conilab @coni — 15.06.2026

我还在牌桌上 Still at the Table

把 FreeGrid 发出去,三小时后 GitHub 上只有我自己点的那颗星。关于预期管理、把每次发布当成实验、没有反馈也是一种反馈,以及为什么我还在牌桌上——顺手也复盘了小说为什么没跑出结果。

Three hours after I shipped FreeGrid, the only star on GitHub was the one I'd clicked myself. On managing expectations, treating every launch as an experiment, why no feedback is still feedback — and why I'm still at the table. With a side post-mortem on the novel that never landed.

这个世界很操蛋,但是你要笑着玩下去,哄也要哄自己玩下去。

很多事情并不遵循"付出就有回报"的逻辑,也不会按你预期的方向发展。

事实上,大多数事情都不会如你所愿——这不是例外,这是常态。

你认真做了,结果平平;你投入很多,反馈寥寥;

这些都是真实世界会出现的,所以,预期管理真的很重要。

一、被打脸这件事

预期管理,是一件比想象中更重要的事。做任何事情,我们内心都会悄悄架起一个预期,无论是主动设定还是被动生成。

一旦预期过高,结果没有如期而至,痛苦就是必然的。

以 FreeGrid 为例。

我把它发到朋友圈、公众号、群聊,心里盘算:再差也能涨十几个 Star,也许能收到几笔打赏,再乐观一点,凑够 688 元的开发者费用也未可知。

现实是,发布三小时后,GitHub 首页上只有我自己点的那颗星。

那一刻,我多少有点难受。

怎么说呢,像是自己给自己放了一场烟花,结果发现观众席空空荡荡,只有我本人站在原地鼓掌。

不过现在写下来,反而释怀的笑了。

二、为什么这次特别难受

这周末,我的状态其实一直不太好。

也许是前段时间能量太高,现在进入了低能量回转期;

也许是几件事叠加在一起,让我开始重新质疑自己:我做这些事情到底有什么意义?我到底有没有在变好?我是不是只是在自我感动?

倒不完全是因为 FreeGrid——它只是压倒这一切的最后一根稻草。

往前看:4 月底启动的 AI 小说创作项目,做了两个月,没什么结果;

实习那边被派了一个陌生领域的活,花了大力气才完成,心气也被磨掉了一截;

看了笑来关于英语学习的文章,被"每天至少三小时训练"吓退,拖拖拉拉不敢开始;

5 月份立下的"本月写一篇三千字文章"的 flag,也没有达成。

这些未完成的事情像账单一样累积,终于在这个周末一并结算,情绪崩了。

理性上,我当然知道,下一步应该分析问题:为什么结果不如预期?是产品定位有问题,宣发方式有问题,目标用户不清晰,还是渠道太窄?当下的困境和难点到底是什么?有没有更精准的改进方法?要不要换一个方向继续试?

这些问题都该问,也都值得问。

但情绪低落的时候,人很难立刻进入理性分析模式。尤其当几件事情同时压过来时,心气会被一点点磨掉。

三、但我真的什么都没拿到吗?

不是的。

这两个月里,我至少在用户运营实习侧证明了自己的策划和执行能力。我学习了如何统筹任务,如何更好地完成别人派发的工作,如何更高效地汇报进度。在这个过程中,我的策划能力、执行能力和反馈能力都得到了训练。

与此同时,我也精进了不少技能,尤其是在和 Claude Code 协作方面,我摸索出了很多新的工作方式。我和 AI 高效率协作完成任务的能力,确实在持续进化。

我创建了个人博客,创建并变现了 Claude Code 教学手册,也从一个 idea 出发,把 FreeGrid 打磨成了一个真正可以展示的 App。

这个过程中,我学到了很多关于编程、GitHub、产品介绍页、README、宣发文案和产品表达的知识。

相比今年年初,我已经走了很远。

所以,不能被失败单方面定义自己。

成功与失败并存时,更该问的是:成功是怎么发生的?什么路径是可复用的?哪些事情失败了?失败暴露了什么问题?下一次怎样调整?

