AI News
Daily AI industry news — funding, products, policy, and major moves from global AI companies and startups.
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
Qwen3.6-27B Doubles Inference Speed via Multi-Token Prediction
Qwen3.6-27B introduces Multi-Token Prediction to double inference speeds. The model supports a native context window of 262,144 tokens. Loca
Anthropic Launches Claude Code and Opus 4.7 to Shift AI Coding to Verification
Anthropic released Claude Code and Opus 4.7 to enhance developer autonomy. New memory structures and Dreaming features improve agentic task
GPT-5.5 Hits IQ 136 in New AI IQ Bell Curve Mapping
GPT-5.5 ranks first with an estimated IQ of 136 on aiiq.org. The system maps 50 models using 12 benchmarks across four reasoning areas. Anth
Claude Code and the Dangerous Allure of Vibe Coding
Claude Code enables a new trend called Vibe Coding where developers skip reviews. The shift in productivity moves the engineering bottleneck
Meta AI on Threads Sparks Backlash Over Missing Block Button
Meta is testing @MetaAI tagging in five countries on Threads. Over one million users protested the inability to block the AI. Meta suggests
Why Senior Developers are Splitting AI Code into Speed and Scale Versions
Senior developers are separating AI code into Speed and Scale versions. The Speed version validates market demand using rapid AI generation.
Claude Design Project Access Vanishes After Subscription Cancellation
Anthropic users report losing project access after canceling Claude Design. Compensation credits disappear immediately upon subscription ter
The Full-Stack Strategy Powering OpenAI and Anthropic's Commercial Lead
OpenAI and Anthropic are shifting focus from model intelligence to commercial speed. The US leads through a vertically integrated stack from
Anthropic Restores External Agent Access via New Agent SDK Credits
Anthropic introduced Agent SDK credits for paid subscribers. External agents like OpenClaw now have tiered monthly credit limits. The system
Anthropic Overtakes OpenAI With 34.4% Enterprise Market Share
Anthropic reached a 34.4% US enterprise adoption rate in April. Claude Code now generates 4% of all public GitHub commits. High token costs
Anthropic Targets $950 Billion Valuation as Claude Shifts to Proactive Automation
Anthropic is pursuing a valuation of $950 billion to surpass OpenAI. The company is transitioning Claude from a reactive tool to a proactive
GPT-5.5 Codex: The 1.43x Cost Jump That Unlocks Domain Modeling
GPT-5.5 Codex improves code quality as reasoning effort settings increase. The high setting provides the best balance of cost and review pas
Claude API Proxy Scams: 90% Discounts Masking 37% Accuracy
Fraudulent proxy services offer Claude API access at 10% of the cost. These services drop medical benchmark accuracy from 84% to 37%. Stolen
Disrupt 2026 Unveils 6-Stage Blueprint for AI Startup Survival
Disrupt 2026 will host over 10,000 attendees at Moscone West. The event features six specialized stages focusing on AI implementation. Sessi
The 34.4% Market Share That Put Anthropic Ahead of OpenAI
Anthropic has overtaken OpenAI with a 34.4% corporate market share. Data from Ramp shows a surge from 9% to 34.4% within one year. Corporate
Anthropic Targets the 44% of US GDP Held by Small Businesses
Anthropic launched Claude for Small Business to automate local operations. The tool integrates with QuickBooks, Canva, and PayPal for workfl
Alexa for Shopping Replaces Rufus to Automate External Purchases
Amazon launched Alexa for Shopping to automate personalized commerce. The new AI agent can execute purchases on external websites via Buy fo
Meta AI Secret Mode Aims to Remove Privacy Barriers in WhatsApp
Meta is launching a Secret Mode for Meta AI within WhatsApp. The feature uses Muse Spark and a private processing infrastructure. Side Chat
Statewright Boosts Local Model Performance 5x via State Machines
Statewright uses state machines to restrict AI agent tool access. Local models saw SWE-bench scores jump from 2/10 to 10/10. The engine inte
Qwopus3.6-35B-A3B-v1 Hits 81.1% HumanEval with 3B Active Parameters
Qwopus3.6-35B-A3B-v1 achieves over 80% accuracy in coding benchmarks. The model uses a hybrid MoE architecture with only 3 billion active pa
Why AI Agents Make Rust More Productive Than Zig for Systems Code
AI coding agents are shifting systems programming preferences toward Rust. Rust's strict type system automates verification for AI-generated
Everything-Gemini-Code Ported 183 Claude Code Skills to Gemini
Everything-Gemini-Code ports 183 specialized skills from Claude Code to Gemini. The project implements an automated drift detection system t
Why Claude Code Needs 90% Test Coverage to Scale to 970,000 Lines
Garry Tan used Claude Code to generate 970,000 lines of code and 665 tests. Test coverage above 85% increases defect removal rates to nearly
How Matt Pocock's Claude Code Skills Replace Vibe Coding with Engineering
Matt Pocock released a modular skill library for Claude Code. The toolkit replaces monolithic AI frameworks with discrete tools. Specific pr
DeepSeek-TUI Automates Terminal Coding with 1M Token Context
DeepSeek-TUI integrates AI coding agents directly into the terminal. The tool features a 1M token context window and automatic model routing
Loop Intelligence: Solving the Organizational Learning Gap in the AI Era
AI tool adoption does not automatically improve organizational skill. Loop Intelligence captures individual AI wins to scale them company-wi
Unitree GD01: The $537,000 Rideable Robot Entering Mass Production
Unitree launched the GD01 rideable robot priced at $537,000. The machine features both bipedal and quadrupedal movement modes. Unitree is pu
Claude Platform on AWS Bridges the Gap Between Native APIs and Cloud Scale
Anthropic launched Claude Platform on AWS with native API capabilities. The service integrates AWS IAM and CloudTrail for enterprise securit
Anthropic Open-Sources 10 Financial AI Agents to Automate Workflows
Anthropic released 10 open-source financial AI agents under Apache 2.0. The agents integrate with 11 data providers via the Model Context Pr