English(EN) We probably just expected too much out of LLMs, entire teams of well-trained teams can be capable of writing entire code bases of slop, why would we expect much

作者声称大型语言模型（LLM）写的代码不会比人类写的更差

作者 PulseAugur 编辑部 · [1 个来源] · 2026-09-07 05:07

作者认为，鉴于人类团队编写糟糕代码库的历史，期望大型语言模型（LLM）编写出完美代码是不现实的。他们以《军团要塞2》（Team Fortress 2）的代码库以及应用商店的普遍状况为例，证明了人类编码的不足。文章认为，LLM编写的代码可能不会比现有的人类生成代码更差，而劣质代码的普遍存在是一个自找的麻烦。 AI

影响表明当前的大型语言模型（LLM）与现有的人类生成软件相比，在代码质量上可能不会显著恶化。

排序理由一篇评论文章，讨论了大型语言模型（LLM）在编码方面的能力和局限性，并与历史上人类的编码实践进行了类比。

在 Mastodon — mastodon.social 阅读 →

观点

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 2 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Commentary; 一篇评论文章，讨论了大型语言模型（LLM）在编码方面的能力和局限性，并与历史上人类的编码实践进行了类比。
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: opinion; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

Mastodon — mastodon.social TIER_1 English(EN) · [email protected] · 2026-09-07 05:07

我们可能对LLM的期望过高了，训练有素的整个团队都能写出垃圾代码，我们还能期待什么

We probably just expected too much out of LLMs, entire teams of well-trained teams can be capable of writing entire code bases of slop, why would we expect much better from anything we train, even if was magically deterministic? Just ask anybody who has worked on the team fortres…

本文如何被排名

报道来源 [1]

我们可能对LLM的期望过高了，训练有素的整个团队都能写出垃圾代码，我们还能期待什么

相关实体

相关话题