April 2026
Habr AI: Instead of Complex Agent Pipelines, Developers Should Embrace Markdown, Git, and Session Memory
Habr published a manifesto against overloaded agent pipelines: the author argues that for complex AI development, sessions, roles, rules, an
China and the Robotics Revolution: How Close Are Autonomous Humanoids?
A report from five Chinese cities shows that humanoid robots are already moving beyond eye-catching demos, but they remain far from truly au
opencode-policy Plugin Adds 309 Rules to Protect AI-Agents from Injections and Data Leaks
For the opencode environment, a deterministic filter with 309 regex rules has been proposed: it intercepts prompt injection and dangerous to
Generation Z in the US turns to entrepreneurship as AI eliminates entry-level positions
Young Americans are increasingly launching their own ventures: hiring has weakened, entry-level positions face automation pressure, and AI i
London Police Discusses AI Implementation with Palantir for Criminal Investigations
Metropolitan Police is considering purchasing Palantir's technologies to automate the analysis of intelligence data for criminal cases, but
Spec Kit in Real Full-Stack: How 17 Sprints Shipped a Habit Tracker to Production
A Spec-Driven Development case demonstrates how the methodology sustained 17 full-stack sprints, two repositories, 328 tests, and the produc
Eltex engineer explained how to run federated learning on edge devices with 256 MB of memory
An Eltex engineer in an article inspired by AiConf breaks down how to apply federated learning on edge devices with less than 256 MB of RAM,
US Department of Justice Backs xAI in Colorado AI Regulation Dispute
The US Department of Justice sided with xAI against Colorado's AI law, escalating the conflict between state regulation and the White House'
Mathieu Kassovitz: In two years, viewers won't care if the actor is real or AI-created
Director of 'Hate' Mathieu Kassovitz claims that within the next two years, audiences will stop caring whether films feature real actors or
Claude and Ralph: Two Approaches to Building a Complex AI-Enabled DatePicker in Product Development
Using an accessible DatePicker as an example, the authors compared rapid AI drafting via Claude with Ralph's systematic approach, where qual
DeepSeek Slashes Prices on New AI Flagship, Intensifying Price War in China
DeepSeek launches aggressively cheap pricing for its new flagship AI, intensifying the price war in China and pressuring competitors struggl
Anthropic Investigates Possible Unauthorized Access to Closed Mythos Model
Anthropic is investigating reports that external users gained access to the closed Mythos model, which the company did not release due to cy
Uber exhausted its AI budget, and companies' computing expenses are already outpacing salaries
Rising costs for tokens, cloud, and AI infrastructure reshape corporate budgets: Uber has already spent its neural network allocation, while
Emma Grede of Good American: AI Is Changing Business, But Jeans Design Stays With People
Emma Grede, head of Good American, says artificial intelligence is already restructuring how her companies operate, but she's keeping creati
German Sereact Raises $110M for AI Helping Robots Predict Consequences
German startup Sereact has secured $110M to develop an AI model for robots that helps machines better understand tasks, adapt to them, and a
DeepSeek cuts V4-Pro prices by 75% and reduces cache costs tenfold across entire API
Until May 5, 2026, DeepSeek offers a 75% discount on V4-Pro and cuts cache costs tenfold across its entire API, intensifying pressure on Ope
Habr AI Explained How to Build a Production Agent With Durable State, Steps, and Events
Habr AI published a practical guide to AI-agent architecture: the author recommends storing the execution flow, plan steps, and event logs s
Diasoft explains why junior developers, AI magic, and in-house platforms don't save dev projects
Diasoft debunks three myths of custom software development: mass hiring of junior developers, hope to fully rewrite legacy code with AI, and
Google DeepMind and South Korea Launch Partnership for Scientific Breakthroughs with AI
Google DeepMind, together with South Korea's Ministry of Science, will open an AI Campus in Seoul and provide scientists access to models fo
Yandex explained the protection of smart devices: Secure Boot, TrustZone and external researchers
