I’m sick of AI “Thinkslop” in my PRs

The author, a mechanical engineer turned software developer, is frustrated with AI-generated code that looks clean but has logical flaws. To combat this, they've developed a tool called NexaVerify that runs code through multiple AI models to identify potential issues. This approach helps catch AI hallucinations and ensures code reliability. The author suggests treating AI as a junior developer prone to errors, rather than an oracle, and forcing it into a consensus through multi-model pipelines.