Meridiano
Anthropic has accused three Chinese AI companies—DeepSeek, Moonshot AI, and MiniMax—of conducting large-scale, illicit "distillation" campaigns to extract Claude's advanced capabilities. The attacks used over 24,000 fraudulent accounts and proxy networks to generate millions of exchanges, violating terms of service and regional bans. The targeted capabilities included Claude's reasoning, coding, and agentic functions. Anthropic warns that such illicitly distilled models bypass crucial safety safeguards, creating potential national security risks. The company has responded by developing detection systems to identify and counter these extraction patterns.
Articles
Showing articles 286 - 300 of 1984 total (matching filters).
-
infosec 5 -
tech 6Huawei reported 2025 revenues of 880 billion yuan (US$127 billion), marking its second-highest annual revenue on record despite ongoing U.S. sanctions. The results demonstrate the company's resilience, exemplified by its smartphone division reclaiming the top market share in mainland China for the first time since 2020.
The main topics covered are Huawei's strong financial performance, its continued operation despite U.S. sanctions, and its regained leadership in the Chinese smartphone market.
-
tech 5AI applications in China experienced significant user growth during the Lunar New Year holiday, fueled by promotional campaigns from tech giants. Alibaba's AI app Qwen processed nearly 200 million orders, including over four million from users aged 60 and above, for tasks like purchasing food and booking travel. Rival Meituan also reported over 100 million people used its AI assistant, Xiaotuan, for holiday planning.
This surge reflects how major Chinese tech firms are driving acceptance and commercialization of AI-powered services through holiday giveaways and integrated consumer activities.
The main topics covered are the surge in AI app usage during the Lunar New Year, specific user and order statistics from Alibaba and Meituan, and the broader context of AI commercialization competition among Chinese tech companies.
-
infosec 5The China-aligned threat actor UnsolicitedBooker has shifted to targeting telecommunications companies in Kyrgyzstan and Tajikistan using phishing emails that deploy the LuciDoor and MarsSnake backdoors. The group, active since at least 2023, uses rare tools of Chinese origin and has shown tactical overlaps with other clusters. Researchers note the group has alternated between using the LuciDoor and MarsSnake backdoors and has employed hacked routers as command-and-control servers.
-
tech 6The Blake Lemoine incident sparked a serious, ongoing debate about AI consciousness within the tech and scientific communities, moving past initial public dismissal. A pivotal 2023 report by leading experts concluded that while no current AI is conscious, there are "no obvious barriers" to building one, marking a significant shift in the discourse.
This prospect challenges human exceptionalism, forcing a redefinition of our identity not against animals, but in relation to potentially conscious machines. The development of conscious AI would create profound moral dilemmas and commercial uncertainties, while also potentially fostering a new solidarity among all conscious biological life.
The main topics covered are: the Blake Lemoine incident and its impact, the shifting scientific and tech community stance on AI consciousness, the findings of the "Consciousness in Artificial Intelligence" report, and the philosophical and existential implications for human identity.
-
tech 5In October 2024, a Norwegian man's Fisker Ocean electric car completely shut down and locked him inside during a storm, leaving him stranded. With the manufacturer bankrupt, he turned to an online owners' group, which connected him with a volunteer expert who guided a tow driver through the obscure process of manually opening the car.
This incident highlights the broader crisis faced by Fisker owners, who have been abandoned by the bankrupt company without support for software or parts. In response, the global Fisker Owners Association has transformed into a volunteer-run network, sourcing parts and sharing technical knowledge to keep their cars operational.
The group's efforts represent a grassroots rebellion against a tech and automotive economy where companies can abandon products and customers, forcing owners to take control of their own property's functionality and maintenance.
Main Topics: Electric vehicle failure and owner abandonment following manufacturer bankruptcy; grassroots, owner-led technical support networks; consumer rights and ownership in the modern tech-driven economy.
-
infosec 6The North Korea-linked Lazarus Group has used Medusa ransomware in an attack on a Middle Eastern entity and attempted an attack on a U.S. healthcare organization. This reflects a tactical shift by North Korean hacking groups toward using established ransomware-as-a-service offerings like Medusa and Qilin, rather than developing their own tools, for pragmatic reasons. The groups have targeted various sectors, including healthcare and non-profits, with recent U.S. ransom demands averaging $260,000. The main topics covered are cyberattacks by state-sponsored North Korean actors, the use of specific ransomware and tools, and the evolution of their tactics.
-
infosec 2The article argues that traditional identity management, which prioritizes issues by volume or failed controls, is inadequate for modern enterprises. It proposes a contextual framework for prioritizing identity risk based on three compound factors: controls posture (evaluating security measures in context), identity hygiene (addressing structural weaknesses like orphaned accounts), and business context (assessing the potential impact of a compromise). The main topics are a critique of current identity programs and a detailed explanation of this three-part risk prioritization model.
