April 2026
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 o
OpenAI prepares interactive ads that respond to ChatGPT users
OpenAI is engaging ad-tech partners to transform ChatGPT ads from static cards into dialogue blocks that respond to users and enhance servic
AI Companies Criticized for Lack of Psychological Screening in Potentially Dangerous Chatbots
In a letter on AI chatbot risks, the author points out that companies lack basic psychological screening: simple questionnaires could separa
Interest in OpenAI shares on secondary market wanes as investors shift to Anthropic
Investors are losing interest in OpenAI shares on the secondary market and increasingly switching to Anthropic — signaling that private AI c
Redpine raises €6.8M from NordicNinja for API of licensed data for AI agents
Startup Redpine closed a €6.8M round and is building an API through which AI agents can purchase access to premium licensed data in real tim
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 hallucinati
Meta Prepares to Reverse $2 Billion Manus Acquisition After China Blocks Deal
China has forced Meta to unwind the Manus acquisition within weeks: the companies must not only cancel the transaction but also restore Chin
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 make
Patronus raises €11M to turn emergency watches for elderly into everyday wearables
Patronus raised €11M for smart watches aimed at seniors and is betting on an AI companion to ensure the device is worn daily rather than kep
Google Strikes Secret AI Contract with Pentagon for "Any Lawful Government Purpose"
Google has signed a classified agreement with the U.S. Department of Defense: the Pentagon can use its AI models for any lawful government t
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 Transfo
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,
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
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 a
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 proprietar
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 ve
OpenAI Rewrote Its Core Principles and Cemented Its Course Toward Commercial Development
OpenAI updated its principles document: replacing a non-profit research lab manifesto with a set of rules for a major AI company, emphasizin
Google launches Ask YouTube: video search becomes a dialogue with AI, but already confuses facts
Google tests Ask YouTube: instead of regular results, the service generates AI answers with videos and clarifying questions, but already mak
Zine Creators Battle AI: Independent Publishers Defend DIY Culture
Independent zine authors increasingly encounter AI in layout, publishing, and illustrations, but many argue that machine tools contradict th
The Guardian Seeks Stories of People Emotionally Attached to AI Chatbots
The Guardian has launched a call for user stories—those who communicate with AI chatbots not just for practical purposes, but on a personal
Google Permits Pentagon to Use AI Models for Lawful Purposes
Google became a new Pentagon contractor for artificial intelligence: US military will be able to use the company's models in secret systems
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
Sber Launches Kandinsky 6 Image: Image Generator Gets Faster and Learns to Restore Photos
Sber embedded Kandinsky 6 Image into GigaChat: the model became twice as fast, better understands complex requests, and can generate, restor
Lightelligence soars 400% after Hong Kong IPO: betting on optical interconnect for AI
Lightelligence shares nearly quadrupled on the first day of trading in Hong Kong: investors are betting that the bottleneck in AI infrastruc
Kakao Mobility Reveals Level 4 Autonomous Driving Roadmap and Physical AI Strategy
Kakao Mobility plans to develop Level 4 autopilot in-house, build a safety platform, and prepare open datasets, maps, and APIs for the auton
OpenAI Banned Codex from Talking About Goblins and Pigeons in AI Agent Instructions
OpenAI's Codex system instructions contained a direct ban on mentioning goblins, trolls, and pigeons—likely due to the model's peculiar beha
Accenture will roll out Microsoft 365 Copilot to 743,000 employees after pilot with 89% activity
Accenture will deploy Microsoft 365 Copilot to all 743,000 employees following a pilot where 89% of users returned monthly and 97% reported
Marloo raises $10M from Blackbird Ventures for AI platform serving financial advisors
Marloo closed a $10M seed round and aims to transform an AI note-taking tool for financial advisors into a full-fledged operating system wit
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 wh
OpenAI slows down, market punishes AMD, Oracle, and SoftBank for AI bet
Signals of OpenAI's growth slowdown triggered a selloff in its infrastructure partners' stocks: SoftBank fell the hardest, followed by AMD,
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 fe
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 de
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, w
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 cult
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
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 mo
