本篇文章来源于https://mp.weixin.qq.com/s/Go5gvCiNZa6YqhU4ZEYouA,如有侵权请来信告知,本站将删除。13951431913#139.com
我是做产品的,平时干活习惯分屏——一块屏上常常叠着五六个窗口:文档、代码、聊天、浏览器。
日程和待办通常也只是其中一个窗口。于是每次想确认「今天还剩什么没做」「下午几点开会」,都得先把它从一堆窗口里翻出来,切过去瞟一眼,再切回来。一天下来,这套「切窗口—看一眼—切回来」重复很多次,每次都在打断手头正做的事。
干活时一块屏上常叠着五六个窗口,找一条信息得先切来切去
手边正好有台吃灰很久的旧 Kindle。某天冒出个念头:能不能让它专门当一块日历 + 待办屏,常亮摆在桌上,把这些高频要看、又不需要操作的信息挪出主屏——抬眼就能看到,不用再切窗口。顺手也算给这台闲置设备找个正经用处。
先交代一句:整个过程里的脚本和网页,我几乎没自己敲代码,都是把想法讲给 GPT,让它帮我写出来的。所以这与其说是技术改造,不如说是一次「动嘴指挥 AI 干活」的记录。
真正让我意外的是:这么一件纯属折腾的小事,最后竟照出了我另一个做产品的老毛病。这个放到最后说。
第一版:我想当然地把日历「原样搬」了过去
最开始的想法特别直接:我平时在 Notion Calendar 里看日历,那让 Kindle 也打开这个视图不就行了?
结果一上手就翻车。
Kindle 的浏览器能开网页,但它不是现代浏览器——屏幕小、刷新慢、交互弱,地址栏还占掉一块。
桌面上那个看着很舒服的月视图,搬到 Kindle 上就成了一堆挤在格子里的小字。7 列乘 5 行,每格再塞几个任务标题,最后就是四个字:能看见,但不想看。
这一版让我停下来想:问题不在 Kindle 太弱,而在我又一次默认「把所有东西都放上去」。
所以第一个结论是:
不要把桌面日历原样搬到 Kindle,要为 Kindle 重新决定「什么值得被看见」。
第二版:只留「今天 + 接下来两周」
我试了几种布局,最后定下来的版本很克制:
- • 顶部:今天的日期。
- • 中间:今天的事项,大字列表。
- • 下面:接下来 14 天,两行横向日期卡片。
- • 最底部:一个小月历,但不再塞标题,只标每天有几项。
它比完整月视图清楚太多了。因为 Kindle 上真正要回答的,从来不是「这个月每一天都发生了什么」,而是三个问题:
- • 今天有什么?
- • 接下来几天有什么?
- • 这个月哪几天会忙?
那一刻我有点恍惚:这不就是我写需求文档时,一直想做却总有时候会忽略的事吗——先想清楚用户需要回答什么问题,再决定放什么,而不是一股脑全铺上去。
数据源:不读 Calendar,读它背后的库
这里有个小坑。Notion Calendar 里看到的内容,其实来自好几个地方:Google 日历、任务库、项目库、会议库。
所以我没去「截图」或「复刻」Calendar 本身,而是换了个思路:
用 API 读日历背后的数据,再生成一个 Kindle 专用网页。
这部分我依然「动嘴不动手」:把要读哪几个库的 Notion 页面链接给到 GPT,抓取、合并、生成页面全交给它写脚本。目前接了三个库——项目上线跟踪库(客户名称 + 计划上线日期)、daily-todo(任务名称 + 截止日期)、会议库(名称 + 活动时间)。
说得再直白点:电脑当「厨房」把菜做好,Kindle 只当「餐盘」负责端出来。这样 Kindle 不用登录 Notion,token 不外泄,页面也够素,正好喂得动它那个老浏览器。
一些「够用就好」的取舍
更新方式:一开始想过每 15 分钟自动刷新,后来觉得没必要。我的日程不是每分钟都变,Kindle 也不是监控屏。于是改成手动触发——给 GPT 一句指令更新日历,重新生成页面,Kindle 刷新即可。需要时更新,不需要时它就安静待着。
屏幕常亮:肯定有人问,Kindle 不是很快休眠吗?我翻了设置,休眠时间最长只能设到 1 小时,系统没有「永不休眠」开关(除非越狱)。我的做法很简单:上限设 1 小时,最多 1 小时睡一次,点一下就醒,页面直接停在最新日历。对一块桌面日历来说,这已经够了。
真正被改造的,不是那台 Kindle
折腾完我才反应过来:有意思的从来不是「技术上能不能实现」,而是 Kindle 逼我回答了一个我职业里天天遇到、却常常糊弄过去的问题——
什么信息,值得被看见?
这问题我其实太熟了。这些年我写需求文档,有个改不掉的共同点:总想写全,生怕漏一个场景、少一个字段、缺一句边界说明,好像写得多才等于负责——这毛病我私下管它叫「信息贪心」。
桌面软件总鼓励你要更多:更多视图、更多颜色、更多标签、更多入口。可 Kindle 很慢、很素、很克制,它逼你把不重要的删掉,只留真正该被看到的。最后那块小屏幕上只剩三样:今天、接下来、本月忙闲。反而比原来的日历更像一个「生活仪表盘」。
这也是我这些年做产品、写文档一直在补的那一课:「多」从来不等于「好」,「准」才是。 判断力不体现在你放了多少,而体现在你敢删掉多少。
一台旧 Kindle,没越狱、没刷机,只用自带浏览器,就把这句话又敲了我一遍。
你手边有没有那种「还能用、却一直不知道拿来干嘛」的旧设备?评论区一起聊聊。
This article is from https://mp.weixin.qq.com/s/Go5gvCiNZa6YqhU4ZEYouA . If there is any infringement, please notify us via email, and the site will delete it. 13951431913#139.com
I work in product design, and I'm used to splitting screens when working—often with five or six windows stacked on one screen: documents, code, chat, browser.
