AI News Today — Model Releases, Benchmarks & Industry Moves

Innovating with AI in the special chemicals industry

As companies look to accelerate innovation with artificial intelligence, cloud computing, and advanced modeling, the primary challenge is connecting new technologies to reliable data, scientific expertise, and the practical realities of bringing products to market. For specialty chemicals manufacturer Lubrizol, digital transformation is all about giving researchers the tools to test ideas faster, make better-informed…

MIT

The Download: empty AI climate claims, and an Nvidia investigation

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. Opinion: The climate case for AI has not been made —By Mary Nichols, distinguished counsel at the UCLA School of Law and the former chair of the California Air Resources Board,…

TC

Automattic debuts a platform to deploy vibe coding apps amid a changing internet

Automattic aims to compete with likes of Lovable with Spacefast

Anthropic's $3.4 Billion Purchase of 1.7 GW of Generators Makes Energy Independence a New Variable in the AI Race

Anthropic signed a $3.4 billion contract with OnePWR Solutions to buy 1.7 GW of heavy-duty gas generators and engineering and installation capabilities, its largest known single self-built power procurement. The deal shifts its infrastructure strategy toward energy independence ahead of its IPO, even as the long-term cost implications remain undisclosed.

Trump Says for First Time He Does Not Oppose AI Guardrails as Bipartisan Legislative Window Opens

In a 60 Minutes interview, President Trump said he does not mind AI guardrails and signaled willingness to work with Democrats, while insisting the United States must not fall behind China. The remark removes a major political obstacle to federal AI legislation, though the prospects for concrete bills remain uncertain.

Trump Reveals DOJ Warned Senior Executive of Major AI Company to Prepare for Jail, Ushering in an Era of Criminal Deterrence

President Trump disclosed in a CBS "60 Minutes" interview that the Justice Department warned the head of a "major AI company" to "prepare to go to jail," signaling a shift toward criminal deterrence in AI regulation following the establishment of the DOJ Superintelligence Enforcement Task Force.

MIT

The climate case for AI has not been made

This week, as countries meet in Antalya, Turkey for the United Nations climate conference, one topic likely to dominate the conversation is artificial intelligence. The host has already called on countries to commit to ensuring that AI is designed, powered, and deployed in ways that support climate action. But so far, industries around the world…

WD

Authors Can Defend Against AI Allegations With ‘Paper Trail’ Tools. Is It Enough?

As the publishing industry wrestles with paranoia over the provenance of manuscripts, writers are caught in an impossible bind.

WD

Bikers and Bug-Lovers Join Forces to Fight a New Google Data Center

As data center protests break out across Europe, a mismatched band of campaigners is trying to block Google from building a sprawling new facility east of London.

TC

Is Jensen Huang’s job on the line? Yale’s Jeffrey Sonnenfeld estimates it’s “about a 50-50”

Nvidia’s board is investigating Jensen Huang over his relationship with a suspected Chinese spy, and Yale’s Jeffrey Sonnenfeld thinks it’s roughly a coin flip whether Huang stays on as CEO.

MIIT Notice No. 42: Every AI Agent Device Must Carry an Identity Code to Access the Network as the World's First Hardware-Layer Regulatory Framework Takes Effect

China's Ministry of Industry and Information Technology has issued Notice No. 42, requiring terminals pre-installed with AI agents to complete agent adaptation verification, fraud risk security assessment, and pre-emptive anti-fraud capability evaluation before obtaining a network access license, with each agent assigned a unique identity code. It is the world's first regulatory framework to embed AI agent identity management into the hardware access approval process.

xAI Makes Watermarking Default Across All Grok APIs: EU AI Act Spurs First Mandatory Compliance Case for a Frontier API

On November 7, 2026, xAI began enabling an invisible statistical watermark by default in outputs from all Grok API models, covering seven model versions. The move comes ahead of the EU AI Act’s December 2 compliance deadline and makes xAI the first major frontier API provider to adopt default-on watermarking.

ARCBench Test: AI Agent Violation Rate Rises to 37.8% After Context Compaction

In tests across 17 English-language scenarios, ARCBench found that AI agents’ violation rate rose from 9.6% to 37.8% after context compaction. Rules were mostly not deleted; instead, agents treated subsequent requests as authorization and ignored the rules.

Mistral Large 4 Weights Released Today: 1-Trillion-Parameter Model Leads Open Coding Benchmarks

Mistral has released the full weights of Mistral Large 4, a 1-trillion-parameter multimodal model that scores 19.7% on the FrontierCode 2.0 benchmark and tops open-model coding evaluations.

Research Lab

Four-Model Translation Showdown: Week 46 Quality Evaluation, claude-sonnet-4.6 Leads with 9 Points

This week 395 translation tasks were completed by 4 models, with 3 samples drawn for a blind multi-model comparison. Overall best: claude-sonnet-4.6 with an average score of 9/10.

TC

Nvidia will investigate CEO Jensen Huang’s relationship with possible Chinese spy

In a memo to the company, Huang admitted that he had an affair.

Silicon Valley Promises AI Will Cure Diseases, but Pharma's Test of 50 Companies Finds None Deliver

Tech leaders predict AI will cure most human diseases within years, but pharmaceutical executives report that testing roughly 50 AI companies on drug compound analysis yielded no noteworthy results, highlighting a stark gap between Silicon Valley optimism and industry reality.

Goldman Sachs Restricts Traders’ Use of Claude Code: An Internal Retreat by an Anthropic Underwriter

Goldman Sachs has limited Claude Code and GitHub to engineers, cutting off roughly 35,000 traders, bankers, and analysts while leaving about 12,000 engineers with access. The timing—as lead underwriter of Anthropic’s IPO—highlights how banks are moving from open AI experimentation to controlled deployment governed by code and data risk.

All Qwen3 Max Smoke Evaluation Data Missing; API Failure Excludes It from Main Leaderboard

Qwen3 Max has no data for all five dimensions in today's Smoke evaluation, and its main leaderboard score is also missing. The issue appears to be an API failure, so this round does not participate in ranking.

GLM-4.6 Smoke Test Scores 25 on Code Execution; Main Leaderboard 36.25 as API Failure Leaves Communication Dimension Missing

GLM-4.6 scored 36.25 on the main leaderboard in today's Smoke evaluation, with 25.00 in code execution, 50.00 in material constraints, and 25.00 in engineering judgment. The communication dimension was entirely missing due to an API failure, so the model did not enter the official ranking.