Latest publications
How Boris Cherny manages 2,000 overnight AI agents from a smartphone
Boris Cherny, creator of Claude Code, launches 2,000 AI agents for overnight code automation and manages them through a smartphone — he spoke about it in an interview.
Kling AI and AI Video entered App Store top charts after Sora's shutdown
After Sora's exit, two AI video generators rose to fifth and sixth place in Apple App Store downloads, taking the top spots in their categories.
Anthropic integrates Claude into QuickBooks, HubSpot and Canva for small businesses
Anthropic launched Claude for Small Business, a package of agent workflows that embed AI into QuickBooks, PayPal, HubSpot, Canva, Docusign and Google Workspace. No additional costs.
Apple to allow AI agents in the App Store under strict restrictions
Apple is developing its own standards for allowing autonomous AI agents into the App Store, seeking to retain control over platform safety and revenues.
Anthropic: AI of the future will anticipate user needs
Anthropic manager Kat Wu explained how AI will learn to work proactively — anticipating and solving user tasks without explicit commands.
Microsoft reveals the true scale of its OpenAI investment — over $100 billion
Microsoft invested in OpenAI far more than had previously been publicly reported. The company’s actual spending had already exceeded $100 billion by June this year.
Microsoft rushes to buy AI startups to avoid dependence on OpenAI
Microsoft is grappling with competition from OpenAI and is revising the terms of their joint deal. The company can now invest in AI startups on its own and is looking for acquisition targets. The goal is to reduce depend
Apple defended Google: criticism of EU plan to open Android for AI
Apple sent a letter to the European Commission criticizing the proposal to force Google to open Android components to competing AI services. This is a rare case of defending a competitor.
SK hynix catches up to Samsung: AI boom pushes Korean manufacturer toward $1 trillion
SK hynix has approached a $1 trillion market capitalization thanks to global demand for memory chips for artificial intelligence systems, following its rival Samsung Electronics.
Samsung Galaxy Z Fold8 to Be First to Get Gemini Intelligence
Google announced Gemini Intelligence, a suite of AI features for smartphones. Samsung Galaxy Z Fold8 and Z Flip8 will get them first this summer, ahead of the standard flagship Pixel 10 and Galaxy S26.
Anthropic revealed that Claude writes 90% of the company's code
Anthropic's CFO said that its AI, Claude, now writes more than 90% of the software code — a record figure among tech giants in development automation.
By 2027, 80% of premium smartphones will get AI agents — on Qualcomm and MediaTek chips
Counterpoint Research analysts predict that by 2027, most flagship smartphones will offer autonomous AI agent features capable of handling multistep tasks without user intervention.
Google to penalize sites for trying to manipulate AI answers in search
Google has added new anti-spam rules: attempts by publishers to appear in AI answers will be punished with lower rankings in regular search results.
OpenAI merges ChatGPT and Codex into an agent platform ahead of IPO
OpenAI has reorganized, combining ChatGPT and Codex into a single agent platform. Greg Brockman has taken over the company's entire product organization as part of preparations to go public.
Alibaba, Baidu and Huawei move away from Nvidia in favor of their own AI accelerators
Major Chinese tech companies have accelerated the rollout of their own AI accelerators, building a strategy of independence from US exports — even amid rumors of Nvidia's return to the market.
YouTube makes search for deepfakes using your face available to all users
YouTube has expanded access to its tool for searching AI deepfakes: now any user over 18 can find fake videos created using their face.
Micron and the memory market: industry participants do not expect significant price declines for several more years
Memory market participants warn that the AI server boom is breaking the usual pricing cycle: after prices rise, a drop may not follow, and chips may stay expensive for years.
Google launches Gemini on Samsung Galaxy S26: AI will order taxis, food, and purchases
Google has begun rolling out Gemini's agentic feature on the Samsung Galaxy S26 in the US and South Korea: the AI gets access to third-party services and can place orders and bookings on the user's behalf.
Meta delays launch of flagship AI model Avocado until at least May 2026
Meta has shifted the release of its new AI model Avocado from March 2026 to at least May after internal evaluations showed the flagship trailing the best systems from competitors.
Yandex has moved all Station and TV Station devices to a more powerful Alice model
Yandex has opened Alice’s advanced AI model to all Station speakers and TV Station televisions: answers have become longer, more accurate, and noticeably less likely to require a repeated question.
Interpol: AI scams earn fraudsters 4.5 times more than traditional scams
Interpol assessed the profitability of AI-enabled fraud: such schemes already earn criminals up to 4.5 times more because they are easier to scale and personalize.
Micron to build second memory packaging facility in Taiwan next to Powerchip plant
Micron is expanding capacity in Taiwan: next to the facility it acquired from Powerchip, the company will build another memory packaging site to meet growing demand from AI systems.
Microsoft launched Copilot Health to analyze data from smartwatches and medical records
Microsoft has launched Copilot Health, a new section in Copilot that analyzes health data from wearables, medical records, and lab tests.
ByteDance postponed the global launch of Seedance 2.0 over copyright infringement risks
ByteDance has pushed back the global release of the Seedance 2.0 video generator: the company fears that familiar images and storylines in AI videos could lead to copyright claims.
DeepSeek prepares V4: a multimodal model with a context window of up to 1 million tokens
DeepSeek is preparing V4: the new multimodal model is said to have a context window of up to 1 million tokens and a noticeable gain in capabilities over the previous version.
Meta AI agent triggered an internal failure and gave engineers access to restricted data
Meta's internal AI agent posted advice on a corporate forum without permission, after which employees accidentally opened access to sensitive systems and data for two hours.
Jeff Bezos seeks $100 billion for Project Prometheus — an AI startup for industry
The founder of Amazon and Blue Origin is raising up to $100 billion for Project Prometheus, a startup that aims to introduce AI into production processes and boost the efficiency of entire industries.
Google to add conversation branching to Gemini for parallel chat continuation
Google is preparing a conversation branching feature for Gemini: users will be able to branch off from any message and develop multiple paths in parallel without losing the shared context.
AMD and Samsung expand HBM4 partnership for next-generation AI accelerators
AMD and Samsung signed a new memorandum of understanding in Pyeongtaek on HBM4, the memory set to become a key component of AMD's future AI accelerators.
Sam Altman's World tests AgentKit to link AI agents to iris scans
World has begun limited beta testing of AgentKit, a system that links AI-agent actions to the owner's identity verified through an iris scan.