# Сопутствующие статьи по теме AI

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

Intel Soars 20%, CPUs Return to Center Stage in the Agent Era

Intel's stock surged 20% after reporting exceptional Q1 2026 results, with revenue of $13.6 billion (up 7% YoY) and non-GAAP EPS of $0.29, beating expectations by 29x. The rebound is driven by the resurgence of CPUs in the AI era, particularly as workloads shift from training to inference and agent-based applications. Intel’s Data Center and AI (DCAI) division hit a record $5.1 billion in revenue, up 22% YoY, marking a U-shaped recovery since mid-2025. The growth is attributed to strong demand for Xeon 6 "Granite Rapids" processors and increased AI infrastructure refresh cycles. While NVIDIA and AMD dominated the AI training phase, Intel is benefiting from the focus on AI agents, where CPU performance becomes critical—accounting for 50-90% of workflow latency in agent orchestration. This shift, coupled with management changes and strategic refocus on CPUs (including canceling the Falcon Shores GPU project), has repositioned Intel. New CEO Lip-Bu Tan emphasized that CPUs are re-establishing themselves as essential infrastructure in the AI era.

marsbit04/24 04:39

Intel Soars 20%, CPUs Return to Center Stage in the Agent Era

marsbit04/24 04:39

DeepSeek V4 Finally Released, Breaking the Strongest Closed-Source Monopoly, Explicitly Partnering with Huawei Chips

DeepSeek AI has officially released DeepSeek-V4, available in two versions: the high-performance **DeepSeek-V4-Pro** (49B activated parameters, 1.6T total) and the more efficient **DeepSeek-V4-Flash** (13B activated parameters, 284B total). Both support a 1M context length, making long-context capability a baseline feature rather than a premium offering. The Pro version rivals top closed-source models in agent capabilities, world knowledge, and reasoning performance. It outperforms Claude Sonnet 4.5 in agentic coding and approaches Claude Opus 4.6 (non-thinking mode) in quality. The Flash version offers competitive performance at a lower cost, though it lags in highly complex tasks. A key technical innovation is a new attention mechanism that reduces computational and memory demands for long contexts. The models are optimized for agent frameworks like Claude Code and OpenClaw. API services are available with support for both OpenAI and Anthropic-style interfaces. DeepSeek also announced upcoming support for Huawei’s computing hardware in the second half of the year. The models are open-sourced on Hugging Face and ModelScope.

marsbit04/24 04:21

DeepSeek V4 Finally Released, Breaking the Strongest Closed-Source Monopoly, Explicitly Partnering with Huawei Chips

marsbit04/24 04:21

More and More 'Model Supermarkets' Are Opening: ByteDance, Alibaba, and Tencent Compete to Integrate

Chinese tech giants like ByteDance, Alibaba, and Tencent are accelerating the rollout of integrated AI model subscription services—dubbed “model supermarkets”—to provide developers with bundled access to multiple leading domestic large language models (LLMs). ByteDance’s Volcengine recently upgraded its "Coding Plan" by adding newer models like GLM-5.1, Minimax M2.7, and Kimi k2.6, allowing subscribers to use various top models under a single monthly fee starting at ¥40. However, user feedback reveals significant issues, including rapid consumption of usage limits (e.g., hitting caps within hours), frequent server errors (like HTTP 429), and slow response times during peak hours. Complaints about misleading deduction rates—where calls to advanced models consume more quota—are also common. The trend is industry-wide: Alibaba, Tencent, and Baidu have all launched similar multi-model coding plans. While these platforms reduce trial costs for developers, they also expose challenges in balancing affordability with service quality and computational stability. Amid this shift, independent AI companies like Zhipu, MiniMax, and Moonlight Face (Kimi) are developing strategies to avoid becoming mere “pipes” in this ecosystem—focusing on vertical applications, autonomous agents, and long-context models to retain competitiveness. Analysts suggest that, while platform aggregation may pressure model firms in the short term, specialized and vertical AI capabilities will remain differentiated in the long run.

