Latest publications
All AI Bots Invent the Same Character — the Elias Thorn Phenomenon Explained
Dozens of different AI bots independently insert the same character — Elias Thorn — into their generated stories. The reason turned out to be safety mechanisms built into training.
ChatGPT reaches one billion monthly users—faster than any service in history
By the end of May 2025, ChatGPT reached one billion monthly users, surpassing all previous growth records for internet services, including Google Maps.
Bezos: AI will not destroy jobs, but will bring humanity a 'golden age'
Amazon founder is convinced that fears of technological unemployment are unfounded — AI leads humanity to a new era of prosperity.
Google launches background AI-agents in search — for now only for Ultra subscribers
Google has begun rolling out search AI-agents announced at the May conference: agents work in the background, with access currently limited to Google AI Ultra subscribers.
KPMG published a report on AI agents with hallucinations: fabricated quotes and non-existent cases
KPMG published a report on AI agents with false quotes and non-existent cases — it appears the document itself was written using AI without proper fact-checking.
SpaceX leases entire Colossus 1 data center: engineers unable to handle operations
SpaceX failed to effectively operate its Colossus 1 data center in Tennessee and decided to fully lease it out — despite acute shortage of computational power in the market.
Anthropic blocked access to flagship models due to US pressure
Startup Anthropic has closed access to advanced AI models for all users — a new conflict with US authorities has escalated beyond constructive dialogue.
Meta Stuck on Advertising: Why Diversification Isn't Working
Meta Platforms remains 97.6% dependent on advertising revenue despite diversification attempts — a classic success trap for internet giants. *Meta is recognized as an extremist organization and banned in Russia.
Munich Court Orders Google to Take Responsibility for Errors in AI-Generated Search Summaries
A Munich court ruled that Google must be held accountable for inaccurate information in AI-powered search summaries, despite the company's claims that they are accurate in the vast majority of cases.
Amodei agreed: governments should have the right to block dangerous AI
Anthropic CEO Dario Amodei stated that governments should be able to block the distribution of certain AI models if they pose a security risk.
Visa Allows OpenAI's AI Agents to Make Purchases on Behalf of Users
Visa has integrated into ChatGPT, giving AI agents built-in access to the payment network for conducting transactions and purchases on behalf of users.
Igor Babushkin from xAI Founded River AI Startup for Personalized AI
Igor Babushkin, one of the founders of xAI, announced the launch of River AI startup, which specializes in developing personalized artificial intelligence.
Students Have Lost the Ability to Read Long Texts — Teachers Blame AI
An American professor observed that his students can no longer manage a 20-page article, lose focus, and are unable to process complex texts — a problem that has already become systemic.
Google introduces DiffusionGemma — a model with an alternative approach to text generation
Google released an experimental DiffusionGemma model that uses a diffusion approach instead of a transformer — a fundamentally different text generation mechanism compared to modern chatbots.
OpenClaw Failed Phishing Tests: AI Agent Failed to Recognize Fraudulent Emails
Cybersecurity researchers tested the OpenClaw email AI agent and discovered a serious vulnerability: the agent failed phishing attacks just like ordinary office workers.
Meta to develop a smart necklace with a voice AI assistant for smartphones
Meta is preparing a smart necklace with voice control to compete with Apple, according to internal company documents disclosed in the press. *Meta is recognized as an extremist organization and banned in Russia.
Google to augment AI Overviews with citations from public forums and social networks
Google is revamping AI Overviews in search: the neural network will now insert direct quotes from Reddit, forums, and blogs, as well as links to articles from users' paid subscriptions.
Google launches cloud AI agent Gemini Spark for $100 a month
Google has introduced the cloud AI agent Gemini Spark, available in the premium Google AI Ultra plan. The agent works 24/7 and helps automate tasks in the browser and the cloud.
AWS Unveils OpenSearch Serverless — A Database for AI Agents
AWS announced OpenSearch Serverless — a managed database optimized for AI agents with vector search support and zero configuration.
Anthropic Confirmed Claude Mythos Release — But Delayed Due to Safety
Anthropic officially confirmed plans to launch Claude Mythos for a wide audience, but delayed the release, fearing risks to open-source code and commercial systems.
Investors Raised $36 Billion in Debt for Google Chips for Anthropic
Financiers attracted $36 billion in borrowed funds to purchase Google TPUs needed by Anthropic to develop an AI platform competing with OpenAI.
TrendForce Raises Memory Market Forecast to $1.28 Trillion Due to AI Agents
TrendForce analysts increased their memory market capacity forecast for 2027 by $437 billion—due to explosive demand from AI agents.
Company spent $500M on Anthropic Claude in a month without limits
Due to the lack of usage limits on the service, a company spent half a billion dollars on Anthropic Claude in just one month — one of the costliest mistakes in corporate AI history.
Lenovo stock soared 109% in a month — best result in 27 years
Lenovo stock rose 109% in a month following a financial report showing revenue growth from AI-system equipment despite component price increases.
Amazon banned employees from using AI just for show
The company discovered that many employees use AI not for genuine efficiency gains, but to appear engaged to management amid the broader AI trend.
DeepSeek Makes 75% V4-Pro Discount Permanent
DeepSeek has decided to make the 75% discount on V4-Pro permanent instead of ending it in late May — likely thanks to access to Huawei chips.
Anthropic Negotiates with Microsoft on Maia 200 Chips for AI Inference
Anthropic is negotiating with Microsoft about potentially using Maia 200 chips for running production AI models, seeking alternative sources of computational power.
Lam Research Integrates AI into Equipment to Reduce Chip Defects
Equipment manufacturer Lam Research integrates AI into semiconductor manufacturing equipment to reduce defects and increase chipfab productivity.
Alibaba Releases Free Qwen 3.7-Max — Rival to Claude Opus 4.6 Max
Alibaba has opened access to Qwen 3.7-Max — an AI-agent model that can plan tasks for hours and control robots, matching Claude Opus 4.6 Max capabilities.
Anker Launches Soundcore Liberty 5 Pro Earbuds with Custom AI Chip Thus
Anker equipped its new Soundcore Liberty 5 Pro wireless earbuds with the custom AI chip Thus, which improves noise cancellation and voice clarity during calls.