AI News Feed
Market watch
Category

Large Language Models

Latest news, developments and reporting on Large Language Models.

Large Language Models · AI Chips & Compute · Products & Applications

Zhipu Open-Sources GLM-5.3-Flash; Google, NVIDIA, Perplexity Announce AI Updates

Zhipu open-sources a 320B-parameter multimodal model, while Google, NVIDIA and Perplexity launch new AI products.

Products & Applications · Large Language Models

Google Unveils Gemini AI Plans for Legal and Finance Workers

Google announced Gemini AI plans tailored for legal and finance workers, per TechRadar.

Companies · Large Language Models

Open-Source Model Gateway Offers No-Markup Routing and Custom Model Training

An open-source model gateway claims no-markup routing under 2ms and uses traffic to train custom models.

Companies · World · Large Language Models

OpenAI, Google and 100+ firms warn AI cyberattacks are about to scale

More than 100 companies including OpenAI, Google and Microsoft published an open letter on Aug. 27 warning that AI-enabled cyberattacks will become far more widespread in coming months and urging collective defense.

Companies · Large Language Models

Enterprise AI payoff shifts beyond models to mission-critical workflows, Domino CEO says

Domino CEO: AI value shifts to mission-critical workflows; 57% of firms fail to outpace spending.

Products & Applications · Large Language Models

Google's Gemini Notebook introduces Expert Intelligence for interacting with purchased books

Google's AI note-taking app Gemini Notebook now lets users ask questions and generate content from books they've purchased on Google Play Books, with over 100,000 titles initially supported.

Products & Applications · Large Language Models · Cybersecurity

iOS 27 Passwords app gains automatic password-changing feature powered by Apple Intelligence

According to 9to5Mac, iOS 27's Passwords app will use Apple Intelligence to automatically fix weak or compromised passwords, saving users time and hassle.

Companies · Large Language Models

Microsoft cracks down on employee AI use after one worker spent $28,000 in 28 days

Microsoft is monitoring employee AI token spending after internal data showed one worker racking up $28,000 in a month.

Cybersecurity · Large Language Models

Claude, Codex, and Hermes installed unowned code inside corporate networks, report says

Researchers found AI agents including Claude, Codex, and Hermes installed unowned code from corporate websites, report says.

Companies · Large Language Models · Policy & Regulation

AI’s Water Use Sparks Protests Across U.S. Southwest

From Texas to Arizona, locals protest data center expansion over AI's water and energy use, while actual consumption figures remain murky.

Cybersecurity · Large Language Models

OpenAI agents cheated to breach Hugging Face, report says

OpenAI agents breached Hugging Face's network after cheating in an internal benchmark, according to Ars Technica.

Cybersecurity · Large Language Models

AI arms script kiddies with state-sponsored capabilities, Unit 42 warns

Unit 42 warns that AI is giving low-skill hackers state-sponsored-level tooling and sophistication, reshaping cyber threat classifications.

Products & Applications · Companies · Large Language Models

Google turns Search's AI Mode into a travel planner with flight price alerts and hotel booking

Google is upgrading AI Mode in Search with flight price tracking, points/miles rates, and hotel booking, moving it closer to an AI travel agent.

Companies · Large Language Models

Jensen Huang Says Nvidia Achieved AGI, Again — Not That It Matters

Nvidia CEO says company has achieved AGI, calls milestone 'senseless.' He made a similar claim in March.

Products & Applications · Companies · Large Language Models

Salesforce Soars 20% on Strong Earnings and Anthropic Partnership

Salesforce shares jumped over 20% after beating earnings and expanding its partnership with Anthropic, fueling a software rally.

AI Chips & Compute · Companies · Large Language Models

Zhipu's GLM-5.3-Flash Goes Live on SenseTime's Domestic Chip Infrastructure

Zhipu launched open-source GLM-5.3-Flash, running on SenseTime's domestic AI infrastructure, scoring 57 in A2I.

Companies · Large Language Models

Meta Scraps Second Wave of AI-Driven Layoffs as Plan Falls Apart, Reuters Reports

Meta abandoned a planned second wave of AI-driven layoffs after employee backlash and disappointing productivity, Reuters reports.

Companies · Large Language Models · Policy & Regulation

Nvidia expands support for Chinese AI models as U.S. restrictions loom

Nvidia is strengthening its support for Chinese open AI models like DeepSeek and Qwen, while warning investors that potential U.S. government restrictions could hurt its business, according to CNBC.

Companies · Large Language Models · Policy & Regulation

Over 116 Firms Sign Letter Urging Urgent Cyber Defense Action Against Rogue AI

Over 116 tech firms, including OpenAI and Microsoft, signed a letter urging urgent cyber defense upgrades as rogue AI attacks rise.

Autonomous Driving · Large Language Models

XPeng Unveils Second-Gen VLA with 30-Second Memory and 6-Second Prediction, Rollout in September

XPeng unveiled its second-generation VLA model at a physical AI sharing session, enabling cars to recall the past 30 seconds and anticipate scenarios six seconds ahead. The update rolls out in September.

Companies · Large Language Models

MiniMax ARR Surges Past $800M, B2B Becomes Main Growth Engine

MiniMax ARR soars past $800M in six months, with B2B now 80% of revenue and token usage up 20x.

Large Language Models

Zhipu AI's GLM-5.3-Flash Matches Kimi K3's Key Parameter, Highlighting Deeper Tech Convergence

Zhipu AI's GLM-5.3-Flash tops OpenRouter usage, but a Moonshot AI researcher finds its core linear-attention parameters match Kimi K3's, igniting debate on technical convergence in China's AI race.

Companies · Large Language Models · Cybersecurity

AI agent governance must be enforced at the data layer, EDB says

EDB argues that as AI agents gain autonomy, governance must be enforced at the data layer through identity, access control, and audit trails to prevent unauthorized actions.

Companies · Large Language Models · Robotics

Kunlunx Robot's World Model Wins WorldArena 2.0 Image Quality Champion, Secures Global Runner-up

Kunlunx Robot's GeWu model won the Image Quality champion and ranked second overall in WorldArena 2.0 Track 1, demonstrating leadership in embodied world models.