The Promises and Perils of LLMs in Building End User Software - Jonas Erthal - Blog

The author argues that while large language models have benefited application development through speed and pattern matching, irresponsible adoption has led to software bloat, severe context switching, and declining quality. Without intentional values and guardrails, heavy reliance on AI risks de-skilling developers and producing mediocre end-user software. The piece suggests companies establish principled stances on AI use similar to infrastructure pioneers. Open article →