📄 中文摘要
关于大型语言模型（LLMs）在编程中的应用，存在一个反复出现的担忧，即这些模型可能会使我们的技术选择趋向于那些在其训练数据中表现良好的工具，从而使新兴的、更优秀的工具难以脱颖而出。几年前，当询问模型有关Python或JavaScript的问题时，得到的结果明显优于较少使用的语言。然而，随着最新模型在良好的编码代理环境中运行，这种情况似乎并没有持续下去。最近，我在使用一些全新工具时，得到了非常出色的结果，表明这些工具的潜力可能被低估了。
或许并非无聊的技术
出处: Perhaps not Boring Technology after all
发布: 2026年3月9日
📄 English Summary
Perhaps not Boring Technology after all
A recurring concern regarding large language models (LLMs) in programming is that they may steer technology choices towards tools that are well-represented in their training data, making it challenging for new and better tools to gain traction. This was particularly evident a couple of years ago when asking models for help with Python or JavaScript yielded significantly better results than inquiries about less commonly used languages. However, with the latest models operating in effective coding agent environments, this trend may no longer hold. Recent experiences with brand new tools have shown excellent results, suggesting that the potential of these tools might have been underestimated.