# Doubts About AI Set In > Source: > Published: 2026-08-29 07:00:00+00:00 Before I designed my own **agentic AI system** (an MCP server) I tried to keep an open mind about AI. I mean, how can you say something is not worthwhile unless you have tried it? I eventually made an AI system that can **translate** certain natural language phrases into command-lines. My system works well. It is fast enough that it doesn't even need a GPU; it tends to **accurately call tools** based on what I tell it. I prefer using my agentic AI system over directly entering the command lines. It also reduces errors in the process. But: • I don't feel it adds much value (if any) over the old system, even though it is preferable. • It may be that the command line tasks I had were not a good candidate for AI automation. So in this enormous AI bubble I have managed to develop a system that adds nearly **zero value** to my life. It seems possible that a **large corporation** would have more tasks to automate with AI that are much more valuable. It may also be that things like **agentic coding** are the most valuable uses of AI. In any case, I feel increasingly **doubtful** about the AI era we are supposed to be entering into.