I Gave Up on No-Code Workflow Engines (malvik.de)

The article discusses the author's disillusionment with no-code AI workflow engines like n8n, OpenAI Agent Builder, Make, and Zapier, which, despite their promises, still present complexities that can overwhelm users. While they work well for simple workflows, the need for handling conditions and errors significantly complicates the setup process, requiring a depth of knowledge that can be a barrier for many non-programmers. The author outlines an ambitious project to create a "second brain" using these tools, but ultimately finds them frustrating due to connectivity issues and insufficient feedback mechanisms. In contrast, after switching to coding with AI assistance from Claude Code, the author experienced a smoother process for implementing the same functionalities. The interaction with Claude allowed for clearer specifications and planning, ultimately leading to a more effective and gratifying solution. This reveals a significant implication for the AI/ML community: as AI coding assistants gain sophistication, they may prove to be more accessible and efficient than traditional no-code tools, potentially reshaping the approach organizations take towards automating workflows. The author expresses skepticism about investing in workflow engines given these challenges, highlighting a critical discussion point within the community about the true effectiveness of no-code solutions versus coding with AI support.