Kyiv: Ukraine managed to shoot down only one of 27 ballistic missiles due to a shortage of missiles for Patriot systems

The Large Hadron Collider (LHC) potentially could have detected signs of dark matter due to minor malfunctions in its operation. Research conducted by physicists Xunyu Liang and Ariel Zhitnitsky from the University of British Columbia in Canada suggests that these unusual events might point to the existence of this invisible substance. The article was accepted for publication in the scientific journal Physical Review D.

Ground Vibration and Accelerator Sensors

Scientists refer to these disturbances as 'UFOs' (Unidentified Flying Objects). In the context of the accelerator, this term refers to sudden proton beam losses caused by microscopic dust particles. The physicists were interested in what exactly causes this dust to detach within such a high-precision structure.

A possible answer lies in a hypothetical form of dark matter known as Axion Quark Nuggets (AQNs). Unlike traditionally sought subatomic particles, AQNs are hypothesized to be macroscopic and dense objects with a mass ranging from five grams to one kilogram.

As an AQN passes through the ground up to 100 kilometers from the laboratory, it could create a small acoustic shockwave capable of causing vibrations in the LHC. Researchers compare this device to a high-precision clock subjected to external vibrations. Liang and Zhitnitsky told ScienceAlert that 'a dark matter particle passing through the Earth over tens of kilometers can create a small vibration, similar to a light tapping on a table under a clock.'

This vibration, in turn, could lead to the shedding of dust, which then crosses the particle beam. The hypothesis arose from the coincidence of the sizes of the dust particles and AQNs, leading to the development of a new search model. The authors noted: 'Initially, we expected to find a direct link between the mysterious LHC 'UFOs' and dark matter. Later, we realized that we presented the first suggestion for using the LHC in the search for macroscopic dark matter.'

To identify the trace of an AQN, the data must show a very specific pattern along the machine's 27-kilometer tunnel. Calculations indicate that the passage of an object must provoke at least three 'UFO' events at different points on the ring within a total of two seconds.

This search does not require equipment upgrades or additional costs for the European laboratory. The analysis can be performed by reviewing the beam loss monitor data already installed on site. The physicists explained that 'our case is particularly attractive because the LHC already has excellent integrated monitoring systems installed to protect the machine from damage,' emphasizing that these devices may be sensitive to signals for which they were not originally designed.

The authors did not analyze the accelerator files directly, but the study serves as an invitation for the scientific community to re-examine old LHC records. If the pattern is confirmed, this discovery will expand the boundaries of physics by proving that dark matter can manifest on a macroscopic scale.

OpenAI, the company that sparked a global boom in artificial intelligence, has faced a loss of leadership to its main competitor, Anthropic, and is now actively working to recover its lost ground.

The growth rate of OpenAI's key product—the ChatGPT chatbot—has slowed down. Furthermore, there has been a change in management after Fiji Simo, who was considered a successor to CEO Sam Altman, left the company. Meanwhile, the sales department is forced to conduct an expensive campaign to attract corporate clients, offering them significant discounts and special terms.

According to information from The Wall Street Journal, leading OpenAI investors are expressing concern over the company's excessively high expenses relative to its growth rate. Some of these investors prefer to diversify their risks by directing funds to Anthropic. At the same time, Anthropic has demonstrated higher revenue growth rates than OpenAI, and its market capitalization has approached one trillion dollars thanks to the great success of the Claude Code coding tool.

In light of this, Anthropic is accelerating preparations for its initial public offering (IPO), scheduled for the autumn, and is negotiating with potential investors, emphasizing its advantage over the developer of ChatGPT.

Acknowledging Difficulties and Shifting Focus

OpenAI CEO Sam Altman acknowledged the difficulties, stating on the social network X that the past year had not been the most successful for the company due to his fault. He assured that the team is developing new solutions and intends to change the current situation. OpenAI's problems stem from the initial mistaken assessment of market development by the management. Altman focused primarily on increasing the number of ChatGPT subscribers, believing that ordinary users would increasingly use the chatbot in daily life.

However, the rapid success of Claude Code showed that the main profit is generated by providing specialized tools for programmers and large enterprises. While OpenAI was distracted by additional projects, such as the Sora video generator, developing its own chips, and user gadgets, the more compact and focused Anthropic seized the vacant niche with its code writing product.

To catch up with the competition, OpenAI released a series of new models aimed at professional tasks and programming. Company President Greg Brockman led the product line update process. In addition, the company signed an agreement with Amazon to sell its AI tools to the cloud giant's customers and hired Denis Dresser, former CEO of Slack, as Chief Revenue Officer.

Negotiations and Monetization Issues

Concurrently, Sam Altman is negotiating in Washington with representatives of the Donald Trump administration and lawmakers, presenting a new model in the context of discussions about industry regulation. OpenAI's management is also trying to determine a pricing strategy for the corporate product Codex, considering lowering the cost to capture market share, which could potentially negatively affect profits before the planned IPO. Consequently, OpenAI's IPO may be postponed until next year.

In a statement following Anthropic's filing, the company clarified that the process may take time, as some issues are easier to resolve in private company status.

Historical Issues with Codex and Competitive Pressure

The roots of the problems with Codex date back to the autumn of 2024, when OpenAI introduced reasoning models. These models successfully handled school Olympiad programming problems but proved insufficiently suitable for complex daily software development. In contrast, Anthropic released the Sonnet 3.7 model in February 2025, initially aimed at solving real business problems. Within OpenAI, the use of early versions of Codex also failed to meet expectations.

Management's attention was diverted by other issues, including the poaching of personnel by Meta's Mark Zuckerberg and the worsening relationship with the main investor, Microsoft. In December, OpenAI hastily released the GPT-5.2 model and struck a deal with the startup Cerebras to prevent Anthropic from collaborating with the chip manufacturer. However, these actions could not stop the user exodus.

While Altman was on vacation in the Caribbean, developers were massively switching to Anthropic's new Opus 4.5 model. An additional blow to OpenAI was the slowdown in ChatGPT's growth due to the growing popularity of Google's competing chatbot. The company failed to timely reach the goal of one billion weekly active users, although it later recorded this figure.

Attempts to Regain Initiative

In March, the situation caused serious dissatisfaction among OpenAI employees, who began questioning management about why the smaller Anthropic was setting technological and cultural trends. The departed Fiji Simo noted that OpenAI had lost direction due to secondary projects and had yielded to its rival in both research and products. Currently, OpenAI is trying to restore its dominant role. The company launched a superapp that combines Codex, ChatGPT, and a web browser, which increased the product's audience to ten million users, and also released the GPT 5.6 Sol model. Experts believe that OpenAI has significantly improved the user experience for developers and has a chance to regain its leadership.