AI writes pretty good code these days and it doesn't really matter

The author critiques the overblown hype surrounding AI in software development, arguing its utility is limited compared to other tools. While acknowledging AI's potential, they emphasize its environmental and societal costs, questioning its worth given the marginal efficiency gains. The author contends that generative AI's current capabilities are comparable to a middle schooler's work, struggling with complex codebases and context. Their daily enterprise development tasks involve meetings, reviews, refactoring, and documentation, areas where AI offers minimal assistance. Tools like language servers, detailed documentation, strong typing systems, and linters provide more substantial and consistent benefits. The author believes the focus on AI stems from an unfounded optimism about future AGI, overlooking its stagnation and present limitations. They recount personal experiences where AI hallucinated Jira tickets, showcasing its reliability issues. The disproportionate veneration of AI grants undue influence to companies like OpenAI, potentially leading to unethical practices. The author advocates for a more realistic assessment of AI's value, considering its costs and comparing it to other valuable developer tools.