KO EN

AX BRIEF

AI news, benchmarks & engineering blog curation

● LIVE
지능1. Claude Fable 5.1 (max with fallback) 100 pt2. Claude Fable 5 (with fallback) 96 pt3. Claude Opus 5 (max) 95 pt4. GPT-6 Astra (max) 95 pt5. GPT-5.6 Sol (max) 91 pt코딩1. Claude Fable 5.1 (max) 100 pt2. Claude Opus 5 (xhigh) 94 pt3. GPT-6 Astra (max) 91 pt4. Muse Spark 1.3 (xhigh) 81 pt5. Grok 4.5 (high) 81 pt이미지1. GPT Image 2 (high) 100 pt2. MAI-Image-2.6 93 pt3. GPT Image 1.5 (high) 91 pt4. Reve 2.1 88 pt5. Muse Image 85 pt비디오1. Gemini Omni Flash 100 pt2. Minimax H3 Max (post-trained by fal) 94 pt3. Dreamina Seedance 2.0 720p 91 pt4. Wan 3.0 85 pt5. gemini-omni-1.1-flash 85 pt가격1. Claude Fable 5.1 (max with fallback) 100 pt2. GPT-6 Astra (max) 100 pt3. Claude Fable 5 (with fallback) 100 pt4. Claude Opus 5 (max) 49 pt5. GPT-5.6 Sol (max) 39 pt속도1. Gemini 3.5 Flash-Lite 100 pt2. Gemini 3.8 Flash (high) 84 pt3. Muse Spark 1.3 (max) 63 pt4. gpt-oss-120b (high) 50 pt5. GPT-5.6 Luna (max) 30 pt

AI News

Daily AI industry news — funding, products, policy, and major moves from global AI companies and startups.

Elon Musk vs. OpenAI: The $10 Billion Legal Battle Over AI Governance

Elon Musk vs. OpenAI: The $10 Billion Legal Battle Over AI Governance

Elon Musk challenges OpenAI over its shift from non-profit to for-profit status. Microsoft's $10 billion investment serves as the focal poin

Runway's $5.3 Billion Pivot From Video Generation to World Models

Runway's $5.3 Billion Pivot From Video Generation to World Models

Runway is expanding its vision from AI video tools to comprehensive world models. The company leverages observational data to understand phy

Why Claude for Legal Relies on Managed Agents API for Law Firm Workflows

Why Claude for Legal Relies on Managed Agents API for Law Firm Workflows

Anthropic released Claude for Legal with a Managed Agents API. The system uses MCP connectors to verify legal citations in real-time. A cold

WhichLLM Matches Local LLMs to Your Specific Hardware Specs

WhichLLM Matches Local LLMs to Your Specific Hardware Specs

WhichLLM automatically detects hardware to recommend the best local LLMs. The tool integrates real-time benchmarks to rank models by actual

The Closed AI Pivot: OpenAI and Anthropic Restrict Security Model Access

The Closed AI Pivot: OpenAI and Anthropic Restrict Security Model Access

OpenAI and Anthropic are limiting access to their cybersecurity models. Model distillation and national security drive this shift toward clo

Why Claude Code Swapped RAG for Agentic Search in Massive Codebases

Why Claude Code Swapped RAG for Agentic Search in Massive Codebases

Anthropic introduced Claude Code to manage massive monorepos locally. The tool replaces traditional RAG with a live agentic search system. P

Ontario AI Scribe Audit Reveals 60% Medication Error Rate

Ontario AI Scribe Audit Reveals 60% Medication Error Rate

An audit of 20 AI Scribe vendors in Ontario found medication errors in 60% of systems. Procurement metrics prioritized local business presen

The 40% Grade Gap Signaling LLM Penetration at the University of Chicago

The 40% Grade Gap Signaling LLM Penetration at the University of Chicago

LLMs have permeated exams and student journalism at the University of Chicago. A 40 percentage point gap exists between take-home and in-per

The 20-Minute LLM Training Pipeline That Demystifies GPT Architecture

The 20-Minute LLM Training Pipeline That Demystifies GPT Architecture

A new workshop enables users to train a small LLM in 20 minutes. The project optimizes a 10M parameter pipeline for local hardware. Develope

Why OpenDesign and GPT-5.5 Struggle to Match Claude Design's Polish

Why OpenDesign and GPT-5.5 Struggle to Match Claude Design's Polish

Developers are testing OpenDesign with GPT-5.5 as a Claude Design alternative. The open-source combination lacks the pixel-perfect polish of

Recursive Superintelligence Raises $650 Million for Self-Improving AI

Recursive Superintelligence Raises $650 Million for Self-Improving AI

Recursive Superintelligence secured $650 million to build self-redesigning AI. The startup features a team of veterans from OpenAI and Googl

Raindrop AI Workshop Brings Local SQL Telemetry to Agent Debugging

Raindrop AI Workshop Brings Local SQL Telemetry to Agent Debugging

Raindrop AI released Workshop as an open-source local debugging tool. The tool stores agent telemetry in local SQL files to ensure data sove

OpenAI Codex Integration Turns the ChatGPT App Into a Remote Dev Console

OpenAI Codex Integration Turns the ChatGPT App Into a Remote Dev Console

OpenAI integrated Codex into the ChatGPT app for mobile users. Developers can now monitor and approve live code environments remotely. The u

OpenAI Weighs Breach of Contract Suit Over Apple's ChatGPT Integration

OpenAI Weighs Breach of Contract Suit Over Apple's ChatGPT Integration

OpenAI is reviewing legal options against Apple for breach of contract. The dispute centers on poor visibility of ChatGPT within Apple's OS.

