April 2026
T-Bank improved AI Code Completion quality with a filter and removed unnecessary suggestions
T-Bank implemented a two-layer filter for AI Code Completion: first, an LLM evaluates whether a suggestion is relevant, then CatBoost refine
Meta cuts external moderation and shifts content review to more advanced AI
Meta plans to significantly reduce its reliance on moderation contractors in the coming years and hand over the detection of fraud, illegal
LG and NVIDIA discuss physical AI: robots, data centers, and mobility
LG CEO Ryu Je-chol met with NVIDIA's director of Omniverse and Robotics in Seoul — talks show what infrastructure is needed to deploy physic
BotHub, GPTunnel, and ruGPT: how Russian AI aggregators are dividing the market in 2026
A comparison of BotHub, GPTunnel, ruGPT.io, Syntx AI, and AI Wiz found that Russian AI aggregators compete not on price, but on convenience,
Cursor prepares a new AI model for coding to compete with OpenAI and Anthropic
Cursor is preparing to release a more efficient AI model for software development, trying to keep up with OpenAI and Anthropic in the rapidl
Machine Learning Mastery identified five major barriers to scaling agentic AI in 2026
Machine Learning Mastery compiled five problems preventing agentic AI from transitioning from impressive demos to stable production: from or
OpenAI plans to acquire Astral and double down on Python tools for developers
OpenAI is preparing to acquire Astral, a startup that builds tools for Python developers, to deepen its presence in the fast-growing market
Oasis Security raised $120 million to manage access for AI agents and service accounts
Startup Oasis Security, which helps companies manage access for AI agents and other non-human accounts, raised $120 million from major inves
DoorDash Starts Paying Couriers for Data to Train AI and Robotics
DoorDash in some regions pays couriers not only for deliveries but also for videos and digital tasks, turning gig work into a tool for train
AtScale: corporate AI agents for analytics need guardrails, not larger models
AtScale believes that corporate AI agents in analytics fail not due to 'weak' models, but due to poor governance: without unified semantics
Jan Lane: why AI threats are undermining cyber defense even with companies' record budgets
Visio Cyber AI founder Jan Lane believes companies are losing to AI attacks not because of a lack of budgets, but because of fragmented tool
World Bank shifts focus to AI-resilient industries to boost employment
The World Bank is changing its development strategy: when launching employment programs, it wants to give more weight to AI’s impact and sup
Mike Pepi in the Guardian proposed a 1% tax on AI-slop to protect creative labor
A proposal in the Guardian would impose a 1% levy on the largest AI companies for generative content, to offset the flood of AI-slop and sen
Habr AI described the architecture of a reflective agent: less autopilot, more self-checking
Habr AI showed how a reflective agent works: with drafts, pauses, rollback, and mandatory self-checks before acting on code, data, and busin
Habr AI: How prompt engineering in development saves hours but does not replace understanding the tasks
Habr AI examines practical prompt engineering for development: good instructions really do save hours, but they only work together with an u
Google DeepMind Introduces AI co-clinician System for Doctors and Telemedicine
Google DeepMind launched a research AI co-clinician system that helps doctors with clinical questions and is being tested in telemedicine, b
OpenAI shows how ChatGPT Images 2.0 can redesign app interfaces
A journalist uploaded screenshots of two of his products to ChatGPT and received a UX analysis, new mockups, and ideas that can be passed al
Democratic Alliance ministers in South Africa embroiled in scandal over AI errors
In South Africa, two Democratic Alliance ministers have faced the fallout from AI hallucinations: one withdrew a policy draft, the other sus
NSA tests Anthropic Mythos model to find vulnerabilities in Microsoft products
The US NSA is assessing whether the new Anthropic Mythos model can find vulnerabilities in widely used software, including Microsoft product
1X opened a factory in California and wants to produce 10,000 home robots in a year
Backed by OpenAI, 1X has opened a factory in California and plans to produce 10,000 home robots in its first year — one of the most ambitiou
Claude Code hides history: 715 sessions found on disk, though the interface shows 69
