【文章标题】：How much of HN is AI?
【文章标题】：HN 上有多少内容是 AI？
I have a complicated relationship with Hacker News. The site is the most important aggregator of geek news and a major source of traffic to this blog. At the same time, it has a fair number of toxic commenters, making it a dependable source of insults hurled in my general direction; if you want a taste, this article has been called “watered-down” and “slop”.
我和 Hacker News 的关系很复杂。这个网站是极客新闻最重要的聚合器，也是本博客流量的主要来源之一。与此同时，它也有不少有毒的评论者，使其成为朝我这边抛来侮辱的可靠来源；如果你想感受一下，这篇文章就曾被称作“注水”和“垃圾”。
The site is run by geeks and for geeks, so it’s not immune to tech trends; for example, around 2018, it had a fair number of stories focused on cryptocurrencies and NFTs. That said, the recent shift feels more profound: almost every day, it feels that the lineup is dominated by stories focused on AI, written by AI, or commented on by AI.
这个网站由极客运营，也为极客服务，所以它并不能免于科技潮流的影响；例如，2018 年前后，它就有不少聚焦加密货币和 NFT 的报道。话虽如此，最近的变化感觉更深刻：几乎每一天，首页内容似乎都被聚焦 AI、由 AI 撰写或有 AI 参与评论的报道所主导。
That images shows a particularly bad day, so to give a more honest assessment, I also performed a more systematic survey in February 2026, and again in June of the same year.
那张图片显示的是特别糟糕的一天，所以为了给出更诚实的评估，我还在 2026 年 2 月进行了一次更系统的调查，并在同年 6 月再次进行了调查。
Original February 2026 investigation
2026 年 2 月的原始调查
To get a sense of how much of the feed is occupied by AI-related topics, I took a sampling of the daily top #5 for all of February:
为了了解信息流中有多少内容被 AI 相关话题占据，我对 2 月整个月每天排名前 5 的内容进行了抽样：
AI took four out of five spots on Feb 4 and Feb 12, plus arguably the entire line-up on Feb 5 (story #3 was submarine marketing for an AI vendor). The only days without LLM news in the top 5 were February 1 (with the first AI story at #7, then #9), February 9 (first at #8), and February 25 (with AI at #6, #9, #10).
2 月 4 日和 2 月 12 日，前 5 名中有 4 个位置被 AI 占据；2 月 5 日可以说整个榜单都是 AI（第 3 条是某 AI 供应商的隐性营销）。前 5 名中没有 LLM 新闻的日子只有 2 月 1 日（第一条 AI 新闻在第 7 位，然后是第 9 位）、2 月 9 日（第一条在第 8 位）和 2 月 25 日（AI 分别在第 6、第 9、第 10 位）。
For the second part of the experiment — figuring out which stories were likely AI-written — I tapped into Pangram. Pangram is a remarkably good, conservative model for detecting LLM-generated text. These detectors have bad rap among techies, but the objections are often based on outdated assumptions or outright misconceptions. For the tools to work, AI writing doesn’t need to be in any way “inhuman”. It’s enough that the default voice of the current crop of LLMs is quasi-deterministic: ask for the same essay twice and you’ll get a stylistically similar result. The individual mannerisms are human-like, but it’s very unlikely that your writing combines the exact same set. I write about it a bit more here.
在实验的第二部分——弄清楚哪些报道可能是 AI 写的——我使用了 Pangram。Pangram 是一款非常好、偏保守的 LLM 生成文本检测模型。这类检测器在技术人员中口碑不佳，但反对意见往往基于过时的假设或彻底的误解。要让这些工具发挥作用，AI 写作并不需要以任何方式“非人化”。只要当前这批 LLM 的默认语气是准确定性的就够了：让它们写同一篇文章两次，你会得到风格相似的结果。个别的行文习惯很像人类，但你的写作不太可能恰好组合出完全相同的一套特征。我在这里对此有更多论述。
To validate the results, I also reviewed all the flagged stories and I think the findings make sense; if anything, Pangram had a couple of false negatives. To give you a sense of what was flagged, have a look at the #3 story on February 19 (“AI is not a coworker, it’s an exoskeleton”). It had 500+ upvotes and 500+ comments. In my opinion, it has a wide range of red flags.
为了验证结果，我还复核了所有被标记的报道，我认为这些发现是合理的；如果说有什么问题，Pangram 反而出现了几例漏报。为了让你感受一下被标记的内容，可以看看 2 月 19 日排名第 3 的报道（“AI 不是同事，而是外骨骼”）。它有 500 多个赞和 500 多条评论。在我看来，它有很多危险信号。
Updated data for June 2026
2026 年 6 月的更新数据
In June, to capture more detail, I used solid black for pure-play AI navel-gazing (vendor announcements, op-eds about the benefits or drawbacks of the technology, etc) and hatched shapes for stories that lean heavily into AI, but have broader ramifications (e.g., the Instagram AI support agent account hack). As before, stories that are only tangentially related to AI (e.g., reports of RAM price hikes) are not flagged.
6 月，为了捕捉更多细节，我用纯黑色表示纯粹的 AI 自我陶醉式内容（供应商公告、关于该技术利弊的评论文章等），用阴影形状表示大量依赖 AI 但具有更广泛影响的报道（例如 Instagram AI 客服账号被黑事件）。和之前一样，只是略微涉及 AI 的报道（例如 RAM 价格上涨的新闻）不会被标记。
In the first half of the month, roughly 60% of the daily HN lineup was AI-related or AI-generated, tapering off to ~50% as we approached the end of the month. This is up from 40% in February.
在 6 月上半月，HN 每日内容中大约 60% 与 AI 相关或由 AI 生成，临近月底时逐渐回落到约 50%。这比 2 月的 40% 有所上升。