wen

According to Hard AI reports, Qianwen is about to launch the Personal Agent product "Qianwen Go!". In user-shared experience videos, the product can access flight booking platforms to find delayed flights and negotiate with customer service multiple times to secure a cash compensation of 70 yuan for users; it can also search through orders on Taobao that meet price protection conditions to recover a price difference of 52 yuan for users.Reports indicate that based on the Agent architecture, Qianwen Go has upgraded its scope of services, capable of handling tasks that require repeated checking, communication, and comparison. The initial application scenarios also include car maintenance inquiry appointments, sports venue bookings, and calculating the final price of refrigerators by combining subsidies, old machine recycling, and installation costs. The product operates in a cloud environment and can continue working even after the user closes the app. During task execution, it needs to determine the next steps based on results, such as changing entry points, platforms, or communication methods.Bank of America research on October 7 suggests that the Agent helps users complete tasks and purchases more smoothly, which is expected to expand overall online activities. Morgan Stanley estimates that the personal-oriented Chinese consumer AI monetizable market could reach approximately 300 billion yuan by 2030 and may reach 1.6 trillion yuan by 2040, with major revenue coming from advertising and transaction commissions. Citigroup raised its CPU market forecast on October 6, expecting the global CPU market size to grow from 29 billion dollars in 2025 to 300 billion dollars by 2030.

Investment bank TD Cowen's Washington research group Managing Director Jaret Seiberg stated in a report released on Monday that the proposed rulemaking advance notice issued by the CFTC last week is a positive step for crypto platforms, but the agency may not complete the rulemaking until early 2028. As the 2028 election cycle approaches, there is a risk that the rule may ultimately not be implemented. Seiberg noted, "Time is running out," and expects that the CFTC will not issue proposals until the end of the first quarter or early second quarter of 2027, with final rules possibly not being finalized until early 2028.Last week, the CFTC proposed a federal regulatory framework for crypto platforms but avoided direct jurisdiction over spot market trading. According to Seiberg, the proposed rules will require platforms to maintain "reserve proof audits" and grant the CFTC jurisdiction over state-licensed platforms. Previously, the U.S. Senate failed to pass legislation that would grant the CFTC broad powers over spot crypto trading in a preliminary vote, leading the CFTC and SEC to intervene in regulating the crypto industry, including the introduction of long-awaited innovation exemptions, but both agencies lack the permanence that legislation could provide.Seiberg also pointed out that if the CFTC decides to expedite the rulemaking process, it may actually increase the risk of court intervention. He stated that speeding up this process would raise the risk of judges blocking the rule on procedural grounds, and even if the agency is cautious enough, legal challenges could extend beyond 2028.

Tsinghua-based AI infrastructure company Wuwen Xinqiong has completed a C round financing of 2.3 billion yuan, bringing the total financing to 4.3 billion yuan. The company was established in 2023, does not develop chips, nor does it build its own data centers. It integrates the computing power of different chips such as Huawei Ascend, NVIDIA, and AMD through software, providing large model training and inference services to customers.Developers can directly call the models via API without needing to deploy them themselves. Bloomberg reported on October 1, citing informed sources, that Wuwen Xinqiong has secretly submitted a listing application to the Hong Kong Stock Exchange, with the earliest possible listing in the first half of 2027, and the pre-listing valuation is reportedly around 14.3 billion yuan.

Coinbase co-founder and CEO Brian Armstrong stated: The company has fine-tuned the open-source weight model Qwen3.5-9B using internal proprietary data, which performs better than cutting-edge models in the specific task of fraud prevention, while also being more cost-effective and faster.Armstrong mentioned: Cutting-edge models are excellent general models, but companies will increasingly use specialized models for narrower tasks, which can operate independently or collaborate with teams. He believes this pattern will become more prevalent within the company.He also noted that Coinbase is making it easier for employees to fine-tune models internally and providing relevant skills in internal tools, with example instructions for fine-tuning Qwen for payment risk audits. He anticipates that some backend services may be completely replaced by this new architecture.

According to on-chain analyst Ai Yi's monitoring, the Binance contract real smart money account "Single Wife" holds 18.29 million Lobster long positions, with an opening price of $0.1277. When Lobster reached its historical high of $0.3143 on September 21, the account's peak floating profit once reached $2.19 million, with a position value of $5.75 million.The price of Lobster subsequently fell, and the account closed its positions on October 2 at an average price of $0.09374, ultimately incurring a loss of $907,000, with all previous floating profits completely wiped out.

According to Onchain Lens monitoring, a newly created address (0xbcd...bbc) went long on 54.1 BTC with 40x leverage on Hyperliquid, with an opening price of $81,778.6, a liquidation price of $80,819.1, an unrealized loss of $1,200, and a total loss of $20,200.

