# Пов'язані статті щодо Tool

Центр новин HTX надає останні статті та поглиблений аналіз на тему "Tool", що охоплює ринкові тренди, оновлення проєктів, технологічні розробки та регуляторну політику в криптоіндустрії.

OpenAI, Open-Sourced

OpenAI has open-sourced its code security tool, Codex Security CLI. The tool, which originated from the private beta project Aardvark in October 2025, is designed to automatically discover, verify, and fix vulnerabilities in codebases. It functions as an application security agent, first analyzing a repository to build a threat model, then identifying and ranking vulnerabilities based on real-world impact, and finally testing them in a sandbox for validation. According to OpenAI, in its first 30 days, the tool scanned over 1.2 million commits, uncovering 792 critical and 10,561 high-severity vulnerabilities, with a reported reduction of over 50% in false positives upon repeated scans of the same repositories. However, initial user experiences on platforms like Hacker News highlighted significant issues, particularly concerning cost and reliability. Developers reported failed scans that consumed substantial portions of API rate limits and incurred high expenses, with one user noting a cost of approximately $13 for an aborted run. The high cost is attributed to the tool's default configuration, which uses the premium GPT-5.6-sol model with inference intensity set to "extra-high." The release follows public statements by NVIDIA's Jensen Huang advocating for open-source AI. While OpenAI has open-sourced the application-layer CLI and SDK, the core AI models remain proprietary. The move opens the door for community development and potential adaptations of the tool.

marsbit07/30 07:43

OpenAI, Open-Sourced

marsbit07/30 07:43

The Fall of Zapper: An Act of God or a Human Error?

The Fall of Zapper: A Post-Mortem of a DeFi Pioneer In July 2026, Zapper, a once-dominant DeFi portfolio tracker, announced its shutdown. Born in 2020 from a merger, Zapper capitalized on the DeFi Summer boom, reaching 2 million monthly users and processing over $13B in transactions, backed by $16.5M in funding from investors like Framework Ventures and Coinbase Ventures. Its core "Zap" feature simplified complex multi-step DeFi operations. Despite its early success, Zapper failed to build a sustainable business model. Revenue from DEX aggregation was minimal due to fierce competition, while maintaining its multi-chain data infrastructure was costly. Furthermore, the DeFi landscape shifted: capital consolidated around top protocols, reducing the need for complex portfolio tracking across numerous platforms. Zapper's user base and core demand eroded. The company attempted multiple pivots, including an NFT-based points system, a social app (Chainchat), and plans for a ZAP token protocol. However, these initiatives—often focused on creating new, speculative C端需求 rather than solving existing pain points—ultimately failed. Critics argue Zapper remained trapped in a "blockchain purist" mindset, prioritizing costly, non-revenue-generating features over its competitive DEX aggregator. Unlike competitor DeBank, which successfully pivoted to its Rabby Wallet, Zapper lacked a diversified revenue stream. Its closure highlights the peril for tooling projects that fail to adapt to market shifts and monetize effectively, serving as a cautionary tale for the industry.

Foresight News07/10 02:11

The Fall of Zapper: An Act of God or a Human Error?

Foresight News07/10 02:11

Claude Code's Shocking Origin Exposed: It Evolved from Safety Alignment, Boris: Only 1% Complete

**"Claude Code's Astonishing Origin Revealed: Born from Safety Alignment, with Only 1% Done"** This article traces the epic development of Claude Code, Anthropic's groundbreaking AI coding assistant. Its origins are surprisingly rooted in an internal safety alignment (Alignment) project. The journey began in 2021 with early prototypes like a VS Code extension, but the project was nearly forgotten due to immense infrastructure challenges in creating a true "agentic" coder. Key breakthroughs came from research teams focused on autonomous software engineering, developing core components like bash tools and code search. An internal CLI tool named "clide" emerged but was too超前 (ahead of its time), being clunky and slow. The project's fate changed in September 2024 when Boris Cherny joined. Tasked with "agentic coding," he built a simple CLI prototype. A pivotal moment occurred when he used `clide` to generate a complete pull request from an issue description, revealing the assembled potential of earlier research. A small team then executed a furious two-week sprint to build the core product. Launched in February 2025 as Claude Code, initial feedback was mixed. However, with the release of the Claude 3.5 Sonnet model, its capabilities skyrocketed, fundamentally altering software development workflows in Silicon Valley. Notably, Boris Cherny himself reached a point where 100% of his coding was done silently by Claude Code in the terminal. Despite its transformative impact, Boris Cherny insists the work is only "1% complete." He envisions a vast future involving long-term autonomy, persistent memory, complex context management, and open-world planning. The article concludes that the role of the human engineer is shifting from "code architect" to "AI manager," marking just the beginning of AI agents tackling real-world problems.

