Latest publications
Google and OpenAI employees supported Anthropic in its dispute with the Pentagon
Anthropic’s conflict with the US Department of Defense over the ethics of AI use sparked a wave of solidarity among leading Silicon Valley engineers.
The xAI founders’ exodus: why Musk’s team is losing key people
Toby Pohlen became the seventh co-founder to leave xAI after its merger with SpaceX. The company Elon Musk created to compete with OpenAI is rapidly losing the people who were there at the beginning.
DeepSeek entrusted the final tuning of its latest model to Huawei, not Nvidia
Chinese developer DeepSeek chose Huawei and other local accelerator makers to optimize its latest AI model. It is a sign that the technological decoupling between the US and China is entering a new phase.
Google challenges Nvidia in the AI chip race: prospects and obstacles
The AI accelerator market has grown used to the Nvidia and AMD duopoly, but Google is scaling its TPU ambitions so quickly that it can no longer be ignored. How real is this threat?
Tech companies are forcing employees to use AI — whether they like it or not
The technology sector is no longer asking workers what they think: companies are imposing strict mandatory-use policies for AI tools, turning artificial intelligence from an option into a corporate norm.
DeepSeek trained a new model on banned Nvidia Blackwell accelerators
China's DeepSeek, according to U.S. sources, used Nvidia Blackwell accelerators subject to sanctions to train its latest AI model — chips that are barred from being shipped to China.
Altman proposed counting the cost of intelligence differently: 20 years of food versus one data center
The head of OpenAI entered the debate on AI energy efficiency with an unexpected argument: to compare the brain and a neural network fairly, you have to account for all the calories a person consumes over two decades of
Google pays cloud providers to switch to its own accelerators
Google has started offering financial incentives to cloud providers to buy its in-house AI accelerators instead of NVIDIA solutions. The move resembles OpenAI’s tactic of «circular deals», but competitors are skeptical.
Altman acknowledged: Chinese AI is advancing «surprisingly fast»
The head of OpenAI publicly noted the rapid progress of Chinese rivals. Behind the compliment lies concern — and possibly a bid for new investment and regulatory easing.
Nvidia Ready to Invest $30 Billion in OpenAI Amid Stalled Megadeal
While the fate of the $100 billion September round remains uncertain, Nvidia is in talks over a new $30 billion investment in OpenAI. What is behind this move, and why it changes the balance of power in the industry.
OpenAI moves into hardware: its first device will be a smart speaker with a camera
OpenAI is preparing its first hardware product — a $200–300 smart speaker with a camera, facial recognition, and the ability to “see” the surrounding world. The company is challenging Amazon and Google on their own turf.
OpenAI and TSMC prepare to launch their own Titan AI chips
Sam Altman's startup is moving into production of its own hardware. The first 3-nanometer processors for training neural networks could roll off the line as early as 2024.
OpenAI targets a record $850 billion valuation
The startup plans to raise more than $100 billion in a new funding round to cement its status as the world's most valuable AI developer.
SpaceX and xAI to develop voice control for combat drones
Elon Musk enters the arms race: SpaceX and xAI will compete for a Pentagon contract to build intuitive voice-based control systems for drone swarms.
Anthropic releases Sonnet 4.6: fewer errors and access for everyone
The new Claude Sonnet 4.6 version brings advanced coding skills, improved reasoning, and long-context support. The update is already available to all chatbot users for free.
OpenAI strengthens ChatGPT protections with new security features
OpenAI introduces tools to combat abuse: Lockdown Mode and a risk assessment system. The new features aim to protect data and prevent cyber threats.
AI market correction: Big Tech firms lose billions after two years of growth
The euphoria around neural networks is giving way to market pragmatism. We explain why the market capitalization of tech giants has started to decline and which companies are holding up.
Pentagon threatens Anthropic with sanctions over ban on military AI use
The US Department of Defense may deem Anthropic a security threat over the company’s refusal to allow the Claude model to be used for surveillance and autonomous weapons.
Alibaba unveils Qwen 3.5: open source and support for 201 languages
The Chinese giant has released a new language model focused on AI agents that outpaces rivals in speed and availability.
OpenClaw creator joins OpenAI
Peter Steinberger, known for the OpenClaw AI agent, has joined the OpenAI team.
Trust crisis: why leading scientists are leaving OpenAI and Anthropic en masse
Leading AI developers are sounding the alarm: the rapid neural network race is pushing the creators of these technologies to leave the largest labs in protest over ignored risks.
OpenAI permanently retires GPT-4o despite widespread user protests
OpenAI has officially ended support for GPT-4o and related models. Despite a wave of community backlash, the company decided to carry the shutdown through to the end.
Anthropic surges to $380 billion: results of a record Series G round
Startup Anthropic raised record funding, cementing its status as OpenAI’s main rival. The company’s valuation surged to $380 billion after the Series G round.
OpenAI introduces GPT-5.3-Codex-Spark based on Cerebras chips instead of Nvidia
OpenAI has introduced the new GPT-5.3-Codex-Spark model. Its main feature is that it runs on giant Cerebras WSE-3 chips instead of the usual Nvidia accelerators. This marks the beginning of a new era in AI infrastructure
GPT-4o shutdown: users mourn the loss of a 'friend
The news about the discontinuation of GPT-4o, one of OpenAI's most emotional and human-like models, has provoked an unexpected reaction among users.
ChatGPT Will Speak Arabic: OpenAI Creates Version for UAE
OpenAI plans to release a special version of its popular ChatGPT chatbot designed for the United Arab Emirates (UAE).
Claude to the Sixteenth Power: AI Agents Reached for the Sacred
Writing a compiler — this is a kind of initiation rite for any serious programmer. It's an incredibly complex engineering task, requiring an understanding of how abstract ideas transform into ones…
CarPlay opens to outsiders: why Apple no longer trusts Siri
Imagine: you're driving down a nighttime highway, hands on the wheel, and you urgently need to find not just the nearest gas station, but a place that serves decent coffee after ten in the evening…
Trillion dollars on the line: why the market stopped fearing AI spending
Imagine you walked into a restaurant, ordered dinner, and at the end received a bill not just for the food, but also for building a new kitchen, buying the adjacent plot of land, and training the…
Footprints in the sand: AI settled paleontologists' long-standing disputes
Paleontology has always been somewhat like fortune-telling with coffee grounds, except instead of a cup you have a multi-ton block of sandstone, and instead of grounds you have a barely visible…