# Сопутствующие статьи по теме Open Source

Новостной центр HTX предлагает последние статьи и углубленный анализ по "Open Source", охватывающие рыночные тренды, новости проектов, развитие технологий и политику регулирования в криптоиндустрии.

U.S. Lawmakers Push to Clarify Crypto Developer Liability Under Federal Law

U.S. Senators Cynthia Lummis and Ron Wyden have reintroduced the bipartisan Blockchain Regulatory Certainty Act to protect cryptocurrency developers from being wrongly classified as financial institutions or money transmitters. The bill clarifies that liability should be based on control of user funds, not merely writing or maintaining code. This addresses growing legal uncertainty and fear among developers, especially following recent Department of Justice cases involving privacy tools. Lawmakers argue that misapplying financial regulations to developers who don’t handle funds is unfair, stifles innovation, and could drive talent overseas. The legislation aims to safeguard open-source development and prevent overreach by ensuring only entities with actual control over assets face strict regulatory obligations.

TheNewsCrypto01/13 08:56

U.S. Lawmakers Push to Clarify Crypto Developer Liability Under Federal Law

TheNewsCrypto01/13 08:56

X Rolls Out Smart Cashtags to Cut Crypto Spam and Ticker Confusion

X is introducing a new feature called Smart Cashtags to reduce cryptocurrency ticker confusion and spam on its platform. The tool will allow users to link cashtags like $BTC to verified token identities or contract addresses, providing access to charts and information directly from posts. This aims to address issues where scam coins mimic legitimate cryptocurrencies and improve clarity in crypto discussions. The rollout, planned as early as next month, comes amid regulatory pressure in Europe, including a recent €120 million fine for Digital Services Act violations. Additionally, Elon Musk has promised to open-source X’s ranking algorithms to increase transparency. Crypto users have criticized the platform for bot dominance and engagement farming, which Smart Cashtags may partially mitigate by reducing ticker exploitation.

TheNewsCrypto01/12 13:46

X Rolls Out Smart Cashtags to Cut Crypto Spam and Ticker Confusion

TheNewsCrypto01/12 13:46

Jensen Huang Announces 8 New Products in 1.5 Hours, NVIDIA Fully Bets on AI Inference and Physical AI

NVIDIA CEO Jensen Huang unveiled eight major announcements during his CES 2026 keynote, focusing on advancing AI inference and physical AI technologies. The centerpiece was the NVIDIA Vera Rubin POD AI supercomputer, which integrates six custom chips—Vera CPU, Rubin GPU, NVLink 6 Switch, ConnectX-9 SuperNIC, BlueField-4 DPU, and Spectrum-X CPO—designed for协同 performance. The Rubin GPU offers 5x higher inference and 3.5x higher training performance than Blackwell, with support for HBM4 memory. The Vera Rubin NVL72 system delivers 3.6 EFLOPS in NVFP4 inference performance in a single rack, with enhanced memory bandwidth. NVIDIA also introduced the Spectrum-X Ethernet CPO for improved power efficiency, a推理上下文内存存储平台 to optimize KV cache storage and reduce recomputation, and the DGX SuperPOD based on Rubin architecture, cutting token costs for large MoE models to 1/10. On the software side, NVIDIA expanded its open-source offerings, including new models and datasets, and emphasized the rise of physical AI. The company open-sourced the Alpha-Mayo model for autonomous driving, enabling reasoning-based decision-making, and announced production-ready NVIDIA DRIVE platforms for Mercedes-Benz. Partnerships with Siemens and robotics firms like Boston Dynamics were highlighted, underscoring NVIDIA’s full-stack approach to AI infrastructure and real-world AI applications.

marsbit01/06 04:36

Jensen Huang Announces 8 New Products in 1.5 Hours, NVIDIA Fully Bets on AI Inference and Physical AI

marsbit01/06 04:36

Jensen Huang's CES 2026 Keynote: Three Key Themes and a 'Chip Beast'

