HackerNews中文版

2020年，在疫情肆虐期间，我和几个朋友决定，显然世界需要另一个演示工具。我们的痛点很简单：我们讨厌和PowerPoint/Google Slides较劲。我们只想写内容，然后得到漂亮的幻灯片，而不用费心去调整字体、字号、对齐方式、颜色，以及19种不同的“标题+副标题”布局，而这些布局看起来都差强人意。所以我们构建了一个原型：一个微型的点击式幻灯片生成器。然后我们觉得点击太慢了，就把它变成了一个完整的markdown转幻灯片工具。输入markdown，输出漂亮的幻灯片。简单。然后，就像每个优秀的副业项目一样，我们开始添加各种功能：图片上传、通过API生成PDF、AI集成等等，以及越来越多没人真正要求的功能……在某个阶段，我们正处于“我们绝对是下一个独角兽”的巅峰状态，通过Stripe Atlas在美国注册了一家公司。为了显得更认真，我们给自己发行了10,000,000股股票。因为真正的初创公司都是这么做的，对吧？然后我们又开始愉快地发布功能，但……并不愉快地没有发布任何付费计划。（快进大约一年半）有一天，一位同事给我们发消息：> “嘿，联邦门户网站说我们因为没有提交纳税申报单，欠了6万美元的罚款。”我们都经历了通常只在生产数据库崩溃时才会出现的全身发热浪潮。（给读者的提示：我们最初来自欧盟，在那里，如果一家公司没有收入，就不需要缴纳任何税款；如果你收到罚款，通常最多也就几百欧元）因此，我们学到了一件有趣的事情：你发行的股票数量会影响各种费用和罚款的计算方式。在你点击10,000,000之前知道这一点是很好的。我们花了几天几夜的时间阅读IRS文件，给人们发邮件，试图理解我们是否刚刚让一个甚至没有定价页面的项目破产了。最后，解决方案非常无聊：1）我们提交了一份零收入的纳税申报单。2）在线系统重新计算了一切。3）我们那令人恐惧的6万美元罚款变成了大约1500美元，我们支付了这笔钱。危机算是解除了。然后，第二年年底临近，我们都有同样的想法：“我们再也不想经历这种事了。”从技术上讲，我们可能可以直接忽略这家公司，让它慢慢地走向法律上的死亡。但我们是那些喜欢“把事情做好”的烦人的人，所以我们又花了大约2000美元请人帮助我们以正确的方式关闭了这家公司。我们“我们绝对是创始人了”时刻的总账单：大约3500美元，大量的压力，以及0美元的收入。在那之后，我们基本上让这个项目休眠了。每个人都有自己的生意，而这已经是一个足够痛苦的教训了。除了……我们从未停止使用它。在内部，我们一直用它来生成所有的幻灯片。老实说？输出效果仍然很好。它速度快，看起来不错，而且一旦你习惯了它，再回去使用Google Slides或Keynote，感觉就像在Excel中设计海报一样。最近，AI帮助我们重构了一部分代码，完善了一些东西，并添加了我们一直想要但从未有精力实现的功能。所以现在这个项目又在slidepicker.com上复活了。从这一切中得出的两个教训是：1. 在你至少有一个付费客户之前，不要注册公司（尤其是在另一个国家）。最好等到这些客户产生稳定的收入后再注册。2. 如果你确实注册了公司，在兴奋地在表格中输入“10,000,000股”之前，请咨询一位真正了解税务的人。如果你只想通过markdown获得漂亮的幻灯片，而不想意外地快速通关美国的公司法，你可以在slidepicker.com上试用一下。

查看原文

In 2020, in the middle of COVID, a few friends and I decided that obviously the world needed another presentation tool.Our pain point was simple: we hated fighting with PowerPoint/Google Slides. We just wanted to write content and get nice slides without babysitting fonts, sizes, alignment, colors, and 19 different "Title + Subtitle" layouts that all somehow looked bad.So we built a prototype: a tiny click-together slide builder. Then we decided clicking was too slow and turned it into a full markdown-to-slides thing. Write markdown in, pretty slides out. Easy.Then, like every good side project, we started bolting on features: Image uploads, PDF generation via API, AI integration and more and more stuff nobody had actually asked for....Somewhere in there, in peak "we're definitely the next unicorn" mode, we incorporated a US company through Stripe Atlas. To feel extra serious, we issued ourselves 10,000,000 shares. Because that's what real startups do, right?Then we went back to happily shipping features and... not so happily not. shipping a single paid plan.(fast-forward about a year and a half)One day a colleague messages us:> "Hey, the federal portal says we owe $60,000 in penalties for not filing a tax return."We all experienced the same full-body heat wave usually reserved for production database drops.(Note for readers: We are originally from the EU, where if a company has no income, it doesn't pay any taxes and if you receive a fine, it is usually a few hundred euros max)And so we learned something interesting: the number of shares you issue can influence how various fees and penalties are calculated. Which is nice to know BEFORE you click 10,000,000.We spent days and nights reading IRS docs, emailing people, trying to understand if we'd just bankrupted a project that didn't even have a pricing page.In the end, the solution was hilariously boring:1) We filed a zero-revenue tax return.2) The online system recalculated everything.3) Our terrifying $60k penalty turned into something around $1500, which we paid.Crisis kind of averted.Then the end of the next year approached and we all had the same thought: "We're not doing this again."Technically, we probably could have just ignored the company and let it die a slow legal death. But we're those annoying people who like things "done properly", so we paid roughly another $2,000 to have someone help us shut the company down the right way.Total bill for our "we're definitely founders now" moment: aprox. $3,500 and a lot of stress and $0 in revenue.After that, we basically put the project to sleep. Everyone had their own businesses, and this was a sufficiently painful lesson.Except... we never stopped using it.Internally, we kept generating all our slides with it. And honestly? The output still feels great. It's fast, it looks good, and once you get used to it, going back to Google Slides or Keynote feels like trying to design posters in Excel.Recently, AI helped us refactor a bunch of the code, polish things, and add features we always wanted but never had the energy to implement.So now the project is alive again at slidepicker.com.The two takeaways from all this:1. Don't incorporate (especially in another country), until you have at least one paying customer. It may be better to wait until these customers generate a stable income.2. If you do incorporate, talk to someone who actually understands taxes before you excitedly type "10,000,000 shares" into a form.And if you just want nice slides from markdown without accidentally speedrunning US corporate law, you can play with it at slidepicker.com.

我们因为一个零收入的副业项目，收到了 6 万美元的税务罚款。