TVL反超dYdX，盘点Hyperliquid生态新项目

长文源：区块律动Опубліковано о 2015-08-24Востаннє оновлено о 2024-08-15

Пов'язані матеріали

UNDP and DFINITY to Collaborate on AI Infrastructure Development

The United Nations Development Programme (UNDP) and the DFINITY Foundation have announced a partnership focused on sovereign cloud infrastructure and decentralized artificial intelligence (AI). They will explore how these technologies can be applied in government projects and digital transformation for public authorities, civil society organizations, and UNDP initiatives. DFINITY provides decentralized cloud and blockchain infrastructure, while UNDP focuses on sustainable development and public sector support. The collaboration will progress along three main tracks. Through the SDG Blockchain Accelerator, UNDP country offices will test sovereign cloud infrastructure, AI development tools, and decentralized AI applications for development projects. The Future of Work Academy will assist civil society organizations in building and testing their own applications. The Government Blockchain Academy will work with governments to develop pilot solutions aligned with local and national priorities. The projects will utilize DFINITY's Cloud Engines and Caffeine AI platform, enabling organizations to build and maintain applications in a secure digital environment managed at the local level. The partners will also examine the technologies' potential for protecting sensitive data, enhancing digital infrastructure resilience, and fostering local technical expertise.

cryptonews.ru7 хв тому

UNDP and DFINITY to Collaborate on AI Infrastructure Development

cryptonews.ru7 хв тому

Algorand Introduces Protocol for Managing AI Agents

The Algorand Foundation has introduced AC2, an open protocol designed to enable secure user interaction with and control over AI agents. The protocol is blockchain-agnostic, establishing an encrypted direct connection between a user's wallet or application and an AI agent. AC2 allows agents to request user permission for various actions, such as executing payments, signing code, or accessing APIs. Users review and approve or deny these requests directly within their wallet interface, eliminating the need to share private keys with the AI. The protocol supports multiple operational scenarios. For instance, a software agent can seek approval before signing, while a payment agent can send transaction details to a user's wallet for verification. Shopping agents can operate within pre-set user-defined limits on permissible items and spending amounts. For messaging, AC2 utilizes DIDComm v2.0, and for authentication, it employs passkeys via Liquid Auth, which is based on FIDO2 and WebAuthn standards. The protocol supports text and voice interaction, signature delegation, and can function without a centralized intermediary or even a blockchain.

cryptonews.ru9 хв тому

Algorand Introduces Protocol for Managing AI Agents

cryptonews.ru9 хв тому

U.S. Bonds, AI, Inflation: You Can't Have All Three – What's Bitcoin Betting On?

Title: U.S. Treasury Bonds, AI, and Inflation: A Trilemma Facing Washington, and Where BTC Fits In Core Thesis: The U.S. government faces a trilemma: stabilizing the Treasury market, sustaining the AI investment cycle, and controlling inflation. The likely political choice is to prioritize the first two, accepting higher inflation for longer. This creates a sustained tailwind for both gold and Bitcoin by injecting liquidity and shifting duration risk away from the private sector. Key Pressures on Treasuries: * Long-term yields (e.g., 10-year ~4.70%) are near multi-decade highs due to persistent inflation risks, high fiscal supply, weaker demand for long duration, and new competition from AI infrastructure financing. * AI investment, requiring massive capital (estimated ~$300B in investment-grade bond issuance), is now competing directly with the Treasury in long-duration debt markets. * Geopolitical tensions (e.g., Iran conflict) add upward pressure on oil prices and defense spending. Policy Response: The Treasury, under Secretary Besant, is focusing on market stability through measures like supporting the Yen (to prevent forced Japanese Treasury sales), increasing repo operations for long-dated bonds, and potentially shifting issuance toward the short end. This "aggressive Treasury issuance" can function like "stealth QE," easing financial conditions. Impact on Assets: * **Gold** has already benefited (up ~90% from Aug 2024 to Aug 2026), driven by declining trust in the USD, inflation fears, and a "monetary debasement" narrative. * **Bitcoin** is now showing stronger correlation with this theme. Recent outperformance vs. gold (BTC +22.2% vs. Gold +5.9% in late Aug) coincided with Treasury's enhanced repo actions, suggesting the market may be pricing it as a liquidity/ debasement hedge, not just a speculative tech asset. Conclusion: If high inflation, large deficits, and AI capital demands persist, the cost of stabilizing bonds and growth will be shorter debt maturity, constant liquidity support, and tolerance for higher inflation—a favorable environment for gold and BTC. This view would be challenged only by a return to ~2% inflation, credible fiscal consolidation, or AI becoming self-funding.