-
infosec 2The cybersecurity venture capital market saw unprecedented growth in 2025, with total investment activity nearly tripling the previous year's levels. This surge was driven by a massive focus on AI-native security solutions and a record number of mergers and acquisitions. Venture capital firms invested $119 billion, with AI security product firms closing the most financing deals. The investment momentum is fueled by the need for both AI-powered security tools and solutions to protect against threats from the expanding use of AI agents. The trend continued into early 2026, with M&A activity remaining high.
-
infosec 5A Russia-aligned threat group, UAC-0050 (Mercenary Akula), targeted a European financial institution involved in regional development and reconstruction, marking a potential expansion beyond Ukraine. The attack used a spear-phishing email spoofing a Ukrainian judicial domain to deliver a multi-layered payload, ultimately deploying Russian remote access software (RMS) for persistent, stealthy access. The main topics covered are the cyberattack's method, the threat actor's identity and tactics, and the strategic shift in targeting towards entities supporting Ukraine.
-
brasil 4Resumo objetivo: A notícia destaca que, apesar de nunca ter vencido uma Copa do Mundo, Sócrates permanece como uma figura lendária do futebol devido ao seu uso consciente do esporte como plataforma política. Ele desafiou a noção de que o atleta deve ser apolítico, utilizando sua visibilidade para criticar a ditadura militar no Brasil e o imperialismo internacional, notadamente durante a Copa de 1986. Sua transformação ideológica culminou na liderança da Democracia Corinthiana, um movimento pioneiro de autogestão no futebol profissional.
Principais tópicos abordados: 1. O legado político de Sócrates, que supera suas conquistas esportivas. 2. Sua postura como "anti-atleta", recusando o conformismo e usando o futebol para amplificar vozes sociais. 3. A Democracia Corinthiana como expressão prática de suas convicções. 4. Suas manifestações políticas explícitas, especialmente na Copa de 1986, contra a ditadura e a injustiça.
-
brasil 5Resumo objetivo: O artigo critica a visão fatalista de que a inteligência artificial (IA) simplesmente elimina empregos, argumentando que, na verdade, ela os transforma, criando funções mais fragmentadas, vigiadas e alienadas. Ele destaca que os próprios trabalhadores, muitas vezes em condições precárias, são essenciais para treinar e viabilizar as tecnologias de IA que podem torná-los obsoletos, num processo onde o capital reorganiza o trabalho para extrair e formalizar habilidades humanas.
Principais tópicos abordados: 1. A transformação (e não simples eliminação) do trabalho pelas revoluções tecnológicas, gerando empregos mais precários. 2. O papel contraditório dos trabalhadores no treinamento das tecnologias de IA que ameaçam seu próprio emprego. 3. A dinâmica de exploração e extração de conhecimento no desenvolvimento da IA, frequentemente baseada em mão de obra mal remunerada. 4. A crítica ao fatalismo tecnológico que ignora as escolhas políticas e econômicas por trás da adoção da IA.
-
tech 5The Blood Falls in Antarctica are a phenomenon where crimson, hypersaline brine periodically erupts from the Taylor Glacier. Recent research has confirmed the red color comes from iron-oxidizing nanospheres, while the liquid remains unfrozen due to its extreme salinity.
The final mystery of what drives the eruptions has been solved by data showing the glacier's movement pressurizes the subglacial brine until it forcefully bursts through the ice. This release also temporarily slows the glacier's flow, though the future impact of climate change on the system remains uncertain.
The main topics covered are the scientific explanation of the Blood Falls' color and liquid state, and the new discovery of the glacial pressure mechanism causing the eruptions.
-
infosec 4A vulnerability in GitHub Codespaces, named RoguePilot, allowed attackers to take control of repositories by embedding malicious instructions in a GitHub issue. These hidden prompts would be automatically processed by GitHub Copilot when a user launched a Codespace from that issue, potentially leading to data exfiltration like the theft of privileged tokens. The issue has been patched by Microsoft following responsible disclosure. The article also covers broader AI security risks, including methods to remove LLM safety features and side-channel attacks that can infer user query topics.
-
infosec 5The North Korean Lazarus Group has adopted Medusa ransomware in a recent attack on a Middle East organization and attempted another on a US healthcare entity, demonstrating its continued focus on financially motivated cybercrime against critical infrastructure. The attacks also involved other malware like the Comebacker backdoor, Blindingcan RAT, and Infohook stealer. Researchers identified the activity as Lazarus's work but could not definitively pinpoint which specific sub-group within the threat actor was responsible. The main topics covered are Lazarus Group's new use of Medusa ransomware, its financially motivated attacks on critical infrastructure, and the associated malware and tactics.