A look inside Claude Code showed that the app stores hundreds of old sessions locally but does not show them in the interface: 48 MB of hist
Eurozone finance ministers to discuss Anthropic's Mythos model amid rising concerns
Finance ministers of eurozone countries will return to the topic of Anthropic's Mythos on May 4, as regulators and banks try to understand h
Roblox prepares the Reality platform with AI rendering and a push for photorealistic graphics
Roblox is building Reality, a hybrid platform where a cloud engine runs the game and AI adds textures, lighting, and environments to signifi
Elon Musk testifies for a third day in court in case over OpenAI’s founding history
On April 30, 2026, Elon Musk’s case against Sam Altman entered a new phase: after a tough cross-examination, the court will continue examini
BlackRock speeds up product launches with AI: prototypes for Aladdin are now built in days
BlackRock COO Rob Goldstein said the company is already using AI to build working prototypes in days instead of months and to improve transp
OpenAI revealed the source of the «gremlins» in ChatGPT and showed how to remove the restriction in Codex
OpenAI explained why GPT models started inserting goblins and gremlins into responses, linked it to the Nerdy personality mode, and even sho
AI platform developer Andata avoided bankruptcy after layoffs and automation
Andata repaid its debt to the Federal Tax Service a day before the bankruptcy hearing, cut part of its team, and introduced AI into key proc
Anthropic launched Claude Security: AI service finds code vulnerabilities and suggests patches
Anthropic brought Claude Security to public beta: the Opus 4.7-based service scans repositories, reduces false positives, and suggests which
Goodfire releases Silico, a tool for debugging language models during training
Startup Goodfire unveiled Silico, a mechanistic interpretability system that helps look inside a language model, find failures, and change i
Sapphire CFO Solutions on why AI complicates financial decisions for startups
Sapphire CFO Solutions founder Heather Hall says that AI not only automates finances but also complicates tool selection, budgets, and manag
AI boom buoyed the US economy and softened the blow of war-driven inflation in the first quarter
Growing AI investment helped US GDP accelerate to 2% in the first quarter, even as the war involving Iran drove up oil prices, inflation and
Google allowed ads in Gemini after previous commitments against advertising
Google no longer rules out ads in Gemini: the company is testing formats in AI Mode and allowing their transfer to the chatbot, while OpenAI
Anthropic withholds Mythos from public release over the risk of attacks on critical infrastructure
Anthropic has restricted access to Mythos because the model is too effective at finding vulnerabilities and, in the hands of malicious actor
Claude and Ollama helped an Ozon seller process certificates in a day and restore 1,000 SKUs
An Ozon seller used Claude and a local Ollama model to recognize 231 certificates, upload them in bulk via the API, and put more than 1,000
AWS showed how Amazon Bedrock AgentCore Gateway connects to private APIs and services
AWS showed how Bedrock AgentCore Gateway, via Resource Gateway, gives AI agents access to private APIs and services inside a VPC in managed
Amazon unveiled an agentic analytics architecture based on SageMaker, Athena, and Quick
AWS described an architecture in which Amazon Quick, with agentic AI on top of SageMaker, Athena, and S3, lets business users ask questions
Musk admitted: no written agreement existed with OpenAI at its founding
Elon Musk confirmed in court that when he made his initial contribution to OpenAI more than a decade ago, there was no written agreement on
OpenAI gains freedom from Azure: Microsoft rewrites contract with key partner
Microsoft and OpenAI have revised their long-term partnership: OpenAI can now sell models through AWS, Google Cloud and other providers — Az
How Sun Finance and AWS accelerated document verification and reduced fraud risk
Sun Finance built an AI pipeline on AWS for identity verification: data extraction accuracy rose to 90.8%, verification time fell to under 5
AWS introduced a system for migrating and upgrading LLMs in production with prompt optimization
AWS introduced the Generative AI Model Agility Solution — an approach for migrating and upgrading LLMs in production with prompt conversion,
