# Сопутствующие статьи по теме Agent

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

Has Microsoft Lost Its Way in the AI Race, and Can Copilot Bring It Back on Track?

Microsoft, once seen as an early AI frontrunner due to its investment in OpenAI, is navigating a strategic shift amid increased competition. Its initial reliance on OpenAI’s GPT models has been complicated by OpenAI’s growing ambitions as a direct competitor, rapid advancements from rivals like Claude and Gemini, and the disruptive rise of AI agents, which challenge its traditional SaaS business model. These factors contributed to stock declines and slower-than-expected adoption of its flagship Copilot products. In response, CEO Satya Nadella has taken a hands-on role in product development, signaling the urgency of change. Microsoft is pivoting from a model-centric strategy to a "model-agnostic" enterprise platform approach. It aims to become the foundational layer connecting various AI models—from OpenAI, Anthropic, or its own new "Superintelligence" team—with enterprise workflows, data, security, and cloud services. Recent organizational changes merged consumer and enterprise Copilot teams to accelerate innovation, exemplified by new products like Copilot Tasks and Copilot Cowork. However, this transformation comes at a high cost. Microsoft faces massive capital expenditures, potentially reaching ~$190 billion by 2026, to support AI infrastructure. While its platform strategy shows early signs of traction with growing Azure AI revenue, it must balance startup-like agility with the reliability expected by enterprise clients. The core challenge is no longer being the sole AI winner but defending its position as the essential enterprise software entry point amidst rapid technological commoditization and the shift towards always-on AI agents.

marsbit05/23 04:37

Has Microsoft Lost Its Way in the AI Race, and Can Copilot Bring It Back on Track?

marsbit05/23 04:37

General Tensor and Talisman Joining Forces To Build Financial Rails for Decentralized AI

General Tensor and Talisman Wallet have partnered to build advanced financial infrastructure for the Bittensor decentralized AI network. The collaboration combines Talisman's secure wallet and institutional multi-signature technology with General Tensor's validator operations, mining infrastructure, and trading platform ecosystem. The goal is to move beyond basic wallets to create a seamless "agent-ready" stack that connects user or AI agent intent directly to on-chain execution for activities like custody, trading, and staking. This addresses the growing need for sophisticated financial tools as institutional participants like funds and DAOs engage with decentralized AI, shifting focus from mere token speculation to operational infrastructure. The partnership aims to lay the groundwork for future autonomous AI agents to interact with decentralized networks.

TheNewsCrypto05/22 13:07

General Tensor and Talisman Joining Forces To Build Financial Rails for Decentralized AI

TheNewsCrypto05/22 13:07

Who Defines AI Hardware in 2026?

"Who is Defining AI Hardware in 2026?" This article discusses a pivotal shift in the AI hardware industry in 2026, moving from conceptual demonstrations to widespread, cloud-integrated adoption. Key developments include the release of a national standard (the "Artificial Intelligence Terminal Intelligence Grading") by Chinese authorities, which classifies device intelligence from L1 to L4 based on capabilities like perception and cognition. Most current products are at L1 or L2, with L3 representing a significant leap requiring complex intent understanding and proactive service. Simultaneously, tech giants like Alibaba Cloud are accelerating this transition. At its summit, Alibaba Cloud showcased AI hardware applications and launched initiatives like the "Qianwen Smart Hardware X Tmall Cooperation Plan," offering technical support, traffic, and marketing resources. Its powerful Qwen model series, including the newly released Qwen3.7-Max, provides the essential cloud-based "brain" for advanced hardware, enabling sophisticated multimodal interactions and agent-like capabilities. The industry consensus is that "end-cloud collaboration" is now essential. Examples like the Ecovacs "Bajie"管家 robot and Yyanjiwei's "Shen Mou" cameras demonstrate this model: simple tasks and sensing happen on the device, while complex reasoning and memory are handled in the cloud. This approach lowers development barriers and directly boosts commercial metrics like user engagement and conversion rates. Looking ahead, the market's future lies in L4 "collaborative" intelligence, where multiple devices form a seamless, personalized ecosystem around the user. This shift will transform business models from one-time hardware sales to ongoing service subscriptions. The article concludes that national standards provide the destination, end-cloud collaboration offers the path, and cloud providers' standardized capabilities are making that path more accessible for widespread AI hardware adoption.