At CES 2026, NVIDIA CEO Jensen Huang unveiled the Rubin AI computing platform, a major leap in AI infrastructure designed to make AI "think longer" and operate more cost-efficiently. The Rubin architecture includes the Rubin GPU, Vera CPU, NVLink 6, and BlueField-4, working together to significantly reduce inference costs—up to 10x lower than Blackwell—and support longer context and multi-step reasoning. A key innovation is the Inference Context Memory Storage Platform, which uses BlueField-4 and Spectrum-X to manage AI context data at scale, improving token throughput and energy efficiency. NVIDIA also introduced the DGX SuperPOD, scaling the Rubin platform to 576 GPUs for large-scale AI clusters. Huang emphasized the shift toward Reasoning/Agentic AI and announced major updates to NVIDIA’s open-source ecosystem, including models like Nemotron and tools for RAG, safety, and speech. In physical AI, NVIDIA launched Cosmos for physics-aware video generation and prediction, and Alpamayo—an open-source vision-language-action model for autonomous driving, now in production with Mercedes-Benz. The event also highlighted growing adoption of NVIDIA’s robotics platform, Isaac GR00T, by companies like Boston Dynamics and LG, signaling broader commercialization of AI in real-world applications.

marsbit01/06 02:08

Jensen Huang's CES 2026 Keynote: Three Key Themes and a 'Chip Beast'

marsbit01/06 02:08

Racing to Be the First Stock: The Substance, Capabilities, and Ambition of China's Largest Independent Model Company

Zhipu AI, China's largest independent large language model (LLM) company by revenue, has passed its listing hearing on the Hong Kong Stock Exchange with a valuation of RMB 24.377 billion. Its IPO filing provides the first clear look at the financials of a major Chinese LLM player. From 2022 to 2024, Zhipu's revenue grew at a 130% CAGR, reaching RMB 310 million in 2024. Nearly 85% of its revenue comes from on-premise model deployments for enterprise clients, with the remainder from its MaaS (Model-as-a-Service) platform. Despite rapid revenue growth, the company reported significant adjusted net losses, driven overwhelmingly by R&D expenses which reached RMB 1.59 billion in H1 2025. A major portion of these costs is attributed to computing power, essential for training its flagship models. A key part of Zhipu's strategy is a "land and expand" approach: using strategic price cuts on its MaaS platform to attract a large user base (over 1.2 million enterprise developers) and then converting them into high-value on-premise clients. The release of its powerful open-source base model, GLM-4.5/4.6, which ranks among the top global models in several benchmarks, led to an exponential increase in API calls and token consumption. The company is betting that continued heavy R&D investment is necessary to stay at the forefront of the intensely competitive global AI market. Its leadership believes that possessing a superior base model is the ultimate product and the key to long-term growth, even if it requires substantial short-term losses. As one of the first Chinese LLM firms to file for an IPO, Zhipu's market debut is poised to be a major test for valuing China's independent AI industry.

marsbit12/23 11:13

Racing to Be the First Stock: The Substance, Capabilities, and Ambition of China's Largest Independent Model Company

marsbit12/23 11:13

Lighthouses Guide the Way, Torches Claim Sovereignty: A Hidden War Over AI Allocation Rights

The article "Lighthouse Guides Direction, Torch Fights for Sovereignty: A Hidden War Over AI Allocation" by Zhixiong Pan examines the underlying power struggle in AI development, moving beyond superficial metrics like model size and performance rankings. It identifies two coexisting paradigms: the "Lighthouse," representing state-of-the-art (SOTA), centralized AI systems controlled by tech giants like OpenAI and Google, which push cognitive boundaries but are resource-intensive and create dependency risks; and the "Torch," symbolizing open-source, locally deployable models (e.g., DeepSeek, Mistral) that democratize access, ensure data sovereignty, and enable private, customizable AI assets. The Lighthouse drives innovation and sets technical directions but poses risks in accessibility, control, and single-point failures. The Torch, while shifting security and responsibility to users, offers resilience, cost stability, and compliance for critical applications in sectors like healthcare and finance. The interplay between these models forms a symbiotic relationship: Lighthouses expand capabilities, while Torches disseminate and stabilize these advances, collectively elevating AI’s baseline. Ultimately, the conflict is over AI allocation rights—defining default intelligence, managing externalities, and determining individual control. A dual strategy—using Lighthouses for frontier tasks and Torches for private, reliable deployment—is proposed as the pragmatic path forward, balancing extreme capability with broad, sovereign access. The true measure of the AI era lies not in raw power but in whether individuals possess "a light they don’t have to borrow from anyone."

marsbit12/22 11:13

Lighthouses Guide the Way, Torches Claim Sovereignty: A Hidden War Over AI Allocation Rights

marsbit12/22 11:13

MoveBit Research Release | Belobog: A Move Fuzzing Framework for Real-World Attacks

