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. gemini-omni-1.1-flash 85 pt5. Wan 3.0 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) 85 pt3. Muse Spark 1.3 (max) 47 pt4. gpt-oss-120b (high) 43 pt5. GPT-5.6 Luna (max) 28 pt

AI News

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

The 88% Surge in Screenless Wearables Driving the Slowtech Trend

The 88% Surge in Screenless Wearables Driving the Slowtech Trend

Back Market promotes the iPod Shuffle to highlight zero screen time. Screenless wearables saw an 88 percent spending increase in the US. Slo

Why AI Code Generation is Eroding Developer Cognitive Ownership

Why AI Code Generation is Eroding Developer Cognitive Ownership

AI shifts coding from deterministic abstraction to probabilistic agency. Over-reliance on AI leads to a loss of causal understanding in team

The 130,000 Trajectories in XDOF's ABC Dataset for Robot Learning

The 130,000 Trajectories in XDOF's ABC Dataset for Robot Learning

XDOF raised $70 million to build data pipelines for physical AI. The new ABC dataset provides 130,000 trajectories for robot learning. The c

DeepL Acquires Mixhalo to Master Real-Time Translation in Noisy Crowds

DeepL Acquires Mixhalo to Master Real-Time Translation in Noisy Crowds

DeepL acquired Mixhalo to expand into real-time live event translation. The acquisition helps DeepL enter the US market with a Bay Area offi

The PRD-to-Code Shift: Why AI Engineers Are Becoming Gatekeepers

The PRD-to-Code Shift: Why AI Engineers Are Becoming Gatekeepers

AI is shifting the software engineering bottleneck from production to verification. Technical moats are moving from syntax mastery to high-s

Anthropic's AI-Native Strategy Redefines the Path to MVP by 2026

Anthropic's AI-Native Strategy Redefines the Path to MVP by 2026

Anthropic released a playbook for building AI-native startups by 2026. The strategy guides founders through four stages from idea to scale.

WordPress VIP Finds 60% of US Consumers Averse to AI-Centric Messaging

WordPress VIP Finds 60% of US Consumers Averse to AI-Centric Messaging

A WordPress VIP survey reveals 60% of US consumers feel repulsion toward AI branding. Businesses face a paradox between AI discoverability a

sfs Launches Shared Brain Filesystem to Solve AI Agent Context Loss

sfs Launches Shared Brain Filesystem to Solve AI Agent Context Loss

The sfs filesystem creates a shared brain for AI agents across multiple devices. It uses local-first synchronization with S3 and GCS to elim

DOJ Backs xAI Gas Turbines, Citing AI as a National Security Asset

DOJ Backs xAI Gas Turbines, Citing AI as a National Security Asset

The DOJ supports xAI's use of mobile turbines for its data centers. Officials argue that Grok is essential for mission-critical operations.

VibeThinker-3B Outperforms Gemini 3 Pro in Math Reasoning

VibeThinker-3B Outperforms Gemini 3 Pro in Math Reasoning

Sina Weibo released VibeThinker-3B with high math reasoning capabilities. The model scored 94.3 on AIME 2026, surpassing Gemini 3 Pro. Resea

The 'Fix This Code' Prompt That Triggered Fable 5 Export Controls

The 'Fix This Code' Prompt That Triggered Fable 5 Export Controls

US export controls restricted access to Anthropic's Fable 5 and Mythos 5. A simple prompt change bypassed security guardrails to fix vulnera

AI Coding Agents Boost Output 4x but Spike Review Times 441.5%

AI Coding Agents Boost Output 4x but Spike Review Times 441.5%

AI coding agents have increased raw code output while severely degrading software quality. Review times have surged by 441.5% due to a pheno

The 50% Catastrophe Probability Driving the AI Governance Crisis

The 50% Catastrophe Probability Driving the AI Governance Crisis

AI investment now dwarfs the Manhattan Project by a factor of 100. Experts estimate a 10 to 50 percent chance of AI-driven catastrophe. US a

Wolfram 15 Turns AI Hallucinations Into Precise Computation

Wolfram 15 Turns AI Hallucinations Into Precise Computation

Wolfram released Version 15 of its integrated computational language. The system uses 7,000 primitives to verify AI generated code logic. De

Anthropic Hits 41% Enterprise Share as Government Bans Fuel Demand

Anthropic Hits 41% Enterprise Share as Government Bans Fuel Demand

Anthropic captured 41% of the enterprise AI market in May. US government restrictions on Mythos 5 fueled corporate demand. Anthropic now sur

sandboxd Enables Self-Hosted AI App Builders With Isolated Linux Containers

sandboxd Enables Self-Hosted AI App Builders With Isolated Linux Containers

sandboxd is an open-source engine designed for AI app builder platforms. It creates isolated Linux containers via a single HTTP request for

