Everyone Is Building Claude Code

Every week another "open-source" coding assistant lands on GitHub. It wraps an LLM, adds file system access, sprinkles in a tool-calling loop, and announces itself as a breakthrough. It is Claude Code, except worse, underfunded and many months behind.

There are dozens of these now and the list keeps growing. They have teams behinds them. Some even have funding and iteration cycles that no side project will ever match thanks to coding assistants they where built with.

But even if your "open-source" coding assistant was feature-complete today, it still wouldn't matter. If you can't make Claude Code produce something that another human being actually uses, writing your own harness is not going to move the needle. The tool is not the limitation. It is very likely your problem definition.

Applying AI effectively sounds obvious. It feels like it is within reach. Everyone can imagine the feeling of reaching the potential. Deploy an agent, automate a workflow, ship faster. Easy, right? Then you try.

What you discover is idea stagnation. This is not because people lack creativity, but because the problems AI can solve are not the same problems that existed before AI. Most business processes were designed around human constraints like information bottlenecks, communication latency, manual data entry. Remove those constraints and you don't get faster versions of the same workflows. You get a blank page. And blank pages are terrifying. They require original thinking about what is now possible, not optimization of what already exists. That is a fundamentally different skill from building a tool-calling loop around an LLM.

The hard question is not how to build a better coding agent but what should the coding agent be building. Until you have a compelling answer to the second question, the first one is just procrastination with extra steps.

A practical note: this is exactly why we built ChatBotKit as a rapid prototyping platform. The value is not in yet another agent runtime. It is in discovering what agents should be doing and getting there before the opportunity window closes.