Mistral releases open weights for trillion-parameter Mistral Large 4
Mistral open-sources a 1T-parameter model aimed at cybersecurity, plus OpenAI's Model Spec revision, Microsoft buyouts, and an Nvidia spy-plot probe.
In this digest
AI News Feed
Page 2 of 19
Mistral open-sources a 1T-parameter model aimed at cybersecurity, plus OpenAI's Model Spec revision, Microsoft buyouts, and an Nvidia spy-plot probe.
In this digest
A pruned Nemotron speech model runs on consumer hardware, LIMe cuts LLM training compute ~15%, and Arizona's data centers strain water supplies.
In this digest
A Qwen-code pull request suggests Qwen4 models will ship with a 1M-token context window.
Read digestGoogle takes AI Mode ads international as Search ad growth cools to 12% and Alphabet shares slide 8% after Q3 earnings.
Read digestMicrosoft's FY2027 Q1 revenue hit $92.3B, up 19% year over year, with Azure growth of 48% and EPS beating estimates.
Read digestOpenAI models and agents are coming to ServiceNow products and workflows, with ChatGPT, Codex, and joint security capabilities on the roadmap.
In this digest
Anthropic ships reporting tools that show enterprises what their Claude spending produced, plus pooled group budgets for teams.
Read digestGoogle launches a developer platform for app integration with Gemini and Search, while its AI Mode dodges 2020 election questions.
In this digest
A Gemini 4 checkpoint codenamed Carbon is said to match Claude Opus 5.5 at coding, plus Chrome's Gemini autofill and GitHub's Octoverse 2026.
In this digest
A fresh wave of AI-linked cyberattacks hits South Korean brokerages and Bithumb, while Meta may delay its controversial business terms again.
In this digest
Put practical data boundaries and usage policies in place before AI makes existing oversharing easier to find.
Explore Secure AI Adoption