TD Cowen Managing Director Lance Vitanza stated at the Bitcoin Treasuries conference in New York that institutional investors are no longer debating whether to hold Bitcoin, but rather discussing how to hold it. He predicts that the price of Bitcoin will reach $132,000 by 2027.Vitanza pointed out that Bitcoin is evolving from a single asset into a capital market ecosystem that encompasses common stocks, preferred stocks, bonds, and income-generating products. He analyzed the assessment methods for digital credit and which Bitcoin treasury companies can survive in a market downturn, mentioning Strive, Metaplanet, and Nakamoto, emphasizing the importance of having actual operating businesses.He also discussed the potential impact of MSCI index exclusions on Bitcoin treasury companies, as well as blockchain monitoring, front-running behavior, and trust issues regarding Bitcoin prices. He believes that well-operated Bitcoin treasury companies may perform better than Bitcoin itself.

According to Lookonchain monitoring, trader address 0x914b shorted 14,976 ETH (approximately 40.97 million USD) and incurred a loss of 471,000 USD, then immediately opened a 25x leveraged long position with a size of 23,734 ETH (approximately 64.30 million USD), with the current liquidation price at 2,650 USD.

According to on-chain analyst @ai_9684xtpa's monitoring, a certain address opened a position on Hyperliquid for the first time, going long on 121.23 BTC with 7x leverage, valued at 10.27 million USD, with an opening price of 84,918.9 USD, currently incurring a loss of 20,000 USD.

Conway Research founder Sigil Wen announced the launch of Underdog, a private personal AI that can run on users' existing devices. Sigil Wen stated that Underdog has received support from a16z, Khosla Ventures, Hummingbird VC, Anthology (Anthropic and Menlo Ventures), as well as individuals and institutions in the AI field such as Patrick Collison, Naval, Guillermo Rauch, Noam Brown, Thomas Wolf, and Logan Kilpatrick.According to the announcement, Underdog's mission is to provide free, accessible, reliable, and private AI to billions of people. The team proposed Underdog's Law: that cutting-edge intelligence will reach users' devices within six months. The project will start with high-speed AI that can run on consumer-grade hardware, advancing capabilities, speed, power consumption, and data efficiency through collaborative design of models and reasoning engines.The announcement also stated that Underdog's research covers agent commerce, confidential reasoning, and how AI can reshape the internet economy, believing that privacy and capability no longer need to be mutually exclusive.

According to monitoring by Lookonchain, trader 0xd9A2 went long on 2900 ETH with 20x leverage on Aster DEX, with a position value of approximately 7.97 million USD.Currently, the unrealized profit of this position has exceeded 1 million USD, with a return rate of 251.68%.

According to Lookonchain monitoring, a certain trader went long on 1.38 million ASTER on the Aster platform with 5x leverage, with a position value of approximately 1.05 million USD. Currently, the position has an unrealized profit of over 137,000 USD, with a return rate of about 65.36%.The above profit and loss are unrealized gains and may change with market conditions.

TD COWEN initiated coverage on SpaceX, giving it a buy rating with a target price of $200.

FastBiz officially stated that the project has entered the continuous operation and ecological construction phase. As of now, FastBiz has accumulated a total construction amount of 15.9436 million USD, with over 37,000 participating users. Users can participate in ecological construction using USDT and UNI according to the agreement rules. As the scale of participation continues to expand, FastBiz is accelerating community building, ecological applications, and global market layout, promoting more users and business resources to enter its Web3 business ecology.It is reported that FastBiz is positioned as a new type of business infrastructure connecting real-world commerce and the Web3 value network, committed to creating the next generation of decentralized business networks that connect global goods, supply chains, payments, communities, and digital assets. The project connects real-world business scenarios with the Web3 digital asset system, exploring the integration of diverse scenarios such as goods circulation, supply chain collaboration, digital payments, community participation, and digital assets, further constructing an open value network aimed at global users and business entities.

According to Caixin, poker master and seasoned cryptocurrency investor Hu Zheweng made significant bets during the three waves of cryptocurrency frenzy involving ICOs, algorithmic stablecoins, and AI tokens, but faced consecutive failures. Hu Zheweng claimed to have invested in the blockchain project Stratis, achieving a return of "over 1000 times."According to disclosed information, Hu Zheweng invested approximately 80 million USD in algorithmic stablecoin TerraUSD (UST) and its sister token LUNA from May 2021 to May 2022, with the peak market value of his holdings exceeding 800 million USD, but the price of LUNA subsequently nearly dropped to zero.In addition, Hu Zheweng has filed a lawsuit in Chicago, USA against the globally renowned quantitative trading firm Jump Trading and its related companies and executives in the cryptocurrency business, seeking at least 500 million USD in damages. After experiencing Stox and Terra, Hu Zheweng has not left the cryptocurrency market; last January, he bet on a new project by a Chinese computer professor Zhang Yongfeng. Zhang Yongfeng entered the Computer Science Department of Tsinghua University in 2007 and is currently a tenured associate professor in the Computer Science Department at Rutgers University in the USA. He has been sued by Hu Zheweng on multiple charges including "securities fraud." It is reported that the tokens issued by the organization founded by Zhang Yongfeng, in which Hu Zheweng invested millions of dollars, have fallen over 99.6% from their peak in 2025.

