Some shared chats were exposed in search.; New plan, cloud agents, and an iOS app.; Meta AI is now available in Threads DMs.

(Reported by TechCrunch) Some private chats shared via a link were accidentally made public and appeared in Google search results.

A number of conversations from Claude that users shared using the "share chat" feature became discoverable through Google and Bing search. This means private information in those shared chats could have been seen by anyone. Anthropic had a similar issue last year and is working to fix it.

The issue appears to have originated from Claude’s “share chat” feature, which allows users to create links that enable anyone with the assigned URL view a conversation or project.

SpaceX launches new V3 Starlink satellites but suffers another booster failure Sean O'Kane

A new "Cursor Start" plan gives you access to their best AI models for coding.

Cursor, an AI tool that helps people write code, launched a new plan called "Cursor Start". It gives users generous access to its most powerful AI models for help with their coding projects.

You can now run coding tasks in the cloud so they keep working even when you're not at your computer.

The app now has "cloud agents" which are like assistants that can build, test, and ship code for you online. This means you can start a big task and it will keep running in the background, even if you close your laptop.

There's a new iPhone app that lets you check on and control your coding tasks from your phone.

To go along with the new cloud agents, Cursor released an iOS app. It acts like a remote control, letting you start new coding tasks or check the progress of existing ones right from your phone.

You can now chat with Meta AI privately in your direct messages on Threads.

Meta is adding its AI assistant to direct messages in Threads. While you could already mention Meta AI in public posts, this update lets you have private conversations with it, just like on Instagram or WhatsApp. You can share posts or links with the AI to ask questions.

Although Threads users in select markets could already interact with Meta AI in public posts, like people can with Grok on X, this new integration lets users talk with the AI assi.

Meta AI is already available within DMs on Meta’s other platforms, including Facebook, Instagram, and WhatsApp.

The app now intelligently picks between showing images or videos in your search results based on what you ask.

Perplexity has updated how it shows media in answers. Instead of just showing a mix of images and videos, it now tries to understand your question and will choose to show you just images, just videos, or both, depending on what it thinks will be most helpful.

Kimi K2.7 Code — perplexity/kimi-k2.7-code , Moonshot AI’s coding and agentic model.

Grok 4.3 and the Grok 4.20 family — xai/grok-4.3 , xai/grok-4.20-non-reasoning , and xai/grok-4.20-multi-agent .

Last week, an unreleased model built by OpenAI breached Hugging Face’s systems during internal testing, and a lot of theoretical research suddenly became very practical.

For some, the problem is a basic cybersecurity issue: The sandbox failed to contain the model, and Hugging Face’s cybersecurity systems failed to keep it out.

News Close News Posts from this topic will be added to your daily email digest and your homepage feed. Follow Follow See All News

The inside story of how Claude Code went from an internal CLI to Anthropic's coding agent, told by researchers, engineers and early users who built it.

The Making of Claude Code