MoveBit introduces Belobog, a novel fuzzing framework designed specifically for Move smart contracts to address security challenges beyond syntax and type errors. Unlike traditional fuzzing methods that struggle with Move’s strong type system and resource semantics, Belobog leverages type-guided input generation and mutation. It constructs a type graph to produce semantically valid and executable transaction calls, significantly improving the efficiency and depth of state exploration. The framework integrates concolic execution (combining concrete and symbolic execution) to penetrate complex constraints and branch conditions, enabling deeper coverage of potential vulnerabilities. Evaluated on 109 real-world Move contracts, Belobog detected 100% of Critical and 79% of Major vulnerabilities confirmed by manual audits. It also demonstrated the ability to reproduce full exploit paths without prior knowledge of vulnerabilities. Designed to be developer-friendly, Belobog will be released as open-source to encourage community adoption and extension. The work is currently under peer review for PLDI’26. Preprint: https://arxiv.org/abs/2512.02918

marsbit12/16 09:12

MoveBit Research Release | Belobog: A Move Fuzzing Framework for Real-World Attacks

marsbit12/16 09:12

When Token Issuance Becomes an Assembly Line, Some Are Paying Bitcoin Developers

The article highlights the often-overlooked efforts of organizations and individuals who support Bitcoin's core infrastructure, focusing on a report by 1A1z about Bitcoin Core developers and sponsors. Bitcoin Core, described as Bitcoin's "operating system," relies on volunteer work and external funding for critical tasks like security, network optimization, and code maintenance. The report identifies 13 key long-term sponsors, including OKX as the only exchange consistently funding developers since 2019, with nearly $2 million in contributions to individuals and nonprofits. Beyond core protocol support, the piece emphasizes the importance of user-facing infrastructure, such as OKX Wallet's features that simplify Web3 access by eliminating seed phrases and streamlining cross-chain transactions. It also notes OKX's investments in developer ecosystems, test networks, and security audits—efforts that don't generate immediate returns but are vital for sustainable growth. The article contrasts these long-term, low-profile investments with the industry's trend of rapid token launches and marketing-heavy projects. It argues that such foundational work, though less visible, builds competitive advantages through technical efficiency, user experience, and future capacity—ultimately shaping the industry's trajectory toward real-world adoption.

marsbit12/13 01:56

When Token Issuance Becomes an Assembly Line, Some Are Paying Bitcoin Developers

marsbit12/13 01:56

Seed Round Secures $500 Million, It Aims to Be the 'Apple' of Payment Blockchains

Stripe's Tempo blockchain, a fork of Ethereum optimized for fintech applications, is now live on testnet. Backed by a $500 million seed round from Stripe and Paradigm, it aims to become the "Apple" of payment blockchains through vertical integration and a closed-architecture approach. Key innovations include: TIP-20, a stablecoin standard with prioritized transaction lanes; low, predictable gas fees payable in any stablecoin (converted automatically via a built-in AMM); and native smart accounts enabling batch payments, gas sponsorship, scheduled transactions, and passkey authentication. Unlike Ethereum’s open, permissionless validator set, Tempo uses a private, permissioned consensus mechanism (Simplex BFT) and is fully EVM-compatible. Its design prioritizes user experience, security, and commercial adoption over decentralization, targeting the payments industry as its primary market. The code is open-sourced under Apache/MIT licenses.

比推12/11 18:34

Seed Round Secures $500 Million, It Aims to Be the 'Apple' of Payment Blockchains

比推12/11 18:34

Sentient Launches Open-Source AI Research Agent SERA-Crypto to Tackle the 'Hallucination Crisis' in the Crypto Industry

Sentient Labs, one of the largest open-source AI platforms with $85M in funding, has launched SERA-Crypto—an open-source AI research agent designed for crypto and Web3. It delivers on-chain analytics, tokenomics insights, and protocol risk assessments within 30 seconds, complete with verifiable citations and reproducible reasoning chains. SERA-Crypto addresses the reliability issues of general AI models in finance, particularly critical in crypto where inaccurate information can lead to significant financial loss. Unlike generic models, it uses real-time data (updated within 24–48 hours), parallel tool integration, and structured reasoning to ensure accuracy and depth. Key advantages include temporal relevance, deep research capabilities, data consistency, and support for use cases like wallet risk alerts, exchange integrations, and DAO proposal analysis. It integrates over 50 data endpoints and is built on Sentient’s SERA architecture, optimized for blockchain and DeFi. The model was trained on 220K+ users and 22M+ queries and will be open-sourced in phases. Sentient is backed by Founders Fund, Pantera Capital, and Framework Ventures.

深潮12/11 08:26

Sentient Launches Open-Source AI Research Agent SERA-Crypto to Tackle the 'Hallucination Crisis' in the Crypto Industry