Calendar and to-dos are usually just another window. So every time I want to check "what's left today" or "what time is the afternoon meeting," I have to dig it out from the stack, switch over, glance at it, and switch back. Over a day, this "switch—glance—switch back" routine repeats many times, each time interrupting what I'm doing.
When working, a single screen often has five or six windows stacked; finding one piece of information requires switching back and forth.
I happened to have an old Kindle that had been gathering dust for a long time. One day a thought occurred: could I make it a dedicated calendar + to-do screen, always on, sitting on the desk, moving those high-frequency, no-interaction-needed pieces of information off the main screen—visible at a glance without switching windows? It would also give this idle device a proper use.
A quick note: Throughout this process, I didn't write most of the scripts or web pages myself — I told GPT what I wanted, and it generated them. So this is less of a technical modification and more of a "talking to AI to get stuff done" story.
What really surprised me was that such a trivial tinkering project ended up exposing an old habit of mine as a product person. I'll talk about that at the end.
First version: I naively "copied" the calendar over as-is
The initial idea was very straightforward: I usually view my calendar in Notion Calendar, so why not just have the Kindle open that view?
It failed as soon as I tried.
The Kindle's browser can open web pages, but it's not a modern browser—small screen, slow refresh, weak interaction, and the address bar takes up space.
The monthly view that looks comfortable on desktop became a bunch of small text crammed into grids on the Kindle. 7 columns by 5 rows, with task titles stuffed into each cell—the result was four words: visible, but not worth looking at.
This version made me stop and think: the problem isn't that the Kindle is too weak, but that I once again defaulted to "put everything on it."
So the first conclusion is:
Don't copy the desktop calendar onto the Kindle as-is; decide anew for the Kindle what deserves to be seen.
Second version: keep only "today + the next two weeks"
I tried several layouts and settled on a restrained version:
• Top: today's date.
• Middle: today's items in a large-text list.
• Below: the next 14 days as two rows of horizontal date cards.
• Bottom: a small month calendar, but no titles—just how many items each day has.
It's far clearer than the full monthly view. Because what the Kindle really needs to answer is never "what happened every day this month," but three questions:
• What's today?
• What's in the next few days?
• Which days this month are busy?
At that moment I felt a bit dazed: isn't this exactly what I always want to do when writing requirement docs but sometimes forget—first figure out what questions the user needs answered, then decide what to show, rather than dumping everything on the screen.
Data source: not reading Calendar, but the databases behind it
There's a small pitfall here. The content seen in Notion Calendar actually comes from several places: Google Calendar, tasks database, projects database, meetings database.
So I didn't "screenshot" or "replicate" Calendar itself; instead I changed my approach:
Use the API to read the data behind the calendar, then generate a Kindle-specific web page.
Here I still "used my mouth, not my hands": I gave GPT the Notion page links for the databases to read, and it wrote all the scripts for fetching, merging, and generating the page. Currently I've connected three databases—project launch tracking (client name + planned launch date), daily-todo (task name + due date), and meetings (name + scheduled time).
To put it more plainly: the computer acts as the "kitchen" that prepares the meal, while the Kindle is just the "plate" that serves it. This way the Kindle doesn't need to log into Notion, tokens don't leak, and the page is plain enough to feed its old browser.
Some "good enough" trade-offs
Update method: I initially thought about auto-refreshing every 15 minutes, but later decided it wasn't necessary. My schedule doesn't change every minute, and the Kindle isn't a monitoring screen. So I switched to manual triggering—give GPT a command to update the calendar, regenerate the page, and refresh the Kindle. Update when needed; when not, it just sits quietly.
Always-on screen: Some will surely ask, doesn't the Kindle sleep quickly? I checked the settings—the longest sleep timeout is 1 hour, and there's no "never sleep" toggle in the system (unless jailbroken). My approach is simple: set the max of 1 hour, it sleeps at most once an hour, tap to wake, and the page stays on the latest calendar. For a desktop calendar, that's already good enough.
What was really transformed wasn't that Kindle
After all the tinkering, I realized: the interesting part was never "is it technically possible," but that the Kindle forced me to answer a question I face every day in my profession yet often gloss over—
What information deserves to be seen?
This whole issue really hit home for me. Over the years writing requirement docs, I've noticed a persistent flaw: I always feel the need to include everything — every scenario, field, and boundary — as if being exhaustive equals being responsible. I call it "information greed."
Desktop software always encourages you to want more: more views, more colors, more labels, more entry points. But the Kindle is slow, plain, and restrained—it forces you to delete what's unimportant and keep only what truly should be seen. In the end, only three things remained on that small screen: today, what's next, and this month's busy/free days. It felt more like a "life dashboard" than the original calendar.
This is also the lesson I've been learning over the years in product work and doc writing: "more" never equals "better"; "accurate" is what matters. Judgment isn't shown in how much you put in, but in how much you dare to cut.
An old Kindle—no jailbreak, no firmware flashing, just its built-in browser—drove that lesson into me once more.
Do you have an old device lying around that still works but you never knew what to do with? Let's chat in the comments.



