Mark Seemann reflects on his recent series of articles critical of AI in software development, acknowledging he may be engaging in motivated reasoning — a cognitive bias where conclusions align with personal desires. He admits most of his concerns about LLMs may be irrelevant as AI advances rapidly, but argues that even one valid concern justifies raising them. He also mourns the loss of programming as a craft, noting that managing AI agents holds no appeal compared to writing code directly. He closes by acknowledging economic incentives play a minor role in his skepticism, while maintaining that counterpoints to techno-optimism still have value.
Sort: