LLMs Can't Program

The article argues, as its central thesis, that large language models cannot truly program. Based on the supplied metadata, it appears to be a critical discussion of current LLM capabilities in areas such as code understanding, generation, debugging, or reliable execution. However, the available material contains only the title and a Hacker News listing, with no article details, author information, experiments, benchmarks, or evidence. Its claims should therefore be read as an opinion or thesis requiring verification against the original text and supporting data.