Anthropic launches Claude Tag, sparking debate over AI‑driven software development

Anthropic introduced Claude Tag, a beta feature that embeds its Claude AI model directly into Slack channels as a persistent teammate rather than a separate chatbot. Available now for Claude Enterprise and Claude Team customers, the tool lets users tag @Claude to assign tasks; the AI breaks the work into steps, accesses tools, and posts results back in the thread. Claude Tag operates in an "ambient" mode that monitors channels, flags relevant information, and follows up on forgotten threads without direct prompts. The company says the internal version of Claude now writes about 65 % of its product team’s code, including most of the code that built Claude Tag itself. The feature replaces the existing Claude in Slack app, which will be discontinued on August 3, and Anthropic plans to expand the integration to other platforms in the coming weeks.

An accompanying commentary warns that such AI‑generated code, sometimes called “vibe coding,” is being used to build full‑scale production apps without developers understanding the underlying architecture. While AI can produce clean code for well‑defined problems, the lack of human oversight can create security, scalability, and maintenance risks. The piece argues that even with spec‑driven development, a skilled software engineer is still needed to validate AI output before shipping software that users rely on.