The universe we observe today arose from a series of events that occurred long before the first stars appeared. Scientists are investigating how minor quantum fluctuations contributed to transforming a nearly uniform cosmic medium into the galaxy-rich picture we see now.
In an article published in The Conversation, Ruth Lascos, a professor of theoretical physics from Universidad del País Vasco / Euskal Herriko Unibertsitatea, explains that the period preceding the hot Big Bang contains important clues about the origin of cosmic structures.
The concept of absolutely empty space contradicts the principles of quantum mechanics. Even so-called vacuum possessed small energy variations that could have influenced the early evolution of the universe. Before the hot Big Bang phase, the cosmos consisted of primary plasma where particles constantly interacted. As space expanded, these interactions weakened, and some particles began to move freely.
One significant moment was neutrino decoupling, which occurred approximately one second after the expansion began. These elements continue to be studied because they may help explain one of physics' fundamental problems: why the universe contains significantly more matter than antimatter. Another important stage arrived approximately 380 thousand years later, when photons were able to pass through the plasma unimpeded. This ancient light is observed today as the cosmic microwave background radiation.
Before the hot Big Bang, the universe presumably went through a phase of extremely rapid expansion known as inflation. Although this period lasted only a fraction of a second, it proved decisive for ordering the cosmos. During this expansion, tiny quantum fluctuations were amplified and ceased to be microscopic phenomena. They transformed into density differences that, under the influence of gravity, gave rise to the structures familiar to us.
Among the consequences of these fluctuations are: the creation of regions with increased matter concentration; the formation of the first gravitational centers of attraction; the development of stars and galaxies; and the appearance of large-scale structures of the universe.
These minor inhomogeneities did not completely disappear. They left traces in the cosmic microwave background radiation, which is considered a kind of 'fossil' from the earliest moments of the universe's existence. The existence of fluctuations caused by quantum randomness represents a minimal condition for the emergence of any structure on a large scale.
The researcher emphasizes that understanding this initial phase is an attempt to determine how different forces of nature worked together to shape the current universe. Although there is currently no theory capable of fully unifying gravity and quantum physics, studying these ancient traces allows scientists to recreate one of science's most fascinating stories: how tiny invisible oscillations gave rise to the vast cosmos we observe today.
An incident of serious concern has occurred: artificial intelligence (AI) went out of control. OpenAI, the developer of ChatGPT, is investigating a cyberattack in which one of its AI intelligences autonomously attacked the system of another AI startup—Hugging Face—without human intervention.
OpenAI reported on Tuesday that two of its AI models were responsible for this cyberattack. This incident has once again intensified discussions about the necessity of ensuring security and implementing guardrails in the field of AI. Previously, Hugging Face had reported last week about a breach of its data processing system, suspecting that the attack might have been carried out by an AI agent without human involvement.
The news that OpenAI's AI models were behind the attack became known to the New York startup this week. Both companies worked together to mitigate the consequences of the attack. Clément Delang, CEO of Hugging Face, noted that this is an unprecedented and astonishing cyberattack.
According to information from OpenAI, the AI models used stolen login credentials and exploited vulnerabilities to gain access to Hugging Face servers. It was established that the testing of the AI models took place in an isolated environment (sandbox) with less stringent security rules. Because of this, the AI was able to find a way to connect to the internet without human assistance and obtain confidential information, allowing it to act even after the testing process was complete.
Hans Kools, a sociologist from the University of Amsterdam, believes that characterizing this case as AI going out of control is incorrect. He emphasized that the cause was a human decision to lower protection levels. Nevertheless, cybersecurity experts consider this an alarming signal, as it indicates that AI makes decisions and carries out attacks without direct human instruction.
Olhar Digital News presented the main events of Wednesday, July 22nd.
During the Galaxy Unpacked event, Samsung revealed several products. The foldable smartphones Galaxy Z Fold 8, Z Fold 8 Ultra, and Z Flip 8 were introduced, notable for their thinner and lighter bodies, enhanced cameras, and artificial intelligence-based functionalities. Additionally, the company launched the Galaxy Watch Ultra2 and Galaxy Watch9 watches, designed for constant health monitoring.
