AI News

Meta’s Brain2Qwerty v2 Translates Brain Activity Into Text, No Surgery Required

Meta’s Brain2Qwerty v2 Translates Brain Activity Into Text, No Surgery Required

Meta's non-invasive Brain2Qwerty v2 uses a MEG scanner and an end-to-end AI model to turn brain activity into typed text, reaching 61% word accuracy, roughly eight times better than earlier non-invasive methods.

Read More
Meta’s Brain2Qwerty v2 Translates Brain Activity Into Text, No Surgery Required
Meituan’s LongCat-2.0 Was Trained Entirely on Chinese Chips, the Company Says

Meituan’s LongCat-2.0 Was Trained Entirely on Chinese Chips, the Company Says

Meituan says its new open-source LongCat-2.0, a 1.6-trillion-parameter model, was pre-trained and served entirely on domestically developed Chinese chips, a direct challenge to US export controls on advanced silicon.

Read More
Meituan’s LongCat-2.0 Was Trained Entirely on Chinese Chips, the Company Says
Commerce Department Partially Restores Anthropic’s Mythos 5 Under a Limited Export License

Commerce Department Partially Restores Anthropic’s Mythos 5 Under a Limited Export License

US Commerce Secretary Howard Lutnick has eased export controls on Anthropic's Mythos 5, letting approved organizations use it again under a limited license, while restrictions on the Fable 5 model remain in place.

Read More
Commerce Department Partially Restores Anthropic’s Mythos 5 Under a Limited Export License
Eastern Interconnection to Exhaust Emergency Reserves by 2027

Eastern Interconnection to Exhaust Emergency Reserves by 2027

The Eastern Interconnection is projected to hit zero emergency peak reserves by June 2027, raising the risk of rolling blackouts. Here's what businesses need to know.

Read More
Eastern Interconnection to Exhaust Emergency Reserves by 2027
Qihoo 360’s Tulongfeng: 3,432 Vulns Found, AI Sovereignty Debate Heats Up

Qihoo 360’s Tulongfeng: 3,432 Vulns Found, AI Sovereignty Debate Heats Up

Qihoo 360 unveils Tulongfeng AI vulnerability hunter claiming 3,432 flaws; Zhou frames it as cyber nuclear deterrence, fueling AI sovereignty debate.

Read More
Qihoo 360’s Tulongfeng: 3,432 Vulns Found, AI Sovereignty Debate Heats Up
Z.ai GLM-5.2 Matches Mythos in Cybersecurity Bug-Finding

Z.ai GLM-5.2 Matches Mythos in Cybersecurity Bug-Finding

China's Z.ai open-weight GLM-5.2 model now matches Anthropic's Mythos in cybersecurity bug-finding, narrowing the US-China AI gap and raising new oversight concerns.

Read More
Z.ai GLM-5.2 Matches Mythos in Cybersecurity Bug-Finding
OpenAI Previews GPT-5.6 Sol With Ultra Subagent Mode and Cyber Safeguards

OpenAI Previews GPT-5.6 Sol With Ultra Subagent Mode and Cyber Safeguards

OpenAI begins a limited preview of GPT-5.6 Sol with new ultra reasoning, cyber safety, and government coordination. Here's what it means for AI capabilities.

Read More
OpenAI Previews GPT-5.6 Sol With Ultra Subagent Mode and Cyber Safeguards
Vantage Data Center Generators Spark Noise Revolt in Virginia Town

Vantage Data Center Generators Spark Noise Revolt in Virginia Town

Vantage Data Center's 24/7 backup generators in Sterling, Virginia spark noise complaints, sleep disruption, and calls to rethink off-grid data center siting near homes.

Read More
Vantage Data Center Generators Spark Noise Revolt in Virginia Town
OpenAI Restricts GPT-5.6 Rollout at Trump Administration’s Request

OpenAI Restricts GPT-5.6 Rollout at Trump Administration’s Request

OpenAI has restricted initial access to its GPT-5.6 series, marking the first staggered ChatGPT rollout at the U.S. government's request. Here's what changed.

Read More
OpenAI Restricts GPT-5.6 Rollout at Trump Administration’s Request
Intercept’s $500M Plan to End Common Respiratory Infections Using AI

Intercept’s $500M Plan to End Common Respiratory Infections Using AI

Intercept launches a $500M initiative to develop broad-spectrum preventatives and air-cleaning tech that could end colds, flu, and COVID-like illnesses.

Read More
Intercept’s $500M Plan to End Common Respiratory Infections Using AI
Mini Shai-Hulud Supply-Chain Attack Hits Mistral AI, TanStack Packages

Mini Shai-Hulud Supply-Chain Attack Hits Mistral AI, TanStack Packages

The Mini Shai-Hulud supply-chain campaign compromised Mistral AI and TanStack packages on npm and PyPI, stealing GitHub tokens and cloud credentials from developer machines.

Read More
Mini Shai-Hulud Supply-Chain Attack Hits Mistral AI, TanStack Packages
Survey: 60% of U.S. Consumers Find AI in Brand Messaging a Turnoff

Survey: 60% of U.S. Consumers Find AI in Brand Messaging a Turnoff

New WordPress VIP survey: 60% of US consumers find AI in brand messaging a turnoff. 86% don’t fully trust AI answers. Here’s what that means for trust and AI search.

Read More
Survey: 60% of U.S. Consumers Find AI in Brand Messaging a Turnoff
Qwen3.6-27B Dense Model Beats Qwen3.5-397B-A17B on Coding Benchmarks

Qwen3.6-27B Dense Model Beats Qwen3.5-397B-A17B on Coding Benchmarks

Alibaba's Qwen team releases Qwen3.6-27B, an open-source 27B dense model that beats its 397B MoE predecessor on SWE-bench Verified and other coding benchmarks.

Read More
Qwen3.6-27B Dense Model Beats Qwen3.5-397B-A17B on Coding Benchmarks
OpenAI Partner Network Launches with $150M to Drive Enterprise AI

OpenAI Partner Network Launches with $150M to Drive Enterprise AI

OpenAI launches Partner Network with $150M investment to train 300,000 consultants and deploy enterprise AI. Founding partners: Accenture, Bain, McKinsey.

Read More
OpenAI Partner Network Launches with $150M to Drive Enterprise AI
Cursor Unveils Origin, a Git Forge for AI Agents at 22.6 Commits/Sec

Cursor Unveils Origin, a Git Forge for AI Agents at 22.6 Commits/Sec

Cursor launches Origin, an agent-native git forge that handles 22.6 commits per second. Built for AI coding agents with MCP extensibility and auto merge resolution.

Read More
Cursor Unveils Origin, a Git Forge for AI Agents at 22.6 Commits/Sec