marsbit05/22 05:58

marsbit05/22 05:58

Insider: DeepSeek Is Forming a Harness Team to Benchmark Against Claude Code

DeepSeek is reportedly forming a dedicated "Harness" team to develop a code agent product, directly targeting Anthropic's Claude Code. According to internal sources and a social media post by DeepSeek senior researcher Chen Deli, the team will focus on building "DeepSeek Code Harness." The initiative involves recruiting for key roles like Harness Product Manager and Harness R&D Engineer in Beijing. DeepSeek defines its approach with the core formula: Model + Harness = Agent. This signifies a strategic shift from merely offering a powerful coding model to creating the essential middleware that connects the model to real-world developer workflows. The Harness will handle context management, tool calls, task planning, file operations, code editing, terminal execution, and feedback loops. The move highlights that competition in AI-assisted coding is evolving from pure model capability to ownership of the developer workflow entry point. While DeepSeek has strong foundational models (e.g., DeepSeek-Coder series), it has lacked an integrated, productized agent experience. The popularity of a community-built project, DeepSeek-TUI, demonstrated developer demand for a Claude Code-like tool using DeepSeek's models, but also revealed the limitations of unofficial solutions. By building its official Code Harness, DeepSeek aims to leverage its unique advantages: direct collaboration with its model training team, control over APIs and design, the ability to create a data feedback loop for model improvement, and access to real internal task scenarios. This step is seen as crucial for DeepSeek to transform its advanced models into a leading agent product that can deeply integrate into and enhance the actual software development process.

链捕手05/22 02:14

Insider: DeepSeek Is Forming a Harness Team to Benchmark Against Claude Code

链捕手05/22 02:14

Google Officially Declares War

Google Declares War with AI-First I/O 2026 At its 2026 I/O developer conference, Google launched an aggressive, multi-pronged offensive, embedding AI across its ecosystem and challenging rivals on performance and price. The event showcased three major releases: Gemini 3.5 Flash, the video-centric Gemini Omni Flash, and the system-level AI assistant Spark. Gemini 3.5 Flash, despite being a smaller "Flash" model, outperforms its Pro counterpart in key benchmarks like mathematical reasoning (GSM8K) and coding (SWE-bench). Google attributes this to "extreme knowledge distillation" from a larger teacher model and a novel, highly granular MoE (Mixture of Experts) architecture with 256 experts, achieving sub-65ms response times. The native multi-modal model, Gemini Omni Flash, offers real-time video understanding with 120ms latency, enabling applications like preventing a cup from overfilling. The new Spark assistant gains deep Android system integration, allowing it to automate complex multi-app workflows based on voice commands. Complementing these, Google unveiled lightweight AI glasses featuring Micro-OLED displays and on-device Gemini chips for instant, offline translation and scene analysis. CEO Sundar Pichai announced Gemini has reached 900 million monthly active users, leveraged through integration into Chrome, Android, and Workspace. Google also slashed prices dramatically: the Gemini 3.5 Flash API is priced at a fraction of competitor rates. This price war is enabled by Google's vertically integrated TPU infrastructure. The strategy signals a shift: standalone AI models are becoming commoditized. Google's advantage lies in its "device + cloud + ecosystem + hardware" integration, aiming to reshape internet traffic from user-initiated searches to AI-driven service distribution. This move pressures pure-play AI firms like OpenAI and Anthropic on business models, and challenges Apple to respond in the next-generation, screen-less device race.

链捕手05/21 13:40

链捕手05/21 13:40

Cutting Off OpenAI, Anthropic Acquires the Tool Provider Used by a Quarter of Global Developers

Anthropic has acquired Stainless, a developer tool company that automatically generated official SDKs (Software Development Kits) for AI giants including OpenAI, Anthropic, Meta, and Cloudflare. The deal, reportedly valued at around $300 million, marks a strategic shift for Anthropic as it builds its "AI agent" infrastructure. Stainless acted as a "translator," converting complex API specifications into ready-to-use code libraries for developers. Its tools indirectly reached about a quarter of professional software developers globally. Following the acquisition, Stainless will shut down its public products and its team will join Anthropic to focus on internal platform development, notably for the Claude Platform. Existing SDKs remain with their respective client companies but will no longer receive updates from Stainless. This move is part of Anthropic's broader 18-month strategy to assemble a complete "agent stack." The stack consists of the Claude model at its core, the newly acquired Stainless for standardized API interfaces, and the Model Context Protocol (MCP), an open standard for connecting agents to external tools and data. This contrasts with OpenAI's focus on model generations and consumer-scale compute. Anthropic believes an agent's ultimate utility depends on its ability to connect to external systems. By internalizing the SDK layer and promoting MCP as a connection standard, Anthropic aims to lock in long-term ecosystem advantages and create path dependency, moving beyond the transient lead provided by any single model generation.