CodeBurn Integrates Cost Analysis for 25 AI Coding Tools Locally

CodeBurn Integrates Cost Analysis for 25 AI Coding Tools Locally

CodeBurn enables local cost tracking for 25 different AI coding tools. The utility analyzes session data on disk to avoid using proxy server

Gemma 4 Hits 75% Frontier Performance in Local Coding Loops

Gemma 4 Hits 75% Frontier Performance in Local Coding Loops

Gemma 4 achieves 75% of frontier model performance in local coding. Developers use Pi and LM Studio to build secure local agent loops. Hardw

Show GN Bridges the AI Memory Gap With Graph-Based MCP

Show GN Bridges the AI Memory Gap With Graph-Based MCP

Show GN introduces an open-source MCP for AI memory storage. The tool utilizes Event-Centric Knowledge Graphs to evolve data structures. Use

Claude Code and the Shift Toward Parallel AI Development Workflows

Claude Code and the Shift Toward Parallel AI Development Workflows

Anthropic's Claude Code introduces parallel workflows using worktrees. The tool automates validation via Chrome control and GIF reporting. D

Stanford's DeLM Slashes Multi-Agent Costs by 50% Without a Controller

Stanford's DeLM Slashes Multi-Agent Costs by 50% Without a Controller

Stanford developed DeLM to remove central orchestrators from AI workflows. The system reduces per-task operational costs by approximately 50

Tim Ferriss's 80% Sales Drop Reveals the Death of Prescriptive Nonfiction

Tim Ferriss's 80% Sales Drop Reveals the Death of Prescriptive Nonfiction

Tim Ferriss reports an 80% drop in book sales since the launch of LLMs. Prescriptive nonfiction is collapsing as users shift to personalized

GPT-NL: The 13.5 Million Euro Bet on Dutch Digital Sovereignty

GPT-NL: The 13.5 Million Euro Bet on Dutch Digital Sovereignty

The Netherlands is investing 13.5 million euros to build GPT-NL from scratch. This sovereign LLM aims to eliminate dependence on US-based AI

GLM-5.2 Outperforms GPT-5.5 on SWE-bench Pro With MIT License

GLM-5.2 Outperforms GPT-5.5 on SWE-bench Pro With MIT License

Z.ai released the GLM-5.2 open-weight model for autonomous coding. The model outperforms GPT-5.5 on the SWE-bench Pro benchmark. IndexShare

Databricks LTAP and Lakehouse//RT Remove the AI Agent Data Pipeline

Databricks LTAP and Lakehouse//RT Remove the AI Agent Data Pipeline

Databricks introduced LTAP and Lakehouse//RT to eliminate complex ETL pipelines. The new architecture supports 12,000 QPS with sub-100ms lat

Android 17 Integrates Gemini Omni to Turn Smartphones Into Creative Studios

Android 17 Integrates Gemini Omni to Turn Smartphones Into Creative Studios

Google released Android 17 and Wear OS 7 with deep AI integration. Gemini Omni and Lyria 3 enable multimodal video and music creation. Pixel

Qwopus-3.6-27B Brings Agentic Coding to a Single GPU

Qwopus-3.6-27B Brings Agentic Coding to a Single GPU

Alibaba DAMO Academy released the Qwopus-3.6-27B-Coder model. The model achieves a 67.0% success rate on SWE-bench Verified. Trace Inversion

Plaud Hits $100 Million ARR With 2 Million AI Recorders Sold

Plaud Hits $100 Million ARR With 2 Million AI Recorders Sold

Plaud surpassed $100 million in software ARR after selling 2 million devices. The company uses a hardware gateway to drive high-margin softw

Robinhood Cuts 10% of Staff While Avoiding the AI Narrative

Robinhood Cuts 10% of Staff While Avoiding the AI Narrative

Robinhood is laying off 10 percent of its full-time workforce. The company is cutting staff despite a 15 percent increase in Q1 revenue. CEO

OpenAI's $38.5 Billion Loss Reveals the True Cost of Frontier AI

OpenAI's $38.5 Billion Loss Reveals the True Cost of Frontier AI

OpenAI recorded a net loss of 38.5 billion dollars in 2025. Much of the deficit stems from a complex corporate restructuring. Microsoft rema