这些答案,才是真正值钱的东西。

四、FreeGrid 失败了吗?我的六点复盘

一、是预期管理失败,不是产品失败。

我一开始确实把预期抬得太高了。低预期、高试错,才更适合一个新产品的冷启动。尤其是一个没有经过充分市场验证、没有成熟流量渠道、没有精准用户触达路径的小产品,刚发布就期待立刻获得大量反馈,本身就容易制造痛苦。

更合理的方式是,把每一次发布都当成一次实验。不是“我发了,就应该有人回应我”,而是“我发了,所以我获得了一组真实反馈”。没有反馈,也是一种反馈。它至少说明:当前渠道不够强,当前表达不够锋利,当前触达的人群还不够精准。

二、我被朋友圈的信息流扰乱了,信息管理出问题了。

最近我在朋友圈和网上看到很多人的成功:有人竞赛获奖,有人商业变现,有人创业做出成绩,有人已经有别人为他铺好路径,只需要证明实力;有人到处旅游,有人 IP 做起来了,有人小说变现了。

这些信息看多了,心里难免会被扰乱。我羡慕,甚至嫉妒——凭什么他们行,我不行?

如果我拿到了结果,那还好,我会觉得“你强,我也不差”。可如果我暂时没什么结果,再看到这些信息,就很容易被刺激到。

所以,最好的办法很简单:少看。

如果有些人确实值得观察,那就固定名单、固定时间、快速扫描。

不要把自己长期暴露在别人的高光片段里。

别人发出来的是结果,我正在经历的是过程。拿自己的过程去对比别人的结果,本来就很容易扭曲判断。

三、FreeGrid 本质上是一次"灵光一闪"的变现尝试。

FreeGrid 最初来自老王的一个想法,我把它产品化、工具化、视觉化了。它当然有价值,但它起点上更接近一个灵光一闪的 idea。

它没有做过市场调研,没有验证用户需求——这种 idea,百里能有一个成就不错了。

今天我在小红书刷到一个独立开发者的一个刚好火了的产品帖子,如果它没有经过系统市场调研,那它的成功里面也有很高的运气成分。

FreeGrid 只是那百个里没有被命中的那一个。

四、样本量太小,不足以判死刑。

如果只有几十个人看到过这个产品,就在这里批判它"没人需要",是不公平的。

它没有获得反馈,未必说明产品没人需要,也可能只是样本量太小。假设有几千人、几万人看到,哪怕只有1%的人感兴趣,也会有几十甚至上百个潜在用户。

问题也许不在产品本身,而是触达规模太小。

五、我不会放弃 FreeGrid。

它是我的第一个孩子,我不允许它的故事就止步于此。

接下来要做的,是打通 Marketing 通道——我需要一个真正能让产品被陌生人看见的 IP 或渠道,而不只是在熟人圈子里自娱自乐。

我需要扩大自己的盘。

我要让 FreeGrid 走出朋友圈,走出小群聊,走向真正可能需要它的人。

六、产品本身不是问题,讲好故事才是核心。

FreeGrid 的 GitHub 页面、产品介绍页、README,相比我以前做的东西,已经进步很多了。它并不是一个随手糊出来的产品。

它有设计,有理念,有表达,也有我认真打磨过的痕迹。

所以现在最关键的问题,可能并不在于产品本身有多差,而在于:我还没有把它的故事讲好。

  • 用户为什么需要它?

  • 它解决了什么痛点?

  • 它和普通记账工具有什么区别?

  • 它背后的财富观、自由观、长期主义,如何用更容易被普通人理解的方式表达出来?

  • 怎样的标题、截图、案例、短视频、文章,能够让别人愿意点进来,愿意试一试?

这些才是我接下来要思考和训练的东西。

五、顺手:关于小说失败的复盘

我那两个月的 AI 小说创作,为什么没有跑出结果?我真的沉下心调研过市场吗?我知道当下读者喜欢看什么吗?我知道平台推荐什么、用户点击什么、评论区在骂什么、榜单作品怎么设计爽点吗?

没有。我只是把自己的 idea 往上套,然后期待它被人喜欢——这是在吃 90% 的运气。

正确的做法应该是:先看读者在看什么,再借鉴、模仿,加入一个微创新点,运气占比就能降到 50% 以下。

写小说当然吃运气。这个判断我不会美化它。

所以它更适合作为副业来做。成了,那是意外收入;没成,也可以沉淀表达能力、故事能力、市场感和长期创作经验。

这样看,它依然值得做,只是不能用“必须马上成功”的心态去做。

尾声:

写到这里,我突然又一次感受到了写作的魅力。

原本很差的心情,在文字里被拆开、摊平、分析、重组。

那些混乱的情绪,慢慢变成了材料;

那些难受的瞬间,慢慢变成了方法论。

写着写着,我好像又知道下一步该怎么走了。

真是如笑来老师所说:文字是最好的的生产资料。

文字永远是生产资料,我不打算拱手让给别人。

我不想只看别人的输出,我也要持续生产自己的输出。

我要继续写,继续做,继续试错,继续把自己的失败和成功都加工成经验。

这个世界很操蛋,但我还是要笑着玩下去。

哄自己也好,鼓励自己也好,咬牙也好,自嘲也好。

反正我还在牌桌上。

谢谢笑来老师,也谢谢写作。

我大概知道接下来该怎么做了。

The world is a shitshow — but you keep playing, grinning through it. And if you have to coax yourself into staying in the game, then you coax.

