Qihoo 360 Unveils China's AI Cybersecurity Solution Amid US Restrictions
Qihoo 360 founder Zhou Hongyi introduced "Tulong Feng," an AI-based cybersecurity tool in response to U.S. restrictions on Anthropic’s Mythos, according to Decrypt. The tool reportedly excels in detecting vulnerabilities, aiming to position China as a competitor in AI-driven cybersecurity solutions.
At the ISC.AI 2026 conference held in Beijing on June 24, Qihoo 360's founder, Zhou Hongyi, presented "Tulong Feng," a new AI vulnerability detection tool aimed at bolstering China’s cybersecurity capabilities. Zhou claims that Tulong Feng offers comparable functionalities to Anthropic's proprietary tool, Mythos, which is currently restricted for use in the U.S. market due to export controls. Zhou states that Tulong Feng has successfully detected 3,432 vulnerabilities, with 105 of these confirmed by Chinese regulatory authorities.
Zhou's unveiling comes during a turbulent time for Anthropic, which is in negotiations with the U.S. Department of Commerce over regaining access to its AI models, including Mythos. He described the importance of developing autonomous systems for cybersecurity, likening such systems to "cyber nuclear weapons" in their ability to identify vulnerabilities and execute attacks without human involvement. He stated, "America has Mythos; China also has its own 'Heaven-Sword Dragon-Saber.'" In light of the restrictions placed on Chinese firms regarding access to tools like Mythos, Zhou emphasized the need for domestically developed solutions.
Alongside Tulong Feng, Qihoo 360 introduced Yitian Zhen, an automated defense platform, and announced the formation of a new domestic security coalition termed "Panshi Zhidun," translating to Shield of Bedrock. Zhou criticized the exclusion of Chinese companies from collaboration programs such as Anthropic's Glasswing, which includes prominent partners like Microsoft and Apple. He articulated that providing tools like Tulong Feng to local firms not only fills a significant technological gap but also empowers China’s position in the global cybersecurity landscape.
Meanwhile, Z.ai, a Beijing-based competitor, made a significant move by releasing its own version, GLM-5.2, freely available for download. This open-source version aims to provide capabilities similar to what is offered by Anthropic's tools, highlighting the current competition in the AI cybersecurity field.
As the landscape evolves, observers are keen to see how these developments will shape the future of AI in cybersecurity and how incidents of vulnerability detection and mitigation manifest in practical applications.
Summary based on original reporting by Jose Antonio Lanz at Decrypt, originally published Jun 29, 2026. SolanaWire does not republish source content.
DeepReinforce Launches Ornith, Open-Source AI Coding Model for Agents
DeepReinforce unveiled Ornith-1.0 on June 25, 2026, an open-source coding model designed specifically for AI agents rather than human users, as reported by Decrypt. The model is available in various sizes and excels in automating coding tasks within development workflows.
25 minutes ago·Decrypt·Reported by Jose Antonio Lanz
Meta Introduces Brain2Qwerty AI for Decoding Brain Activity into Text
Meta has launched Brain2Qwerty v2, a non-invasive AI system that translates brain activity into text, achieving 61% average word accuracy compared to 8% for previous methods. The initiative aims to assist individuals with communication impairments and is part of Meta’s broader Digital Brain Project, according to Decrypt.
3 hours ago·Decrypt·Reported by Jason Nelson
Bitcoin Struggles Around $60K Amid Bearish Signals
Bitcoin hovers around $60,104 while facing negative price indicators, according to Decrypt. Key resistance levels and bearish charts suggest potential for further declines, despite some bullish activity at critical support levels.
3 hours ago·Decrypt·Reported by Jose Antonio Lanz
Silicon Valley Bank Reports Shift in Bitcoin Lending to Institutional Era
Silicon Valley Bank indicates that bitcoin lending is undergoing a transformation towards increased institutional participation, stronger risk controls, and lower borrowing costs. The report highlights the rise of overcollateralization and transparent risk management since the 2022 crypto credit crisis, and notes that total crypto-backed lending has reached $67 billion, according to CoinDesk.
4 hours ago·CoinDesk·Reported by Will Canny
Trending this week
Supreme Court Ruling Allows Trump to Dismiss SEC, CFTC Commissioners
The U.S. Supreme Court ruled that President Donald Trump can fire commissioners of key regulatory agencies like the SEC and CFTC at will. This decision, which overturns nearly a century of precedent, could influence ongoing negotiations on crypto regulation, according to Decrypt.
2 hours ago·Decrypt·Reported by Sander Lutz
Ripple Developers Testing New Lending Protocol for XRP Ledger
Ripple has initiated a testing phase for its XRPL Lending Protocol, which could enable on-chain borrowing and lending of digital assets. The proposed upgrades, if approved, aim to enhance liquidity by allowing previously idle assets to be utilized, as reported by Decrypt.
4 hours ago·Decrypt·Reported by André Beganski
Galaxy Cuts CLARITY Act Passage Odds to 50% Amid Senate Time Crunch
Galaxy Digital has lowered its odds of the CLARITY Act passing in 2026 to 50%, down from 60% earlier this month, according to Decrypt. This reduction comes as the Senate runs out of time before its August recess, facing competing legislative priorities.
4 hours ago·Decrypt·Reported by Decrypt Agent
JPMorgan Supports U.S. Crypto Legislation While Highlighting Risks
JPMorgan endorses the U.S. Digital Asset Market Clarity Act but emphasizes the need for effective safeguards, as reported by CoinDesk. The bank warns that the legislation must address existing regulatory gaps to avoid new risks in the digital asset landscape.
4 hours ago·CoinDesk·Reported by Helene Braun