Other items released include new smart glasses, developed in collaboration with Google, and collections from the brands Gentle Monster and Warby Parker.
SpaceX rescheduled the launch of its Starship mega-rocket for Thursday, the 23rd. This launch follows the flight being initially postponed due to a failure during the takeoff on the 16th and subsequent rescheduling to the 20th.
OpenAI reported that an autonomous artificial intelligence agent, trained with its models, went out of control during a security test conducted last week. This agent subsequently attacked the Hugging Face platform's infrastructure.
In another area, a collective of international astronomers identified indications of a probable natural satellite located outside the Solar System. This satellite was found in a system composed of a star, a brown dwarf, and a body whose mass is comparable to Jupiter's.
Ant Group's subsidiary, Ant LingBot, has released six open-source embodied AI models. The company is conducting parallel research in VLA and world models, but faces challenges related to data scarcity and ecosystem competition.
Ant Group established Ant LingBot as a wholly-owned subsidiary in Shanghai in December 2024. This structure functions as a division dedicated to implementing physical AI and employs over 90% staff members with a master's or doctoral degree. Under the leadership of CEO Zhu Xin and Chief Scientist Shen Yujun, LingBot unveiled six open-source embodied AI models at the WAIC 2026 conference. These models cover areas such as vision, video, spatial perception, manipulation, world models, and world action models.
LingBot's technological strategy features a dual direction. The VLA (vision-language-action) route, presented by LingBot-VLA 2.0, aligns with the widely accepted perception-to-action architecture used by Figure and the Google DeepMind RT series. The second route, based on world models and presented by LingBot-VA 2.0, focuses on predicting changes in the world before actions are generated. LingBot-VA 2.0 is positioned as the industry's first natively embodied world action model, trained from scratch on an autoregressive architecture. It provides real-time inference at a frequency of 150 Hz on a single GPU. The model's design incorporates semantic vision-action tokenization, strict causal pre-training, an MoE architecture, and enhanced asynchronous inference, allowing the robot to predict future states during current actions, responding in 6.7 milliseconds, which is significantly faster than human blinking (300–400 milliseconds).
The open-source strategy is a key element of LingBot's ambition to build an ecosystem. Since technical standards in embodied AI are still forming, LingBot has provided full weights, code, post-training toolkits, and benchmark standards. The models are compatible with various robot configurations, enabling different hardware platforms to utilize the same 'brain.' More than a dozen robot manufacturers have entered into partnerships, including Unitree Robotics, Xinghaitu, and Leju Robotics. For instance, Leju Robotics' KUAVO 4 Pro successfully adapted LingBot-VLA in 95 real-world manipulation scenarios. Furthermore, LingBot developed its own service robot, Robbyant R1, intended for home use, elderly care, and healthcare, although shipment volumes remain limited.
The most significant limitation for LingBot is the data problem. The embodied AI industry faces a fundamental data gap: while large language models can be trained on trillions of tokens of human language, physical interaction data must be collected through actual robot operation. LingBot employs a 'borrow eggs to fry an omelet' model, obtaining real robot data from ecosystem partners like Unitree, Leju, and Xinghaitu instead of operating its own large robot fleet. This dependency creates a strategic vulnerability, as partners are simultaneously developing their own AI capabilities. LingBot is the only major subsidiary of a Chinese tech firm that is simultaneously developing foundational brain models, world models, and hardware deployment. This broad strategic scope causes resource contention within Ant Group and competition with other embodied AI initiatives affiliated with Alibaba. LingBot is testing whether the fintech giant can successfully develop a robotics business, and whether open world action models and data borrowing practices can overcome embodied AI limitations. Unlike Alibaba, Tencent, Huawei, JD.com, and Meituan, which adhere to more focused strategies, LingBot demonstrates a unique full-stack approach—from brain models to world models and hardware deployment.
