I changed one setting in Claude Code, and my token burn dropped by 45%

Anthropic has introduced a feature in Claude Code that allows different sessions to message each other with updates and information. This feature, available on macOS and Linux, enables users to share findings, ask questions, and coordinate work across multiple sessions. According to 9to5Mac, users can prompt Claude to send a message to another session, and Claude will write the actual message.

The update comes as users look for ways to reduce costs associated with AI tools. Some services offer lifetime access to multiple AI models at a discounted price, such as ChatPlayground, which is offering a lifetime subscription for $59.25, as reported by Mashable.

In a separate development, Anthropic has also changed the default mode for Claude Code Pro, Max, and Team users to "auto mode", which uses a classifier model to decide when human intervention is needed. This change aims to reduce the number of permission prompts that users have to approve, as research showed that humans approve 97% of prompts and only catch 13.6% of dangerous commands, according to The New Stack.

Brief written by urgent.news from XDA Developers, Mashable, Techmeme, 9to5Mac, The New Stack — 5 reports on this story. Machine-written — read the original for the full account.