Clawdmeter Turns Claude Code Token Usage Into a Physical Dashboard

Clawdmeter Turns Claude Code Token Usage Into a Physical Dashboard

Hermann Haraldsson released Clawdmeter to visualize Claude Code token usage. The device uses an ESP32-S3 AMOLED screen to show real-time API

OpenAI and Palantir: Why Organizational Structure is the New AI Moat

OpenAI and Palantir: Why Organizational Structure is the New AI Moat

AI product replication has made feature-based moats obsolete. OpenAI and Palantir use unique organizational designs to attract talent. Struc

The AI Coding Trap: How Two Years of Prompting Eroded a Developer's Skill

The AI Coding Trap: How Two Years of Prompting Eroded a Developer's Skill

A software developer reports significant skill loss after relying on AI for two years. Dependency on models like Claude led to severe impost

Claude Code's /goals Feature Solves the AI Agent Completion Problem

Claude Code's /goals Feature Solves the AI Agent Completion Problem

Anthropic introduced a /goals feature to Claude Code for better task verification. The system uses a separate Haiku model to evaluate if goa

The Sam Altman Business Probe Threatening OpenAI's IPO Path

The Sam Altman Business Probe Threatening OpenAI's IPO Path

US Republicans have launched an investigation into Sam Altman's business dealings. The probe focuses on potential conflicts of interest as O

Claude Recovers 5 BTC from a Forgotten 11-Year-Old Wallet

Claude Recovers 5 BTC from a Forgotten 11-Year-Old Wallet

A user recovered 5 BTC worth $400,000 using Anthropic's Claude. The AI identified a configuration error in the btcrecover tool. Claude analy

Qwen3.6-35B-A3B Hits 73.4 on SWE-bench Verified for Local Coding

Qwen3.6-35B-A3B Hits 73.4 on SWE-bench Verified for Local Coding

Qwen3.6-35B-A3B uses a Mixture of Experts architecture to activate only 3B parameters. The model achieves a 73.4 score on SWE-bench Verified

Why Claude Code and Codex Now Force 15-Minute Learning Sessions

Why Claude Code and Codex Now Force 15-Minute Learning Sessions

GitHub released a learning repository for Claude Code and Codex. The tool uses retrieval practice to prevent AI-driven skill decay. Develope

Ruflo Brings 100-Agent Swarm Orchestration to Claude Code

Ruflo Brings 100-Agent Swarm Orchestration to Claude Code

Ruflo enables the orchestration of over 100 specialized AI agents for Claude Code. The platform uses HNSW-based AgentDB to accelerate memory

Inside China's AI Labs: Student-Led Innovation and Vertical Stack Control

Inside China's AI Labs: Student-Led Innovation and Vertical Stack Control

Chinese AI labs are integrating students directly into core LLM development. Engineers are prioritizing vertical control over reliance on ex

Markbase Evolves from Time-Series Database to AI Runtime

Markbase Evolves from Time-Series Database to AI Runtime

Markbase introduced an internal script execution environment for data. The database now integrates REST API and MQTT for real-time connectiv

Anthropic Shifts Claude Programmatic Usage to Monthly Credit System

Anthropic Shifts Claude Programmatic Usage to Monthly Credit System

Anthropic is introducing monthly credits for Claude's programmatic tools. The new system separates API-style usage from standard chat subscr

The 90% Expert Agreement Rate Powering Forum AI's Compliance Engine

The 90% Expert Agreement Rate Powering Forum AI's Compliance Engine

Forum AI achieves a 90% agreement rate between human experts and AI judges. The company targets high-risk compliance in finance and geopolit

Clio Hits $500M ARR as Anthropic Disrupts the Legal AI Ecosystem

Clio Hits $500M ARR as Anthropic Disrupts the Legal AI Ecosystem

Clio reached $500 million in annual recurring revenue following AI integration. Anthropic entered the legal market directly with the Claude

How legalQ Uses RAG and MCP to Democratize Korean Legal Search

How legalQ Uses RAG and MCP to Democratize Korean Legal Search

legalQ provides a natural language interface for Korean law and precedents. The system uses RAG and MCP to eliminate LLM hallucinations in l

The Vercel Claude Code Telemetry Setting Most Devs Haven't Disabled

The Vercel Claude Code Telemetry Setting Most Devs Haven't Disabled

Vercel's Claude Code plugin creates permanent UUIDs to track users. The tool sends device IDs and tool calls to Vercel servers by default. D