The people who most need AI for their busywork are the least likely to use it well
热度趋势
百分比基于当前可用热度信号，而非评论数或独立用户人数。
人们发现，那些被重复性文档和演示工作所困扰，最能从AI工具中受益的人，往往最不愿采用这些工具，即使采用了也只发挥了其能力的极小一部分。这种现象部分源于对界面的不适应，也因为有效使用AI工具本身就是一项技能，更像是向人委派任务而非简单的搜索。用户需要知道如何提问、如何核查结果以及何时放弃。那些不习惯清晰表达任务的人往往难以获得良好结果，最终得出工具被过分吹嘘的结论。
An adoption gap I keep running into: the people drowning in repetitive document and presentation work, the ones who would benefit most from offloading it, are often the last to actually adopt these tools, and when they do, they use them at maybe ten percent of what is possible.
Meanwhile the people fluent with the tools tend to already be efficient and technical, so the marginal gain for them is smaller. The result is that the productivity boost lands unevenly, and not in the direction the "this helps everyone" story predicts.
Part of it is comfort with the interface. Part of it is that using these tools well is itself a skill nobody teaches, closer to knowing how to delegate to a person than to running a search. You have to know what to ask for, how to check it, when to throw it out. Someone who has never had to articulate a task clearly struggles to get a good result, gets a mediocre one, and concludes the tool is overhyped.
So the gap is not really about access anymore, the tools are cheap and everywhere. It is about the skill of using them, which is unevenly distributed and largely invisible. That gap seems more durable than the price one.
For people who have tried to get a less technical colleague to actually adopt this stuff for their routine work: what actually worked, if anything?