BioticsAI: how to navigate the FDA, raise money, and motivate a team in healthtech
BioticsAI CEO Robhi Bustami explained how the healthtech startup is handling FDA approval, attracting investment, and keeping the team motiv
X relaunches its AI-powered advertising platform to restore revenue growth
X is relaunching a rebuilt AI-powered advertising platform and betting on campaign automation to win back advertisers and restore revenue gr
Musk admitted under oath: xAI used OpenAI models for distillation when building its systems
Answering lawyers' questions under oath, Elon Musk did not deny that xAI used OpenAI models for distillation and called it standard practice
Stripe introduced Link — a wallet for online payments and purchases via AI agents
Stripe introduced Link, a digital wallet that brings together cards, bank accounts, and subscriptions, and also lets users securely grant AI
Salesforce handed part of its AI roadmap to customers and sped up releases of Agentforce and new features
Salesforce builds AI products not on a rigid plan, but on weekly feedback from customers to ship Agentforce, voice AI, and Slack features fa
Google brings Gemini to cars with Google built-in and will update millions of vehicles
Google has started rolling out Gemini in cars with Google built-in: the assistant will replace Google Assistant, understand natural speech,
Meta confirms leak: AI agent advice exposed sensitive data to employees
At Meta, an internal AI agent suggested a fix to an engineer, after which company employees gained access for two hours to a large volume of
Nvidia CEO Jensen Huang urged AI companies not to scare people with apocalyptic rhetoric
At GTC 2026, Jensen Huang criticized the grim rhetoric around AI, said the main threat is not the technology itself but fear of it, and defe
Claude Code without the magic: Habr breaks down the architecture, context noise, and engineering practices
Habr has published a detailed breakdown of Claude Code: why the agent loses efficiency in long sessions, how context gets consumed, and why
Ollama and LiteLLM: how to turn a Python script into a complete console-based LLM chat
In the second part of the practical guide, the author shows how to turn a basic Python script with Ollama and LiteLLM into a console chat wi
Sber showed how RAG and LLM in the IDE turn manual scenarios into automated tests
Sber described an IDE plugin that uses RAG and LLM to search for examples in the codebase and turn manual scenarios into automated tests, wh
VS Code 1.111 launched Autopilot: AI agent writes, tests, and deploys on its own
Microsoft released VS Code 1.111 with Autopilot mode — the AI agent now works autonomously and completes the entire task without the develop
MIT conference: Karen Hao and Paola Ricaurte on the right course for AI
At an MIT conference, journalist Karen Hao and researcher Paola Ricaurte discussed who controls AI's trajectory — and how to put the technol
Habr explained why language models and classic RAG lose their understanding of relationships
Habr examines why large language models with classic RAG are good at finding text fragments but struggle when they need to reconstruct relat
Apollo.io acquired Pocus to strengthen its AI sales platform and push further into enterprise
Apollo.io is acquiring Pocus and adding signal analytics to its GTM platform to push further into enterprise and bring sales, data, and AI a
OpenAI is devoting its core resources to building a fully autonomous AI researcher
The company is restructuring its research agenda around an autonomous agent that will be able to conduct multi-day research on its own, with
Claude and other AI services help Americans file tax returns
Tax preparation in the US is becoming a new niche for AI: Claude and other tools are already helping users navigate forms, rules, and calcul
Nvidia explained the principle behind DLSS 5: the system only needs a 2D frame and motion vectors
Nvidia revealed the basic principle behind DLSS 5: the neural network reconstructs the image not from a full 3D scene, but from a regular 2D
Alibaba shipped nearly 500,000 AI chips, but admitted it lags behind Nvidia
China's Alibaba said it has shipped 470,000 of its own T-Head AI chips, but openly acknowledged that they trail Nvidia in performance, and i
Generative AI tested in a US Supreme Court case: can it replace lawyers
A lawyer handling a case before the US Supreme Court tested generative AI on materials from his own case, and the experiment again raised th