deep

The U.S. Department of Justice stated on Thursday that North Korean hackers have stolen over $400 million in cryptocurrency assets through a social engineering scheme based on secretly recorded videos and AI deepfakes over the past 18 months, and announced that it has seized the infrastructure supporting this operation. The FBI refers to this method as "FaceStealer," which centers on turning video conference calls into traps: hackers lure targets into meetings using previously stolen accounts from victims, then play pre-recorded footage of trusted contacts to simulate their presence, thereby gaining the target's trust.The group also uses AI to generate deepfake videos of well-known figures in the cryptocurrency and related industries, making the scams more convincing. Once the target lowers their guard, the attackers create false technical issues, such as claiming the other party cannot hear them, and provide a so-called "fix" link that is actually used to implant malware. Afterward, hackers can empty the victim's wallet or exploit their system permissions, sometimes lying in wait for months or even over a year before taking action. The FBI stated that it has seized over 900 segments of camera footage from FaceStealer victims based on federal search warrants.The Department of Justice linked this operation to North Korean-related groups such as Sapphire Sleet, BlueNoroff, and APT38, warning that these hackers are increasingly targeting open-source software developers, seemingly paving the way for supply chain attacks. On the same day, the Department of Justice also reported progress in the Binance theft case, stating that it has obtained court authorization to seize or restrict over $1.2 billion of the approximately $6.4 billion in stolen assets, which is part of the over $2.7 billion that has already been frozen. The FBI attributes this theft to the North Korean group TraderTraitor and claims it is laundering money at a rate of creating a new wallet approximately every 25 seconds. North Korea denies involvement, stating that the FBI's conclusions are a political provocation.

According to Jinshi reports, in the third quarter of this year, the concentration of South Korean exports further deepened. Influenced by the prosperity of the semiconductor industry, the gap between large enterprises and small and medium-sized enterprises widened, with the top ten export companies in South Korea accounting for more than 60% of the total export value, marking the first time since records began in 2010.Data released by relevant departments on Tuesday showed that South Korea's total exports in the third quarter surged by 72.3% year-on-year, reaching 318.6 billion USD, setting a record high for a quarter since records began in 2015. The export value of large enterprises grew by 103.5%, with semiconductor manufacturers like Samsung Electronics and SK Hynix being the main drivers, while the export value of small and medium-sized enterprises only increased by 12.8% and 10.7%, respectively. During the same period, semiconductor exports skyrocketed by 219% year-on-year, reaching 148 billion USD, accounting for 46.5% of the total exports, up from 25.1% in the same period last year.

Adobe announced a deepening of its artificial intelligence collaboration with Google, bringing leading AI models and integrated workflows to creatives and businesses.

According to an interview with "Daily People," a data center operations engineer who has worked in Ulanqab, Inner Mongolia for 5 years stated that local AI data centers are still under large-scale construction, but the hiring demand at his company has significantly decreased. The company hired thousands of people in 2025, but by the time of the interview in 2026, only 3 new hires had been made.The engineer, who goes by the pseudonym Zhang Liwei, originally worked on an assembly line at an automobile factory. When he first switched careers, his internship monthly salary was 2,000 yuan, which increased to 4,500 yuan after graduation. After 5 years of job hopping and promotions, he currently earns over 10,000 yuan per month. In April of this year, DeepSeek was hiring senior operations engineers and senior delivery managers for its data center in Ulanqab, offering a monthly salary of 15,000 to 30,000 yuan, with 14 months' salary. These positions require extensive experience in data center operations and project management. According to Zhang Liwei's observations, local ordinary duty staff still earn around 5,000 to 6,000 yuan per month, while assistant engineers can earn over 10,000 yuan.The entry barriers are also rising. In 2024, some data centers were still hiring newcomers without relevant professional backgrounds; however, this year, some companies have required candidates to have a relevant degree and work experience, and there is even a 35-year age limit for basic duty positions. Zhang Liwei mentioned that a former colleague has started a training class for operations, charging nearly 20,000 yuan per student, and classes only start when there are 60 students, with a single session fee approaching 1.2 million yuan. The so-called internal recommendations are mostly just to help students secure an interview opportunity.

On November 8, Tencent's AI product WorkBuddy released a statement and compensation announcement regarding the DeepSeek-V4.1 Flash discount event. On September 10, WorkBuddy launched DeepSeek-V4.1 Flash and simultaneously started a two-week 50% discount points experience event, which was originally scheduled to end on September 24. Subsequently, WorkBuddy extended the discount until November 5 and restored the original price starting from 14:00 on November 5. Due to the delay in announcing this adjustment to users, WorkBuddy expressed its apologies for the inconvenience caused.WorkBuddy has decided that for accounts that used DeepSeek-V4.1 Flash from 14:00 on November 5 to 23:59 on November 8, all points consumed during this period will be settled at the original price with a 50% discount. The difference, which is 50% of the total points actually consumed, will be fully refunded. No application is required for the refund, as it will be automatically calculated by the system and is expected to be credited within 7 working days.