marsbit04/24 04:07

More and More 'Model Supermarkets' Are Opening: ByteDance, Alibaba, and Tencent Compete to Integrate

marsbit04/24 04:07

OpenAI Launches Workflow Agent, GPTs Begin Countdown

OpenAI has launched Workspace Agents, an evolution of GPTs designed for teams. This new feature allows users to create reusable, shareable agents that automate repetitive workflows. Driven by Codex, these agents operate within their own workspace, can access files, call tools, and run continuously in the background. Users describe a workflow—such as collecting information, making judgments, generating results, and sending outputs—to ChatGPT, which then builds the agent. The entire team can use and refine the agent via ChatGPT or Slack. Once set up, the agent runs autonomously. Workspace Agents are team-shared, code-free, and manageable with predefined permissions and controls. They are suitable for structured, repeatable tasks involving multiple tools and continuous operation, such as software review, feedback整理, report generation, sales follow-up, and risk assessment. The system operates within clear rules, with sensitive actions requiring human confirmation. It aims to transform workflows from being person-dependent to being documented, executable, and reusable—akin to scientific management principles. Competing solutions from Microsoft and Google are also emerging in this space.

marsbit04/24 00:50

OpenAI Launches Workflow Agent, GPTs Begin Countdown

marsbit04/24 00:50

The Richest Fed Chair in History? Three Challenges Kevin Warsh Is About to Face

Kevin Warsh, nominated by Trump as the next Federal Reserve Chair, faces three major challenges as he prepares to take office. With a personal investment portfolio exceeding $130 million, including holdings in crypto protocols and public chains, he is set to become the wealthiest Fed chair in history. Warsh, known for his historically hawkish stance, has recently shifted his position, arguing that AI is a powerful deflationary force that could boost productivity. However, rising CPI and energy prices pose a risk to this outlook. He strongly defended Fed independence, stating that it is eroded not by political pressure but by the institution’s own missteps, such as blurring the lines between monetary and fiscal policy during the 2021-2022 inflation period. Additionally, Warsh advocates for significant balance sheet reduction while potentially cutting rates—a combination that markets find unsettling. He also proposes modernizing inflation measurement with real-time data, including stablecoins and on-chain pricing, to improve policy responsiveness. His confirmation is currently delayed due to a political investigation into current Chair Powell, adding uncertainty to his appointment.

marsbit04/24 00:36

The Richest Fed Chair in History? Three Challenges Kevin Warsh Is About to Face

marsbit04/24 00:36

Rivaling a Documentary, OpenAI President Recounts the 72-Hour Internal Strife

OpenAI President Greg Brockman details the 72-hour internal crisis following CEO Sam Altman's sudden firing by the board. Both were blindsided by the decision, leading Brockman to resign immediately in solidarity. They began planning a new company at Altman’s home, aiming to take all OpenAI employees with them. Despite intense recruitment efforts by competitors, not a single employee accepted an external offer. The situation shifted when co-founder Ilya Sutskever publicly supported the employees' petition, leading to Altman’s reinstatement. Brockman reflects on OpenAI’s origins, its shift from non-profit to capped-profit for computational scale, key breakthroughs like Dota project and GPT, and the importance of resilience and mission-focused leadership. He emphasizes AI’s potential for global benefit and the need for broad access to compute power, while acknowledging risks and the necessity of iterative deployment for safety and alignment.

marsbit04/24 00:28

Rivaling a Documentary, OpenAI President Recounts the 72-Hour Internal Strife

marsbit04/24 00:28

AI "Transfer Station" Earning Millions Monthly? Five Questions Uncover the Truth of Token Arbitrage

The article "AI 'Transfer Station' Earns Millions Monthly? Five Questions Uncover the Truth of Token Arbitrage" explores the emerging business of API token transfer stations, which profit from global AI service price disparities and access barriers. These intermediaries purchase low-cost tokens from overseas AI providers (e.g., OpenAI, Claude) through grey-market methods—such as exploiting enterprise credits, bulk accounts, or subscription benefits—and resell them to Chinese users at a markup. Key drivers include the high cost of using top AI models (e.g., Claude Code costs ~$5 per million tokens), the performance gap between domestic and foreign models, and mismatches between subscription and API pricing. However, the practice carries significant risks: upstream token sources may be unstable or illegal; user data passing through intermediaries can be harvested or injected with hidden prompts; and models might be downgraded without disclosure. The market is evolving, with some operators now exporting cheaper Chinese models (e.g., Qwen3.5 at ~$0.11 per million tokens) to overseas users, leveraging price gaps. Yet, sustainability is low due to compliance crackdowns, instability, and reputational risks. Users are advised to employ detection methods (e.g., prompt adherence tests) and avoid sensitive data usage. The authors caution that while transfer stations offer short-term arbitrage, they lack long-term reliability and security compared to official APIs.