A lot of things don't run on the logic of effort in, reward out. They don't move in the direction you expect them to, either.

Most things, in fact, won't go your way — and that isn't the exception. It's the baseline.

You do the work, carefully, and the result is flat. You pour yourself in, and the feedback is a trickle.

This is just what the real world does. Which is exactly why managing your own expectations matters more than it sounds like it should.

i. On getting humbled

Expectation management is more load-bearing than it looks. Whatever we set out to do, some quiet expectation goes up inside us — whether we plant it deliberately or it just assembles itself while we're not looking.

Set it too high, watch the result fall short, and the pain isn't likely. It's guaranteed.

Take FreeGrid.

I pushed it everywhere — Moments, my public account, the group chats — and ran the math in my head: worst case, a dozen-odd stars. Maybe a few tips. And if I let myself dream a little, maybe even enough to cover the ¥688 developer fee.

Reality: three hours after launch, the only star on the GitHub page was the one I'd clicked myself.

That stung, a little.

How to put it — like setting off fireworks for yourself, then turning around to find the stands empty, just you standing there, applauding into the dark.

Though writing it down now, I find I've already let it go. I'm even smiling.

ii. Why it stung extra this time

The truth is I'd been off all weekend.

Maybe I'd been running hot for a stretch and was now swinging through the low-energy half of the cycle.

Maybe it was a few things stacking up at once, until I started second-guessing the whole enterprise: what is the point of any of this? Am I actually getting better — or just moving my own heart and calling it progress?

It wasn't really about FreeGrid. FreeGrid was just the last straw that broke the pile.

Look back a little. The AI-novel project I kicked off at the end of April: two months in, nothing to show. The internship handed me work in a field I knew nothing about, and grinding it out cost me a good chunk of my drive. I read Li Xiaolai on learning English, hit at least three hours of training a day, and flinched — kept stalling, never daring to start. The flag I planted in May — write one 3,000-word piece this month — went unmet.

All these unfinished things piled up like unpaid bills, and this weekend they all came due at once. The mood collapsed.

Rationally, of course, I know the next move is to diagnose. Why did the result miss? Is it the positioning, the way I marketed it, a fuzzy sense of who it's even for, a channel too narrow? What's the actual bottleneck right now? Is there a sharper way to fix it? Do I pivot and keep testing?

All worth asking. Every bit of it.

But when you're down in it, you can't just flip into analysis mode. Especially when several things land on you at once — your drive gets sanded down, a little at a time.

iii. But did I really walk away with nothing?

No.

Over these two months, I at least proved — on the internship side — that I can plan and execute. I learned how to coordinate tasks, how to deliver on work someone else hands me, how to report progress without wasting anyone's time. Planning, execution, communication: all of it got its reps.

Alongside that, I sharpened a fair number of skills — especially working with Claude Code, where I turned up a whole set of new ways to work. My ability to get things done with an AI, at speed, is genuinely still evolving.

I built a personal blog. I built a Claude Code playbook and got it to pay. I took FreeGrid from a single idea to an app I can actually put in front of people. Along the way I picked up a lot — about coding, about GitHub, about landing pages, READMEs, launch copy, and how to talk about a product at all.

Compared to where I stood at the start of this year, I've come a long way.

So I won't let failure define me single-handedly.

When success and failure sit side by side, the better question is: how did the success happen? Which path can I run again? What failed? What did the failure expose? What do I adjust next time?

Those answers are the part that's actually worth money.

iv. Did FreeGrid fail? A six-point post-mortem

i. It was an expectations failure, not a product failure.

I did set the bar too high at the start. Low expectations, high experimentation is the better posture for a cold start. Especially for a small product with no real market validation, no established traffic channel, no precise path to the people who'd want it — expecting a flood of response the moment you ship is a machine for manufacturing pain.

The saner frame: treat every launch as an experiment. Not I shipped it, so someone owes me a reply, but I shipped it, so now I'm holding a set of real signals. No feedback is also feedback. At minimum it tells you something: the channel isn't strong enough yet, the pitch isn't sharp enough yet, the people you reached weren't quite the right people yet.

ii. I let the feed get to me — a failure of information diet.