According to IT Home, on the afternoon of November 7, multiple users reported that DeepSeek experienced service issues, with a dialogue prompt stating "Server temporarily unavailable." The official status page of DeepSeek shows that the issue has been resolved and services have been restored. The duration of the interruption for the webpage and API was from November 7, 15:11 to 15:56, totaling 45 minutes.

According to the "Science and Technology Innovation Board Daily," the artificial intelligence company DeepSeek has once again experienced issues with its chat function, with the chat interface temporarily displaying "Server busy, please try again later." The official status page indicates that starting from 15:11 today, there have been partial interruptions to the website and API, and the official team is investigating the issue to identify the cause.The affected services include the chat service, DeepSeek V4 Pro API service, and DeepSeek V4.1 Flash API service. The latest update on the DeepSeek official website states that the interruption issue has been addressed, and they are currently monitoring the effectiveness of the fix.

Google DeepMind posted on the X platform that it has released a new study demonstrating how AI agents can accelerate stress testing of cryptographic systems. Under human guidance, its model identified two significant vulnerabilities in the traditional 1024-bit RSA encryption system.DeepMind stated that modern 2048-bit systems are not affected by these specific vulnerabilities, but this work indicates that as the era of AI-driven mathematics and computation arrives, reliance on historical assumptions of computational difficulty can no longer be maintained, and continuous stress testing of cryptographic standards is essential.

According to Gelonghui, the DeepSeek Harness team released data on the evening of November 5, showing that as of November 4, the daily active users of DeepSeek Harness, based on users utilizing the official API, have reached 1.54 million, with daily token consumption of approximately 92 trillion.The data indicates that on September 30, the daily active users of DeepSeek Harness (7-day average) were 271,000, reaching 429,000 on October 7, and climbing to 1.54 million on November 4, an increase of over one million in just one month.

According to trader Ansem, Zcash has recently been one of the stronger assets in the cryptocurrency market. If the market experiences a deeper correction, he will pay attention to whether ZEC can maintain its resilience.He believes that as the cryptocurrency industry faces more political and regulatory scrutiny, the narrative around privacy may become increasingly important. Privacy features, concerns about quantum computing, and market interest in transparent blockchain alternatives have drawn his attention to ZEC.

The Science and Technology Innovation Board Daily reporter learned from multiple sources that the venture capital firm Lisi Capital, which participated in the previous round of financing for DeepSeek, continues to secure an investment share in DeepSeek's new round of financing, with the participating entity being its subsidiary Tianjin Lisi Xing Shen Equity Investment Partnership (Limited Partnership) (Xing Shen Fund). As multiple listed companies disclose their funding progress and business information updates, the fund size has increased from the previously disclosed 2.64 billion yuan to 6.02 billion yuan, and was ultimately confirmed at 7.52 billion yuan, with the funds already paid in.Public announcements show that Zhihu announced on September 6 that it plans to invest 1.5 billion yuan through its wholly-owned subsidiary to participate in the Xing Shen Fund, which completed registration with the Asset Management Association of China on September 15; Tongchen Beijian disclosed in late September that it participated with its own funds of 50 million yuan when the fund size was 2.64 billion yuan. On October 9, Zhongwang Fabric announced its intention to invest 50 million yuan, at which time the subscribed amount had increased to 6.02 billion yuan. On October 21, Tongchen Beijian disclosed that it received notice that the fundraising scale had increased from 2.64 billion yuan to 7.52 billion yuan, the fundraising has been completed and all funds have been paid in; Perfect World, Te Bao Biotechnology, and others subsequently confirmed this scale. According to Qichacha, the Xing Shen Fund has a total of 41 partners, including the executive partner Tianjin Lisi Ming Tang and 40 limited partners, with a total subscribed capital of 7.52 billion yuan.The reporter compiled that 13 listed companies and their related funding entities have a total subscribed amount of 3.495 billion yuan, accounting for approximately 46.48% of the total fund size. Zhejiang Innovation Deep Intelligent Venture Capital Partnership has subscribed 1.4 billion yuan, making it the second-largest contributor; along with the two funds of Zhejiang Province, Ke Chuang Zhan Xin and Ke Chuang Wei Lai, each subscribing 50 million yuan, the three Zhejiang state-owned entities have a total subscription of 1.5 billion yuan, while Wuxi Chanfang Guosheng Innovation Development Equity Investment Partnership has subscribed 80 million yuan. A person familiar with the situation stated that Lisi is one of the market-oriented institutions that obtained a larger share, while other market-oriented institutions typically have quotas ranging from 3 billion yuan to 4 billion yuan. Zhihu previously announced that the Xing Shen Fund adopts a blind pool structure, and this person indicated that the actual allocation may still be highly concentrated in DeepSeek. The reporter had previously sought confirmation from Lisi Capital on whether the Xing Shen Fund participated in this round of financing, but had not received a clear response by the time of publication.