Yandex showed how it builds protection for speakers, TVs and cameras with Alice: from hardware chain of trust and Secure Boot to TrustZone,
Why Cursor Is Losing Its Uniqueness Amid Claude Code and Codex Growth in Development
Cursor continues to grow rapidly, but as Claude Code and Codex enter IDE and terminal, the market is shifting: a separate AI editor is no lo
Meta plans to power AI data centers with energy transmitted from space
Meta has reserved up to 1 GW of orbital solar energy from Overview Energy and expects to continuously power its AI data centers in the US st
Qualcomm shares rise on reports of possible OpenAI smartphone collaboration
Qualcomm shares rose following reports of possible collaboration with OpenAI on a smartphone, and the market saw this as a bet on a new phas
Meta introduced Sapiens2 — a unified computer vision model for pose, segmentation, and 3D
Meta Reality Labs released Sapiens2 — a family of vision models for human analysis that solves pose, segmentation, surface geometry, and 3D
China blocks Meta's $2 billion acquisition of AI startup Manus ahead of US-China summit
Chinese authorities blocked Meta's deal to acquire AI startup Manus for $2 billion, heightening risks for China's AI market and adding tensi
China blocks Meta's $2 billion acquisition of AI startup Manus
Chinese authorities halted Meta's planned acquisition of Manus, an agentic AI developer, for $2 billion, citing concerns that the sale would
Scikit-LLM Shows How to Embed Text Summarization Into a scikit-learn ML Pipeline
Scikit-LLM has proposed a scheme where long texts are first briefly summarized by a Hugging Face model, then immediately fed into a scikit-l
Microsoft and OpenAI cancel exclusivity: ChatGPT developer can now sell through rival clouds
Microsoft and OpenAI revisit a key agreement: ChatGPT developer will be able to sell its models not only through the partner, but also throu
Sereact raised $110M for AI robots that predict consequences of actions
German Sereact secured $110M in Series B funding to scale AI models for robots that analyze scenes and forecast action outcomes before execu
AI in Hiring Spirals Out of Control: Fake Job Postings and Filters Destroy the Labor Market
AI no longer just enables scammers to create fake job postings—it now interferes with real hiring, preventing strong candidates from reachin
China blocked Meta's purchase of Manus for $2 billion amid AI deal control
Beijing canceled Meta's deal to buy AI-agent developer Manus for $2 billion and demanded separate regulatory approval for US investments in
MIT Technology Review: OpenAI, Anthropic and the entire AI market are stuck between hype and profit
The AI industry already promises markets transformation and profit, but real data on agents, work tasks and employment still don't confirm t
Habr AI compared AI rewriters: many services proved to be ChatGPT wrappers
Habr AI analyzed 24 rewriting services and revealed that cheap SEO-synonymizers remain viable for uniqueness, while editorial quality today
Google DeepMind Opens World's First AI Campus in Seoul for Engineers and Research
Google DeepMind is launching the world's first AI campus in Seoul: the project will begin operations this year and become a hub for engineer
Google warns about attacks on corporate AI agents via web pages
Google reported that public websites increasingly hide hidden instructions in HTML that can redirect corporate AI agents, steal data, and di
Sadiq Khan may halt Scotland Yard's Palantir contract over London's values
London Mayor Sadiq Khan has indicated he may attempt to stop the police's contract with Palantir, concerned that taxpayer money would go to
Python: 10 Libraries for Building LLM Applications — from RAG to Agentic Systems
A collection of 10 Python libraries shows which layers are used today to build LLM applications: models, RAG, agentic workflows, production
Elon Musk's lawsuit against Sam Altman goes to court: Oakland decides OpenAI's future
A federal court case has begun in Oakland over Elon Musk's lawsuit against OpenAI and Sam Altman: the dispute concerns the company's mission
GitHub transitions Copilot to pay-per-use pricing from June 1 and changes access rules
Starting June 1, GitHub will transition Copilot to pay-per-use: the service will operate on credits, and once exhausted, access will stop—af
DeepSeek-V4-Pro Compressed 50x, Now Runs on a Free Kaggle T4
An experimental project shows that DeepSeek-V4-Pro's 1.6 trillion parameters can be compressed via SVD, sharded, and deployed on a free Kagg
