost

Moonshot AI presented Attention Residuals — an alternative to residual connections in transformers

Apr 30, 2026·3 min
DeepSeek and GLM-5 beat Yandex in a test of 34 AI models for managers without VPN
LLM·Habr AI

DeepSeek and GLM-5 beat Yandex in a test of 34 AI models for managers without VPN

Apr 30, 2026·3 min
Google released Gemini Embedding 2 for multimodal RAG with video, audio, and PDF
LLM·Habr AI

Google released Gemini Embedding 2 for multimodal RAG with video, audio, and PDF

Apr 30, 2026·2 min
Bitrix24 listed eight common mistakes in developing MCP servers for LLMs
LLM·Habr AI

Bitrix24 listed eight common mistakes in developing MCP servers for LLMs

Apr 30, 2026·3 min