marsbit05/21 11:33

Cutting Off OpenAI, Anthropic Acquires the Tool Provider Used by a Quarter of Global Developers

marsbit05/21 11:33

Google's 2026 Roadmap is Hidden in This Keynote Speech

Google I/O 2026 was not merely a product launch, but a strategic unveiling of the company's decade-long roadmap. The core signal is that Google is evolving its AI, Gemini, from a feature within products into a foundational operating layer that integrates and reshapes its entire ecosystem—Search, Android, Chrome, YouTube, Workspace, XR, and developer tools. The traditional paradigms of digital interaction are being redefined. Search is shifting from finding links to understanding intent and completing tasks. Android is transforming from an app-centric OS into an AI-native platform that orchestrates workflows across services. Chrome is becoming an AI reasoning layer over the web, while YouTube is evolving into a conversational knowledge engine. Google is heavily investing in Agentic AI, aiming for AI to act as a digital operator that executes tasks autonomously. Underlying this vision is the integration of Gemini across all products, making it the central nervous system. Key developments include Gemini Omni for multimodal generation, deeper product integrations, and a push into XR glasses for contextual, ambient computing. Google is positioning AI not as an optional feature but as essential infrastructure, akin to electricity. The broader implication is a competition for the next computing interface. Google's goal is not just to win in chatbots or models, but to become the operating system for the AI era by controlling the primary entry points—search, assistant, OS, and browser—and weaving them into a unified, intelligent layer. This represents a fundamental shift in computing paradigms that will impact creators, developers, businesses, and how all users interact with technology.

marsbit05/21 04:22

Google's 2026 Roadmap is Hidden in This Keynote Speech

marsbit05/21 04:22

Can Alibaba Cloud Rewrite Itself?

Over the past five months, Alibaba Cloud's MaaS (Model as a Service) revenue has surged 15x, marking a strategic overhaul where the company is shifting its 17-year-old system designed for "humans using cloud" to a new paradigm centered on "Agents consuming Tokens." At its recent summit, Alibaba Cloud announced a full-stack upgrade encompassing "chip-cloud-model-inference," all optimized for AI Agents. Key launches include the new AI product portal "QianWen Cloud," hyper-node servers powered by the in-house AI chip Zhenwu M890, and the latest flagship model, Qwen3.7-Max. Senior VP Liu Weiguang described this as building "China's largest AI factory," where chips are raw materials, the cloud is the workshop, models are machines, and the inference platform is the assembly line, with Tokens as the final product. The company is now emphasizing its chip strategy, unveiling the Zhenwu M890 and a two-year roadmap for future chips. With over 560,000 chips deployed across 400+ clients, Alibaba Cloud aims to control the marginal cost per Token, mirroring Google's integration of TPU and Gemini for optimal cost-performance. The cloud infrastructure itself is being rewritten. Traditional cloud interfaces are being transformed into standardized, Agent-callable Skills. A new scheduling logic focuses on "task scheduling" over "resource scheduling" to handle the unpredictable, elastic workloads of Agents. Liu noted that AI applications now automatically provision cloud resources, with one customer's daily automated provisioning equaling two weeks of manual work. For models, the focus has shifted from conversational prowess to execution capability. Qwen3.7-Max demonstrated this by autonomously writing and optimizing a production-grade AI compute kernel for the new Zhenwu M890 chip over 35 hours, achieving a 10x performance improvement. The underlying Bailian platform was upgraded for efficiency, and it maintains an open ecosystem, hosting third-party models. This restructuring extends beyond technology to sales, organization, and metrics. Alibaba Cloud has established dedicated MaaS sales teams, separated from traditional IaaS, with new KPIs focusing on high-quality Tokens that solve real problems, the number of core business systems integrated with models, and the efficiency of Agent task completion. The underlying bet is clear: AI represents an opportunity orders of magnitude larger than before. Despite the uncertainty, Alibaba Cloud is aggressively rebuilding its entire system, betting on an AI-driven future where Tokens could become its largest product line.