Lately my Moments and the wider internet have been a parade of other people's wins: someone took a prize, someone monetized, someone's startup is showing results, someone already had the path paved for them and just has to prove they belong, someone's traveling everywhere, someone's brand took off, someone's novel is paying out.

Soak in enough of that and your head gets noisy. I envied them — was jealous, even. Why does it work for them and not for me?

When I've got results of my own, it's fine: you're strong, and I'm not far behind. But when I've got nothing to show yet, the same feed lands like a needle.

So the fix is almost embarrassingly simple: look less.

If a few people are genuinely worth watching, fix the list, fix the time, scan fast.

Don't leave yourself parked, long-term, in front of other people's highlight reels.

What they post is the result. What I'm living is the process. Holding my process up against their results is a distortion engine by design.

iii. FreeGrid was, at bottom, a flash-of-inspiration bet on monetizing.

FreeGrid started as a friend's idea; I turned it into a product — gave it shape, tooling, a visual identity. It has real value, sure. But at its origin it was closer to a spark than a strategy.

No market research, no validated demand. With an idea like that, one in a hundred working out is a decent hit rate.

Today I scrolled past a solo developer's post on Xiaohongshu about a product that happened to blow up — and if it never went through systematic market research, then a big slice of that success is luck too. FreeGrid is simply one of the ninety-nine that didn't land.

iv. The sample is too small to hand down a death sentence.

If only a few dozen people have ever laid eyes on the thing, it's not fair to stand here and convict it of nobody needs this.

No feedback doesn't have to mean no demand. It might just mean the sample is too small. Put it in front of a few thousand — a few tens of thousands — and even at 1% interest you'd have dozens, maybe hundreds, of potential users.

The problem may not be the product. It may be the size of the reach.

v. I'm not giving up on FreeGrid.

It's my first child. I won't let its story stop here.

What's next is opening the marketing channel — I need a brand or a pipeline that actually puts the product in front of strangers, not just a circle of people I already know entertaining itself.

I need to widen the board.

I want FreeGrid out of Moments, out of the little group chats, and in front of the people who might genuinely need it.

vi. The product isn't the problem. Telling the story is the whole game.

FreeGrid's GitHub page, its landing page, its README — they're a real step up from anything I've made before. This is not something I slapped together. It has design, a point of view, a voice, and the fingerprints of someone who actually sweated the details.

So the key question now probably isn't how bad is the product — it's that I haven't told its story well yet.

  • Why does anyone need it?
  • What pain does it actually kill?
  • How is it different from any other expense tracker?
  • The ideas underneath it — about money, about freedom, about playing the long game — how do I put them in a way an ordinary person actually gets?
  • What headline, what screenshot, what example, what short video, what essay would make a stranger want to click in and give it a try?

That is what I need to think about, and train, next.

v. A side note: the post-mortem on the novel

Two months of AI-assisted novel writing — why did it produce nothing? Did I ever actually sit down and study the market? Did I know what readers want right now? What the platform pushes, what people click, what the comment sections are tearing apart, how the chart-topping titles engineer their payoffs?

No. I just draped my own idea over the top and waited to be liked — which is betting on 90% luck.

The right move would have been: look first at what readers are reading, then borrow, imitate, fold in one small twist of my own — and watch the luck component drop below 50%.

Writing fiction is a luck game. I won't dress that up.

Which is exactly why it belongs as a side project. If it hits, that's surprise income. If it doesn't, I still bank the reps — on expression, on story, on a feel for the market, on the long muscle of making things over years.

Seen that way, it's still worth doing. Just not with a this has to work immediately mindset.

Coda

Writing this far, I feel it again — the pull of writing itself.

A mood that started out rotten got taken apart on the page: laid flat, examined, reassembled.

The tangled feelings slowly turned into material. The bad moments slowly turned into method.

Somewhere in the writing, I seem to have found the next step again.

It's exactly as Li Xiaolai puts it: words are the best means of production.

Words will always be a means of production — and I have no intention of handing mine over to anyone else.

I don't want to only consume other people's output. I want to keep producing my own.

I'll keep writing, keep building, keep getting it wrong, keep machining my failures and my wins alike into experience.

The world is a shitshow — but I'm still going to play it with a grin.

Coax myself, cheer myself on, grit my teeth, laugh at myself — whatever it takes.

Either way, I'm still at the table.

Thank you, Li Xiaolai. And thank you, writing.

I think I know what to do next.