According to Investment World citing Quantum Bit reports, DeepSeek has publicly disclosed the technical details of the system DSec (DeepSeek Elastic Compute) used for training Agents, authored by Liang Wenfeng. This system can generate over 5,000 sandboxes per second, reaching 3 million in a day, with a peak simultaneous operation of 380,000; supporting this scale is a single cluster with approximately 160 nodes, 30,000 CPU cores, and 250TB of memory.DSec prepares four types of backends for four categories of tasks: FnCall, Container, MicroVM, and Full VM, with the training side called through a unified Python SDK libdsec. The scheduling chain includes IAM, API Server, scheduling engine, node Edge, network proxy Aether, and components within the sandbox Chronus. The environment is divided into three layers of read-only images: base image, workspace, and toolkit, which are used in combination at startup. Runtime data from the paper shows that the actual read ratios of Python, Java, and C++ container images are approximately 6.0%, 9.2%, and 8.7%, respectively.Starting from DeepSeek-V4.1, the Agent loop has been moved to the DSec worker container, no longer bound to the GPU Pod lifecycle. The security section disclosed reward hacking during training, including actions such as overwriting system files, swapping file data blocks, scanning networks, and triggering kernel crashes. Defensive measures include AppArmor and eBPF-based network filtering, but reports indicate that these measures do not completely resolve the issues.

The Polish prosecutor's office has charged a former police officer in the case of the collapsed cryptocurrency exchange Zondacrypto. The Central Bureau of Cybercrime (CBZC) in Poland detained Artur K. on Tuesday under the orders of the organized crime department of the national prosecutor's office. He faces two charges: one for participating in an organized crime group, and the other for causing significant property losses to BitBay, BB Trade Estonia, and another entity, assisting in the misappropriation of entrusted property, aiding computer fraud, and money laundering. The prosecutor's office stated that a search of his residence uncovered documents related to the operation of the exchange, and he is being questioned as a suspect; the prosecutor's office will decide whether to apply for pre-trial detention.Meanwhile, the police in Katowice are searching and investigating a fuel depot, which is the last known location of Zondacrypto founder Sylwester Suszek before he went missing in March 2022. The case is classified as illegal deprivation of liberty rather than a disappearance, as a suspect has been accused of removing a hard drive from the gas station's surveillance camera.Suszek founded the exchange in Katowice in 2014, originally named BitBay, and relocated its registration to Estonia as BB Trade Estonia in 2020, rebranding it as Zondacrypto, claiming to have 1.3 million customers, most of whom are Polish. The exchange froze withdrawals and went offline in April, and the Estonian Financial Intelligence Unit revoked its operating license on June 29. Since August, at least five people have been charged, including former Polish Olympic Committee president Radosław P. All public respondents have denied the charges, and the related accusations have not yet been tried in court.

According to exclusive reports from "Investment World", Yan Wentao officially joined DeepSeek on Monday and reported to the company, taking on the role of the company's first CFO.Yan Wentao was born in 1991 and graduated from Fudan University. He joined Hillhouse Capital in 2020 and has previously participated in projects such as MiniMax, Zhipu AI, and ByteDance.DeepSeek is finalizing its second round of financing, planning to raise 50 billion yuan, with a pre-investment valuation of 500 billion yuan, and has already hired CITIC Securities to prepare for a potential Sci-Tech Innovation Board IPO.

According to LatePost, the Alibaba Cloud Shadow team is developing an AI device called QwenBook, positioned as a native intelligent agent computer, resembling the Qianwen tablet. The project started in April this year, with a team of about 150 people, integrating models including Qwen3.8-Max and Qwen3.8-Flash. The complete product is expected to be released by the end of this year or early next year.The report states that QwenBook is similar in concept to the Doubao phone, both attempting to push the model to the system's core. After the launch of the first generation Doubao phone, WeChat, Taobao, Meituan, Pinduoduo, and others successively prohibited it from replacing user operations through screen reading and simulated clicks. The Nubia NaviX Ultra, which went on sale on September 16, retains the GUI approach, attempting to allow applications to open services to the Agent through MCP, A2A, or dedicated interfaces.