marsbit07/07 12:31

Claude Code's Shocking Origin Exposed: It Evolved from Safety Alignment, Boris: Only 1% Complete

marsbit07/07 12:31

Exporting to Domestic Sales: The Chinese-style Outbound Journey of an AI Short Film

From Export to Domestic Boom: The Chinese-Style Overseas Journey of an AI Short Film The story begins with PJ Ace, a prominent Hollywood AI filmmaker, launching a public search on X for the creator of a stunning AI-generated short film titled "Zombie Scavenger." The film, featuring a robot cowboy in a post-apocalyptic wasteland, impressed Ace with its quality, which he estimated would have cost $500,000 and six months pre-AI. The trail led back to China. The creator, Mx-Shell, is a self-described amateur from China with a photography and music background. Using ByteDance's AI video tool, Seedance 2.0, he independently produced the short in about ten days for a minimal cost. Ironically, while the film went viral overseas after Ace's endorsement, it initially gained little traction on Chinese platforms like Bilibili. This sparked a "cross-server" search. Ace posted in English on X, while Mx-Shell, who doesn't speak English, posted his QQ email in Chinese comment sections. With netizens' help, they connected. Ace extended an invitation, asking if Mx-Shell was interested in becoming a Hollywood director. The article highlights this as a case of "talent export" or "brilliance going overseas." A creator in China, using domestic AI tools and computing power, captured global attention first. This "export-to-domestic-sales" path succeeded due to China's competitive, low-cost AI video tool market and its vast pool of untapped creative talent. Mx-Shell's success underscores that AI lowers production barriers, but core creativity, aesthetic judgment, and storytelling sense remain uniquely human. His path—individual, low-budget, and quality-driven—contrasts with the industrialized, capital-intensive route of bulk-producing AI short dramas for overseas markets. His story, spontaneous and beyond any corporate marketing plan, serves as powerful validation for tools like Seedance 2.0. The piece concludes that while China has many creators whose traditional barriers (equipment, funds, teams) are being dismantled by AI, the challenge of visibility remains. Until a robust domestic AI creative ecosystem develops, this indirect route of gaining overseas recognition first may continue to be a viable path for Chinese talent.

marsbit05/14 04:24

Exporting to Domestic Sales: The Chinese-style Outbound Journey of an AI Short Film

marsbit05/14 04:24

Agents Have Entered the Harness-Driven Era

The article discusses the significance of the leaked Claude Code from Anthropic, highlighting its revelation of advanced Agent engineering practices centered on "Harness" design. Rather than relying solely on model capabilities, modern AI systems now depend on a structured engineering framework—the Harness—to maximize performance. This framework includes six core components: multi-layered System Prompts, Tool Schema, Tool Call Loop (with Plan and Execute modes), Context Manager, Sub-Agent coordination, and Verification Hooks. The Harness enables tighter integration between training and inference, supports long-chain tool execution, and improves reliability through objective verification. It also drives six key training directions: behavior alignment via System Prompt, end-to-end tool-use training, integrated plan-execute training, memory compression, sub-agent orchestration, and multi-objective reinforcement learning. The shift to Harness-driven development reduces the emphasis on pure prompt engineering, favoring instead multidisciplinary talent with skills in AI, backend engineering, and infrastructure. The market is evolving toward more secure, private, and vertically integrated Agent deployments, with "model shell" companies needing either strong infrastructure or deep domain expertise to compete. Claude Code’s leak underscores that future AI advancements will be shaped by engineering architecture as much as by algorithmic innovation.

marsbit04/15 10:11

Agents Have Entered the Harness-Driven Era

marsbit04/15 10:11

活动图片