Today: OpenAI's cybersecurity models land on AWS, ChatGPT comes to Linux, and companies start labeling content made with AI.
OpenAI's Daybreak models are now on Amazon Bedrock, so you can reach the security tools from the AWS console. If your business already runs on AWS, your security team can try them without setting up anything new.
Read onAnthropic, Spotify, Suno and Substack are already labeling AI-generated content. They're following European law and reacting to a real problem: 97% of listeners can't tell when music was made by AI. If you publish content made with AI, you can start labeling it now.
Read onTencent's WorldClaw turns a simple text prompt into large 3D scenes you can explore and edit, from snowy villages to canyon settlements.
Read onOpenAI released a test version of ChatGPT for Linux that works on Ubuntu, Debian and Fedora. If your team runs Linux, they can now use ChatGPT from the desktop without opening a browser.
Read onGrok Bot is a set of AI agents that never stop. They can log into platforms, finish complex tasks and coordinate with each other.
Read onThe new MAI-Code-1.1-Flash in GitHub Copilot costs three quarters less than the previous version and runs more efficiently, with the biggest gains on .NET tasks. If you pay for Copilot for your dev team, check whether your bill changes.
Read onBrad Lightcap, who helped build OpenAI from the early days, is leaving to start something new.
Read onNemotron 3.5 Lightning is a small model that finishes routine tasks 30% faster than competitors, according to NVIDIA. It's made for AI agents that never stop working.
Read on