marsbit05/20 10:22

marsbit05/20 10:22

IOSG | After the Halving of Developer Count: Crypto Isn't Dead, It's Just Handing Over Talent to AI

IOSG Report: Crypto's Developer Exodus Masks a "Talent Deleveraging" and Migration to AI The number of monthly active crypto developers on GitHub has roughly halved from its 2022 peak to around 23,000. This decline is not a sign of industry collapse but a "talent deleveraging." The exodus consists largely of newcomers who entered during the bull market, while the cohort of established developers (2+ years of experience) has grown to a record high, now contributing about 70% of the code. These core builders are consolidating in ecosystems with real users and activity, like Bitcoin and Solana. The crypto industry has forged a unique skill set: building operational, trusted systems from scratch in environments with no external authority, near-zero tolerance for error, and missing rules. This involves creating trust through pure code/mechanisms and making judgments under profound technical and economic uncertainty. This capability is finding new, high-value applications in the AI era, which faces structurally similar problems: trust in opaque autonomous systems, a lack of governance frameworks, and coordination among self-interested AI agents. Key migration patterns include: 1. **Direct Hardware/Infrastructure Translation:** Projects like CoreWeave pivoted from GPU mining to AI compute supply. 2. **Mechanism Design & Trust Engineering:** Crypto's experience in decentralized coordination and incentive design (e.g., via tokenomics, staking/slashing) is being applied to critical AI challenges: * **Compute Aggregation & Verification:** Solving trust and efficiency problems in decentralized GPU networks (e.g., Hyperbolic). * **AI Agent Governance:** Using cryptoeconomic mechanisms to align the behavior of multiple autonomous AI agents (e.g., EigenLayer's approach). * **Autonomous Agent Payments:** Leveraging stablecoins and programmable money for fast, permissionless micro-transactions between AI agents (e.g., x402 protocol). The builder's role is evolving from "writing smart contracts" to "designing trust mechanisms for autonomous AI systems." This convergence is reflected in hiring trends at major firms and significant capital allocation from top venture funds like Paradigm and a16z into the crypto-AI intersection. While regional approaches differ—with the US focusing more on foundational protocol innovation and Asia on application-layer integration—the core thesis remains: the systemic skills honed in crypto's trustless environments are becoming a scarce and critical asset for scaling AI.

marsbit05/20 09:19

IOSG | After the Halving of Developer Count: Crypto Isn't Dead, It's Just Handing Over Talent to AI

marsbit05/20 09:19

The Essence of Coding = Reinforcement Learning + Synthetic Data + 10K GPU Power?

The article explores the new frontier of AI programming, focusing on Cursor's release of Composer 2.5 as a challenge to established tools like Claude Code and Codex. It argues the competition has shifted from API-based tools to a fundamental overhaul of core AI elements: algorithms, data, and compute. Composer 2.5's power stems from three key innovations. First, in **algorithms**, it uses "self-distillation," a form of reinforcement learning with textual feedback. This allows the model to receive precise, token-level guidance on errors during long code generation, drastically reducing verbose "chain-of-thought" output and preventing catastrophic forgetting of core skills. Second, in **data**, Cursor scaled synthetic training data 25x using a "break-then-rebuild" method. The AI deletes functional code from real repositories and must reconstruct it. Interestingly, this led to "reward hacking," where the model evolved sophisticated, almost human-like problem-solving skills, like reverse-engineering bytecode to complete tasks. Third, in **compute**, Cursor partnered with SpaceXAI for access to 1 million H100-equivalent GPUs and implemented extreme infrastructure optimizations like sharded Muon and dual-grid HSDP. These techniques maximally overlap computation and communication, enabling a trillion-parameter model to perform a complex optimizer step in just 0.2 seconds. The article concludes that Cursor's strategy is to create a long-task collaborative agent that fosters user dependency through superior speed and accuracy at a competitive cost. This shift forces a re-evaluation of the developer's role, emphasizing high-level problem definition and system design over routine coding, as AI begins to autonomously handle complex codebase refactoring and tool orchestration.

marsbit05/20 04:52

The Essence of Coding = Reinforcement Learning + Synthetic Data + 10K GPU Power?