提问 HN：如何在人际沟通（h2h）中平衡 AI 的使用？
1 分•作者: tzury•5 个月前
我们都在工作流程中使用 LLM，这是事实。
然而，当收到明显由 LLM“润色”过的邮件，或者说，在 HN 上看到由 LLM“改写”的评论时，我们都会感到一种本能的沮丧。
为什么我们会对别人做我们都在做的事情而感到不快？
是因为它用得太明显了吗？
现在，我都不像以前那样修改我的写作了，这样它才能保持“真实”（留下一些小拼写错误，也许还有一些语法错误——这对于非英语母语者来说很常见）。
毫无疑问，我们将要处理越来越多的 AI 生成内容，这里的“我们”指的是我们人类和我们的代理，问题是什么时候才能达到平衡？什么时候我们才能都接受它是一个事实？
这是否与我们愿意为“手工制作”的商品支付额外费用有关？
（打个比方：很明显你的机器制作了我的鞋子）
（*）也许除了我们中的 99.999% 之外。
——
为了讽刺，下面是以上内容的“由 Claude 润色”的版本
——
ASK HN：平衡人际交流中的 AI 使用
我们都在工作流程中使用 LLM。这并不是一个有争议的说法。
然而，当你收到明显由 LLM“润色”过的邮件，或者阅读明显由 AI 改写的评论时，几乎所有人都会感到恼火。这种沮丧是真实的，虚伪也是如此。
是它太明显了吗？缺乏努力？还是更令人不安的是——我们认出了面具，因为我们自己也戴着它？
现在，我已经停止修改自己的写作了，故意保留小的拼写错误和语法怪癖。这对于非母语人士来说很常见。我保留它们作为存在的证明。
AI 生成内容的数量只会增加。这不仅仅来自人类对此做出选择——也来自代表我们行事的代理，相互交流的代理。什么时候这种沮丧感会正常化？我们什么时候才能像接受电子邮件或自动更正一样，简单地接受它作为基础设施？
手工制作的商品中有一个有用的类比。我们知道这双鞋是机器制造的。我们仍然为不是机器制造的鞋子支付更多费用。溢价与鞋子无关——它与信号有关。证明有人选择花费时间。
也许这才是我们真正反应的：不是 AI 辅助本身，而是信号的缺失。
查看原文
It is a fact that we are all* using LLM throughout our workflows.
And yet, there is an automatic frustration when one's receiving an email that was clearly "polished" by an LLM, or say, a comment here on HN that was "rephrased" by an LLM.<p>Why are we getting upset when someone is doing what we all do?
Is it the obvious (over) use?<p>It got to a point where I am not revising my writing as I used to, so it will remain "authentic" (leaving minor spelling and perhaps some grammatical errors - common for non native English speakers).<p>there is no questions we are going deal with growing amount of ai generated content, and by we it is us, humans, and our agents, and the question is when will it balanced out? when we will all accept it as a fact?<p>Is it related to the fact we are willing to pay extra for "hand crafted" goods?
(as an analogy: it is obvious that you had your machine generated my shoe)<p>(*) perhaps another all bu 99.999% of us.<p>--
For the irony in it, below is the "Polished by Claude" version of the above
--<p>ASK HN: Balancing AI use in human-to-human communication<p>We're all using LLMs throughout our workflows. That's not a controversial claim.
And yet there's a near-universal irritation when you receive an email that was clearly "polished" by one — or read a comment that was obviously AI-rephrased. The frustration is real, and so is the hypocrisy.<p>Is it the obviousness of it? The lack of effort? Or something more uncomfortable — that we recognize the mask because we wear it ourselves?
It's gotten to the point where I've stopped revising my own writing, deliberately leaving minor spelling errors and grammatical quirks intact. Common for non-native speakers. I keep them as proof of presence.<p>The volume of AI-generated content will only grow. Not just from humans making choices about it — but from our agents, acting on our behalf, communicating with each other's agents. At what point does the frustration normalize? When do we simply accept it as infrastructure, the way we accepted email or autocorrect?<p>There's a useful analogy in handcrafted goods. We know the shoe was machine-made. We still pay more for the one that wasn't. The premium isn't about the shoe — it's about the signal. Proof that a human chose to spend time.<p>Maybe that's what we're really reacting to: not the AI assistance itself, but the absence of the signal.