Human Creativity Is the Real AI Breakthrough

OpenAI deserves credit for building the foundation of LLMs. But they followed the typical chat-based app approach, and that limits what you can do. The chat-only frame of mind.

The engineers who are actually pushing things forward are the ones who master how LLMs work. They understand their strengths, their weaknesses, and get creative building around them. Claude Code, Lovable, OpenClaw, the Ralph Method and endless others. It's all driven by engineering creativity. Better prompting, smarter context window management, automating the boring parts of interacting with LLMs. Creative engineers building great software on top of what already exists.

The use of LLMs is getting more efficient not because the models are getting better, but because human creativity is driving the real innovation around these tools. That's what's exciting.

This is why I'm not worried about software engineers becoming obsolete. The tools keep getting better because humans keep getting creative. Not all is doom.