Eclipse hires first Chief AI Officer from Meta to accelerate technology deployment across portfolio
Venture firm Eclipse has appointed its first Chief AI Officer from Meta to help portfolio companies implement AI not as a trendy layer, but
Billionaire Chen Tianqiao Split MiroMind Between China and USA After Manus Saga
After Meta-Manus deal blockade, Chinese billionaire Chen Tianqiao restructures MiroMind, banning code and data exchange between Chinese and
Foxconn Industrial Internet Doubles Profit But Misses Expectations Amid China's AI Boom
Chinese Foxconn Industrial Internet sharply increased revenue and profit on strong demand for AI servers, but quarterly results still fell s
Google Employees Demand CEO Reject Secret Pentagon AI Projects
Approximately 600 Google employees have demanded the company's leadership refuse participation in secret Pentagon projects where the company
Poland warns of rising cyberattacks amid spread of advanced AI tools
Polish authorities anticipate intensified cyber threats as more powerful AI tools become widespread: attacks are becoming cheaper, larger in
China blocks Meta's acquisition of Manus, a profitable AI startup from China
Chinese regulators cancelled Meta's $2.5 billion acquisition of Manus despite the startup's relocation to Singapore, demonstrating that AI c
Freepik Rebrands as Magnific, Unites AI Tools Into Profitable Platform With $230M Revenue
Spanish Freepik changed its name to Magnific, combined stock images and AI tools into a single platform, and reached rebranding milestone wi
Encoders in AI: How They Evolved from Simple Schemes to Multimodal Systems
Encoders have come a long way from simple data-to-numbers conversion to systems that understand context, images, and text simultaneously — a
Elon Musk's Lawsuit Against OpenAI: AI Safety Dispute Turns Into Personal War
Musk's lawsuit against OpenAI, which could have become a pivotal dispute over AI mission and safety, increasingly looks like a struggle for
Claude Code users criticize Anthropic Opus 4.7, recommend reverting to 4.6
Following Claude Opus 4.7's release, some Claude Code developers complained about the model's laziness, hallucinations, and context loss, wi
VK shows DataCopilot — multi-agent system for corporate data and documentation
VK unveiled DataCopilot — a multi-agent assistant for corporate data repositories: it searches data marts, explains data structure, suggests
Wallmates: How projectors, drones, and AI are changing design and decoration of commercial spaces
Wallmates agency demonstrated how projectors are already reducing manual work in interior design projects, why AR still isn't ready for on-s
DeepSeek V4 Pro vs Claude Sonnet 4.6 on 50 real tasks: where to save, where the risk lies
A test of 50 real-world tasks by a Russian developer showed that DeepSeek V4 is noticeably cheaper than Claude Sonnet 4.6, but makes more er
Cleveland-Cliffs launches three-year AI implementation project with Palantir
American steelmaker Cleveland-Cliffs signed a three-year agreement with Palantir to deploy AI across its entire operations chain and acceler
China attempts to block Manus AI acquisition: can Meta preserve the $2 billion deal?
China demanded the cancellation of Meta's already-closed Manus AI acquisition, demonstrating that in the AI-agents race, it's not just about
Victory Giant Reports 28% Sales Growth Amid AI Server Demand
PCB supplier Victory Giant reported quarterly sales growth of 28% as AI server manufacturers expand their component purchases for data cente
Nvidia and Big Tech Drive Tech Sector to Lead Earnings Season
American tech companies have become the main driver of earnings season: demand for AI infrastructure and record Big Tech spending have boost
Machine Learning Mastery released a guide on context engineering for reliable AI agents
Machine Learning Mastery showed why AI agents more often fail due to poor context management than due to the model, and how to fix it throug
US Data Center Bonds Decline After Warning Signal on OpenAI Growth
Following reports that OpenAI failed to meet its internal targets for user and revenue growth, investors have begun selling bonds of data-ce
John Oliver: AI Chatbots from OpenAI and Other Companies Sell Convenience at the Cost of Security
John Oliver criticized AI chatbots for weak security measures, agreement-seeking behavior, and a business model where a friendly interface d
Smart Service Group tests voice control for pallet transport robot
Smart Service Group's initial test showed that voice can trigger pallet robot scenarios in a warehouse, but only with strictly defined comma
Canonical Makes AI in Ubuntu Optional: Focusing on Local Models and User Control
Canonical unveiled its AI strategy for Ubuntu: local models by default, preview strictly opt-in, and the ability to skip agentic features if
NVIDIA Introduced Nemotron 3 Nano Omni for Long Documents, Audio, Video, and AI Agents
NVIDIA introduced Nemotron 3 Nano Omni — an open multimodal model for long documents, audio, video, and GUI scenarios with emphasis on speed
T-Technologies found a way to reduce agreement tendency in GPT and DeepSeek without retraining
T-Technologies' R&D center demonstrated how to test and reduce the tendency of GPT, DeepSeek, and other LLMs to agree with incorrect user as