AI Briefing
Hacker News news
The latest AI and developer news from Hacker News, with short summaries.
Visit website →Handbook.md: A Benchmark for Long-Context Agentic Instruction Following
A new benchmark called Handbook.md has been released to evaluate whether AI agents perform tasks while complying with lengthy policy documents.
Transformer Transformer: A Unified Model for Motion-Conditioned Robot Co-Design
Proposes a unified diffusion transformer model that simultaneously optimizes a robot's morphology and control policy.
LearnVector – Andrew Ng's AI Company Building 1:1 Learning Experiences
Andrew Ng has founded LearnVector, a 1:1 personalized AI learning agent company, with investment from Coursera.
MCP 2026-07-28 Specification: Transport Shifts to Stateless
The Model Context Protocol (MCP) has shifted to a Stateless protocol to improve scalability and reliability.
Show HN: Ctrlb-decompose: A Tool That Strips Noise from Logs Before Sending Them to an LLM
This is a tool that patterns massive log data to compress it for LLM transmission.
Fast Recovery Is the New Trust Model (The JFrog and OpenAI Zero-Day Discovery Case)
Through the case of a JFrog Zero-Day vulnerability discovered by an OpenAI model, the importance of AI-driven security response was highlighted.
Neutrino-1 8B
Technical specifications have been released for Neutrino-1 8B, a dense decoder-only transformer model built on Qwen3-8B.
A $500 RL fine-tune of a 9B open model beat frontier models on catalog review
Through low-cost RL fine-tuning, a 9B-scale open model outperformed high-performance models on specific domain performance.
Judge Rejects Google's Attempt to Use DMCA as Defense Against Scraping
Google's attempt to use the DMCA to block data scraping was rejected in court.
Show HN: FeyNoBg – Automatic Background Removal Model and Training Library
FeyNoBg, an open-source background removal model and training library achieving SOTA performance, has been released.
MAI-Cyber 1
Microsoft has announced MDASH, a multi-agent system for identifying and resolving security vulnerabilities, along with a new model, MAI-Cyber-1-Flash.
Nvidia's $750 Billion Deal Reignites Concerns Over AI Circular Financing
Nvidia's massive deal is raising concerns about a circular financing structure in the AI industry, where money keeps flowing back into infrastructure.
AI Companies' Washington Lobbying Spending Hits Record High
AI companies are pouring record-breaking amounts of lobbying funds into Washington, D.C. in the U.S. to respond to regulation.
Truth Is Not a Direction: A Tarskian Attack on LLM Probes
Mathematical logic proves that no specific direction in an LLM's embedding space can represent Truth.
Increased Error Rate for Claude Opus 5
The error issue affecting Anthropic's Claude services and API has been resolved.
Chinese Chipmaker's Stock Surges 470%
CXMT, China's largest memory chip maker, drew market attention as its stock surged 470% on its first day of trading.
How AST-grep Rewrote Tree-sitter in Rust to Boost Performance by 30%
AST-grep rewrote Tree-sitter's C core in Rust to improve performance for AI coding agent optimization.
Terence Tao: Mathematics in the AI Era
Mathematician Terence Tao analyzed the impact of AI on mathematical discovery and research methods.
Inflect-Micro-v2: A Complete Voice Model Implemented with 9.36M Parameters
Inflect-Micro-v2, a highly efficient TTS model capable of local execution with fewer than 10 million parameters, has been released.
New AI Traffic Options for Cloudflare's Customers
Cloudflare has introduced a new management framework that classifies and controls traffic based on the purpose of AI bots.