marsbit19 хв тому

U.S. Bonds, AI, Inflation: You Can't Have All Three – What's Bitcoin Betting On?

marsbit19 хв тому

HBM is Being Redefined

HBM is being redefined as AI-driven demand for computing power shifts the performance bottleneck from processing units to memory bandwidth, making High Bandwidth Memory the critical component for advanced AI hardware. At HotChips 2026, Samsung and SK Hynix unveiled new technical roadmaps for HBM4, moving away from the traditional approach of simply increasing DRAM speed and stacking layers. The focus for HBM4 has shifted from DRAM optimization to a comprehensive system-level redesign of the Base Die. Key innovations include upgrading the Base Die to advanced logic processes (e.g., Samsung's 4nm), adopting hybrid bonding for 3D stacking, and exploring heterogeneous packaging solutions like Intel's EMIB alongside the established CoWoS. This marks HBM's evolution from a simple memory component into an integrated system involving "memory + logic + packaging + IP + EDA." This transition brings benefits like better signal integrity, higher I/O bandwidth (up to 2048-bit), and potential for in-memory computing. However, it also introduces significant complexity. DRAM manufacturers now must deeply collaborate with logic foundries, extending design cycles and increasing supply chain challenges. Furthermore, the industry is moving towards multiple advanced packaging paths, with EMIB emerging as an alternative to CoWoS to alleviate capacity constraints, while hybrid bonding is being developed for future ultra-high (16+ layer) stacks. New core barriers have emerged beyond manufacturing. High-speed PHY/SERDES IP, crucial for HBM4's higher data rates, is concentrated among a few established players, creating an entry barrier for smaller firms. Similarly, the industry lacks mature 3D heterogeneous EDA toolchains capable of co-simulating the coupled electrical, thermal, and mechanical effects across the logic base, memory stacks, and interposer. These "soft" capabilities in IP and EDA are now critical determinants of HBM4's performance and yield. In summary, the HBM industry is entering a new era of systemic competition. The standardized approach is giving way to divergent vendor roadmaps, with multiple technologies coexisting. Success will depend on integrated architecture, packaging, process, and ecosystem capabilities, fundamentally reshaping the competitive landscape for AI memory.

marsbit24 хв тому

The Era of Earning 200,000 from a 3,000 Investment Is Over; AI Comic Dramas Are Not a Get-Rich-Quick Trend

The era of turning 3,000 yuan into 200,000 yuan is over. AI-generated comic dramas are no longer a low-barrier gold rush. A wave of sudden account suspensions has hit creators in AI comic dramas, AI art, and AI tutorials, primarily triggered by new content and copyright compliance rules implemented by Tomato Novel on August 17. This "last straw" exposed the underlying issues: tightening platform rules, rising copyright requirements, increasing costs, and severe content homogenization. The business model was fundamentally unsustainable for most creators. One creator detailed the math: producing five minutes of content daily cost around 510 yuan in AI computing and editing, while a video with 60,000 views might only earn 45 yuan. While a few early viral hits like "Huo Qubing" and "Zombie Cleaner" achieved high returns with minimal investment, they were exceptions. The reality for most solo creators was barely covering costs or operating at a loss. The new regulations mandate formal authorization from original authors for adaptations, limit video length and the amount of原著 content used, and ban terms like "Episode 1" in titles. This directly dismantled the previous "produce first, authorize later" model. Faced with the inability to afford版权 costs and the threat of penalty, many creators chose to halt updates. While some fans expressed frustration, the rules protect original authors who often faced侵权 with little recourse. The shutdowns signal not the failure of AI technology, but an inevitable industry shift from wild growth to regulated refinement. The previous model prioritized quick cash over quality content, leading to formulaic, "AI-face" saturated dramas that audiences grew tired of. The future belongs to creators who use AI as a tool to serve solid storytelling, not replace it. Some are now seeking formal licenses, a slower but necessary process. Upcoming national regulations further underscore this move towards standardization, weeding out those seeking only fast money and making space for genuine creators.

marsbit24 хв тому

The Era of Earning 200,000 from a 3,000 Investment Is Over; AI Comic Dramas Are Not a Get-Rich-Quick Trend

marsbit24 хв тому

Торгівля

Спот