According to a report by The Paper, in response to rumors circulating online about the preparation of an AI startup focused on Recursive Self-Improvement (RSI) with a project valuation of up to $2 billion, Google DeepMind senior research scientist Yao Shunyu explicitly denied it, stating, "I have not founded a company and am still an employee of Google DeepMind," and claimed to have no idea where the related rumors originated.Previously, there were reports that Yao Shunyu's startup team had basically taken shape, intending to attract research and engineering talent from leading AI labs such as Google DeepMind, Anthropic, and OpenAI. This is not the first time Yao Shunyu has responded to startup rumors. On October 12, he publicly stated that there are currently no plans to raise funds in China and urged the public not to waste time on related rumors.

On the afternoon of November 3, a reporter from Yicai tested and found that the dialogue function of DeepSeek, a subsidiary of the artificial intelligence company Deep Pursuit, was unavailable, with the chat interface displaying "Server busy, please try again later."The official status page of DeepSeek shows that starting from 16:58 on the same day, there were partial interruptions to the DeepSeek website and API, and the official team is investigating the issue to identify the cause. Affected services include the dialogue service, DeepSeek V4 Pro API service, and DeepSeek V4.1 Flash API service.

According to a report by the Science and Technology Innovation Board Daily on November 3, reporters learned from multiple industry chain insiders that the artificial intelligence company DeepSeek's next-generation large model, DeepSeek V5, plans to be trained based on Huawei's Ascend chips. Previously, the Ascend chips were mainly used for inference tasks of the DeepSeek large model.The report stated that at the end of September this year, DeepSeek open-sourced infrastructure components for the Huawei Ascend computing power platform, covering the TileLang high-level language compilation tool, computing libraries, and distributed communication libraries, which correspond one-to-one with the previously open-sourced components for the NVIDIA platform. The report was written by Huang Xinyi, a reporter for the Science and Technology Innovation Board Daily.

DeepSeek showcased the creator mode of the AI workstation Harness. Harness currently offers four modes: Standard, PTC, Minimal, and Creator. Among them, the Creator mode allows users to directly let AI develop plugins, modify software interfaces, and then install them for use in Harness without needing to write code themselves.Users only need to describe their requirements, and the AI will check Harness's development interface, write the plugin code, complete the installation, and check the running effects. The installed plugins can be saved for continued use the next time they open the application. In addition to adding new interfaces, it can also provide tools for the Agent, enabling it to perform operations it originally could not. For example, users can have AI create a video editing tool with a timeline, a paper reader that supports underlining and annotations, as well as generate poster design tools, interactive 3D models, or even change the Harness interface to a pixel game style.In the official demonstration, the AI created a floating tomato timer, writing over 600 lines of front-end code, completing the installation and operation, taking a total of 5 minutes and 24 seconds. DeepSeek also released the prompts for 7 demonstration cases. Some works were modified through multiple rounds of dialogue, and it is not possible to achieve the same effect by simply inputting a single sentence.

Chinese artificial intelligence company DeepSeek has announced a design plan aimed at enhancing the autonomy of AI coding assistants, allowing them to reduce human intervention while performing multiple tasks. According to a report by the South China Morning Post on Sunday, the plan originates from a paper published by the DeepSeek Harness team, with the core goal of reducing the human effort required to coordinate and monitor multiple AI agents.The team pointed out that managing multiple agents consumes human attention and cited a statement from the paper that "the single-task limit depends on model capability, while the multi-task limit depends on human attention." The paper also stated that one person can only supervise 3 to 5 active AI agent conversations at the same time. The proposed system will allocate tasks and coordinate work among the agents, allowing humans to free up energy to review results and make decisions.This proposal comes at a time of intense competition among AI developers, involving American companies Anthropic and OpenAI. Previously, Cointelegraph reported in October last year that, according to data from blockchain analysis platform CoinGlass, Chinese AI models outperformed their American counterparts in cryptocurrency trading, with DeepSeek being the only AI model in an experiment to achieve a 9.1% positive unrealized return, while the development cost of this model was only a small fraction of that of its American competitors.