marsbit04/24 00:26

AI "Transfer Station" Earning Millions Monthly? Five Questions Uncover the Truth of Token Arbitrage

marsbit04/24 00:26

a16z: Scaling AI Without Cryptographic Verification Is a Dangerous Liability

A16z argues that scaling AI without cryptographic verification is a dangerous liability. As AI agents rapidly evolve from tools into autonomous economic participants, they currently lack standardized, portable identities, verifiable permissions, and programmable payment methods. This creates systemic risks in an economy where non-human entities already vastly outnumber human users in sectors like finance. Blockchain infrastructure offers a solution by providing a neutral coordination layer. It enables verifiable, on-chain credentials for agent identity (a "Know Your Agent" standard), ensures transparent governance to prevent centralized control of AI systems, and facilitates native payments through stablecoins and emerging markets for AI-to-AI commerce. Without cryptographic guarantees—such as auditable transaction records, constrained agent behavior, and proof of origin—scaling AI agents accumulates unmanaged risk. Trust deficit, not intelligence, becomes the bottleneck. The authors conclude that cryptographic verification is essential to maintain user control, ensure accountability, and safely delegate economic activity to autonomous systems.

marsbit04/23 13:25

a16z: Scaling AI Without Cryptographic Verification Is a Dangerous Liability

marsbit04/23 13:25

Yao Shunyu's 88 Days

Yao Shunyu, a 27-year-old AI expert with a background from Princeton and OpenAI, joined Tencent in September 2025. Within 88 days, he led a major overhaul of Tencent’s AI strategy and organization, resulting in the release of Hunyuan Hy3 preview—a MoE model with 295B total parameters and 21B active parameters, supporting up to 256K context length. The launch came after Tencent leadership, including CEO Ma Huateng and President Martin Lau, openly criticized Hunyuan's earlier underperformance—citing slow development, over-reliance on superficial benchmark optimization, and poor generalization in real-world applications. Internal adoption was low, with key business units like WeChat and gaming seeking external AI solutions. Yao reshaped Tencent’s AI approach by integrating previously siloed teams, dissolving the ten-year-old Tencent AI Lab, and establishing new units focused on AI infrastructure and data. Hy3 preview was developed using co-design principles, closely aligned with product teams to ensure practical usability from the start. It has already been integrated into core products like Yuanbao, QQ, and enterprise tools. The release signals a shift from chasing rankings to building usable, scalable AI grounded in Tencent’s ecosystem. While external partnerships (like with DeepSeek and OpenClaw) helped retain users temporarily, the focus is now on making Hunyuan a reliable internal foundation. The real test lies in sustaining this new organizational momentum amid fierce competition from Alibaba, DeepSeek, and others.

marsbit04/23 11:13

marsbit04/23 11:13

This Time, OpenAI Eliminated 90% of Human Designers

OpenAI's latest release, GPT-Image 2, marks a paradigm shift in AI-generated imagery, moving beyond aesthetic quality to logical reasoning and contextual understanding. The model introduces a "thinking mode," where it performs background reasoning—such as mathematical calculations and geographic knowledge—before generating images. This enables highly accurate and context-aware outputs, like a livestream overlay showing precise distance metrics or a brand-aligned poster design. The model excels in rendering Chinese text with remarkable accuracy and aesthetic quality, a significant improvement over previous versions. It supports multi-turn conversational editing via the new Responses API, allowing iterative refinements similar to chatting with a large language model. While GPT-Image 2 demonstrates unprecedented capabilities in commercial applications like marketing material and illustration—potentially displacing many human designers due to its cost efficiency—it still has limitations. Minor artifacts in fine text details persist, and complex prompts can cause extended processing times. Additionally, the technology raises ethical concerns around deepfakes and digital trust. Overall, GPT-Image 2 transitions AI image generation from a novelty to a powerful production-ready tool, redefining industry standards and pushing the boundary of what’s possible in visual AI.

marsbit04/23 10:03

This Time, OpenAI Eliminated 90% of Human Designers