一位Hacker News用户发表评论，对大型语言模型（LLM）在软件开发中的实际价值提出了质疑。该用户指出，尽管AI革命已经进行了4年，但整个行业在软件质量、成本、安全性或创新方面几乎没有取得显著改善[1]。他表示，在投入1.5万亿美元后，业界仍然缺乏独立研究来证实AI相关的顶层生产力提升[1]。
在代码质量问题上，该用户指出LLM生成的代码中真正值得合并的比例仍然远低于人类编写的代码[1]。他还提到，前沿模型曾遗漏明显的问题，后来被业余爱好者发现[1]。该用户认为，推进LLM应用背后存在隐性目标，即使智力劳动商品化，从而削弱工人的议价权[1]。
A Hacker News commentator has voiced sustained doubt about the tangible effectiveness of large language models in the software development industry [1]. Despite four years of what has been termed an "AI revolution," the sector has demonstrated minimal concrete gains in software quality, costs, security, or innovation [1]. As the commentator stated, "We are 4 years into this 'revolution' and the industry has almost nothing to show for it" [1].
The critic contends that despite $1.5 trillion in investment, independent research substantiating significant productivity improvements from AI remains scarce [1]. "We still have almost no independent studies that can attest to top-level productivity improvements associated with AI," the commentator noted [1]. Quality concerns plague AI-generated code, with a substantially lower proportion of machine-written code deemed suitable for merging compared to human-authored work [1].
The commentator also notes instances where frontier models overlooked obvious issues that amateur developers subsequently identified [1]. Beyond technical shortcomings, the author suggests an underlying agenda behind AI advancement: the commodification of intellectual labor and the erosion of workers' bargaining power [1].