header-langage
简体中文
繁體中文
English
Tiếng Việt
한국어
日本語
ภาษาไทย
Türkçe
Scan to Download the APP
Partnered With:

Claude Tag Integrates with Slack: Internally Rolls Out 65% of the Code, Karpathy Labels it as Third-Generation Paradigm

According to DeepBeat Monitoring, Anthropic has launched a new generation of collaborative AI named Claude Tag and has initiated the public beta testing on the Slack platform, enabling enterprises and team users to incorporate AI as a persistent, asynchronous collaborative member into their workflows. The new feature is currently in the beta testing phase, open to Claude Enterprise and Claude Team subscription users, transforming the traditional one-on-one conversation model, allowing multiple individuals to collectively interact with the same AI collaborator within a channel.

Unlike traditional chatbots with one-way responses, Claude Tag can seamlessly integrate into Slack channels, supporting autonomous planning and asynchronous execution. Users mention @Claude in the channel and submit tasks, the model automatically breaks down the task into multiple stages, leverages connected code repositories, databases, or external tools for processing, and responds with the results in a Slack thread upon completion. The model also offers an environment-aware mode, which continuously tracks the progress of channel discussions, proactively provides information for relevant discussions, or follows up on threads that have seen no progress for an extended period.

In terms of security management, channel administrators can configure independent tool calls and data access permissions for different channels, ensuring model memory isolation and supporting global and per-channel Token consumption limits. Currently, Claude Tag has been extensively used internally at Anthropic, with 65% of the product team's code being generated by the internal version. Former Tesla AI Director Andrej Karpathy commented that this represents the third-generation paradigm of large language model interaction UI/UX, as AI evolves from early web-based conversations and local client tools to a persistent, asynchronous collaborative entity with organizational-level tools and context.

举报 Correction/Report
Correction/Report
Submit
Add Library
Visible to myself only
Public
Save
Choose Library
Add Library
Cancel
Finish