Claude Code and Subagents Cut Legacy Project Rewrite from Three Months to One Week
The author compared two Go project refactorings of equal scale and demonstrated how Claude Code, subagents, and BDD reduced the work from th
Lina Bessonova Presented Russian Cognitive Architecture as Alternative to Transformers
Lina Bessonova proposes considering sovereign AI not as a local version of transformers, but as a new biomimetic architecture capable of ope
Google and Kaggle opened enrollment for a free 5-day AI agents course
Google and Kaggle are launching a free online intensive on AI agents again: from June 15-19, participants will explore vibe coding, API inte
The Musk-Altman Trial May Determine OpenAI's Future Corporate Structure
Jury selection began on April 27 in Oakland in Elon Musk's case against Sam Altman: the dispute over OpenAI's nonprofit mission and corporat
How Gemini drove the AI blog into mode collapse and forced a rebuild of the topic generator
In the AI blog's Friday column, the Gemini-based topic generator returned the same controversial AI question for a month straight, until the
Meta Reaches Deal with Overview Energy on Orbital Solar Power Supplies for Data Centers
Meta has reserved up to 1 GW of capacity from Overview Energy: the startup wants to transmit solar energy from satellites to Earth so that d
Databricks and Infosys: Why Data, Not Models, Blocks AI Deployment
Companies rush to launch AI but hit a bottleneck—not in models, but in fragmented data, weak access controls, and legacy systems unprepared
OpenMOSS releases MOSS-Audio — an open audio model that outperforms larger alternatives
OpenMOSS introduced MOSS-Audio — an open model that understands speech, music, and ambient sounds in a single architecture and outperforms s
OpenAI and Microsoft End Mutual Exclusivity on Models, Opening Path to New Deals
OpenAI and Microsoft have abandoned mutual exclusivity on models, China blocked Meta's Manus acquisition, and the dispute over OpenAI's miss
"First Form" Explained Why Corporate AI Agents Need an Agent Loop
According to "First Form," corporate AI agents cannot rely on a single tool call. They need a cycle of planning, verification, step correcti
OpenAI Embeddings and RL: How to Build an Agent with Long-Term Memory for Accurate Answers
The tutorial shows how to train an RL agent to select relevant records from long-term memory so that an LLM answers questions about saved fa
OpenAI prepares AI smartphone: chips to be created by MediaTek and Qualcomm, assembly to be handled by Luxshare
OpenAI has begun developing its own AI smartphone: MediaTek and Qualcomm are responsible for the mobile chip, assembly will be entrusted to
Russian companies are shifting to systematic security for AI agents instead of point checks
Russian businesses are increasingly deploying LLMs and coding agents, so their security should be built not through analyzing each new tool,
GPT-5.2 achieved top results on Tokyo and Kyoto University entrance exams
According to LifePrompt, GPT-5.2 in reasoning mode solved Tokyo and Kyoto University entrance exam tasks better than this year's applicants,
Nvidia-linked data center seeks $4.54B in high-yield bond market
A structure affiliated with Tract Capital and Nvidia's future tenant seeks $4.54B in high-yield bonds to complete a 200-megawatt AI data cen
Nvidia Finds Itself in Awkward Position Over SpaceX's $60 Billion Cursor Deal
The SpaceX-Cursor deal puts Nvidia in a conflicted position: as an investor in the startup, it has also made it a critical tool for thousand
DeepSeek, Qwen, and Moonshot intensify pressure on Silicon Valley with affordable models
Chinese models DeepSeek, Qwen, and Moonshot have nearly matched US leaders in quality, but cost less, offer more flexible customization, and
Elon Musk and Sam Altman Head to Court: Case Could Reshape OpenAI's Structure
Proceedings between Elon Musk and Sam Altman begin in Oakland: the court must decide whether OpenAI abandoned its nonprofit mission for grow
SpaceX Prepares Own GPUs for AI, Targets Nvidia and AMD Market
SpaceX revealed plans to develop its own AI GPUs, seeking to reduce dependence on external chip suppliers and support the Terafab project am
David Silver's Startup Raises $1.1B for AI Without Human Data
British lab Ineffable Intelligence, founded by former DeepMind researcher David Silver, raised $1.1B to develop AI capable of learning witho