This is a weekly newsletter on the Software as a Service world. Learning, building, and shipping. Written by Ethan Mick.

"After you solve the biggest problem, something else becomes the biggest problem." – Balaji Srinivasan

The Best Bits

• Finish your projects. If you're going to read anything, read this.
• A headless, fast, and full-featured autocomplete library.
• The problem with disabled buttons.
• Some nice Framer Motion recipes.

The Allure of Building on a Platform

In the past few months, we've seen two platforms blow up that destroy entire ecosystems: Twitter, and Reddit.

So... if building on top of a platform is so dangerous, why do we keep on doing it?

Because, frankly, it's a great place to start building. You have a captive audience that you can target. Platforms often have deficiencies and gaps in their product that make for great wedges for you to get started. Don't believe me?

• Twitter added the ability to search DMs in 2019, 13 years after it was founded.
• Reddit moderator tooling is so garbage mods use third-party apps.
• Twitter's timeline changed to be algorithmic instead of time-based.
• Analytics are basically non-existent on both platforms.

The list goes on and on. The platform builds the most common and useful features for all its users, leaving tons of niche features out. These features can be built by third-party developers and get instant traction.

Until... you know. The platform explodes.

There is an inherent risk in building on top of a platform. I'm not suggesting you don't do it – just acknowledge the risk. The first risk is the platform kicking you off. We're seeing that today.

The other risk is the platform builds your features directly into it. That also kills your product pretty quickly. In the Apple world, that's called "Sherlocked"

In the mid-2000s, a company named Karelia Software released a search utility for Mac called "Watson". It was named after the assistant of Sherlock Holmes, the famous fictional detective created by Arthur Conan Doyle. Watson was praised for its ability to integrate internet-based information into a user-friendly desktop application, and it quickly became popular among Mac users.

Around a year later, Apple introduced a similar feature in its own operating system, which it named "Sherlock". Sherlock was integrated into the operating system and provided for free, while Watson was a standalone paid application. As a result, the demand for Watson declined and Karelia Software eventually discontinued it.

When this happens, sometimes you can pivot and survive. Sometimes you can't.

So if you are building on a platform, what can you do? I think it comes down to two strategies:

1. Diversify as quickly as you can. De-platform users, diversify your product, and reduce the risk of being on the platform.
2. Milk it for all it's worth. Understand that the going is good right now, and it won't last forever. Make that money.

Either way, do what you can because it won't last forever.

Tech Tip

I'm building a CLI that will build Next.js projects. I'm using it locally for some content, and it needs to run a bunch of commands. In TypeScript, here is a super easy way to run a command with code and get the output.

const util = require('node:util');
const exec = util.promisify(require('node:child_process').exec);

async function lsExample() {
  const { stdout, stderr } = await exec('ls');
  console.log('stdout:', stdout);
  console.error('stderr:', stderr);
}

Docs here.

Go forth and code!

Cloud Chronicles

• YouTube Subscribers: 1,378 (+54 in the last 7 days)
• Newsletter Members: 276 (+22 in the last 7 days)

Weekly Update: v2023.6.12

Last week I moved a few projects forward but felt like I lost some momentum toward the tail end of the week. I accomplished most tasks I wanted to early on and then coasted. To try and fix that, we’re upping the goals and trying to be a little more

Build SaaS with EthanEthan Mick

This week I've been doing some project planning for my freelance work. It was sparked by a call from a client I did some work with earlier this year asking about my availability. And I was like, "Uhhhh, I don't actually know. Let me get back to you." I then did the rounds and asked about several proposals I had out.

Turns out, I'm going to be quite busy! In a good way.

Two projects are going to happen in July, which will keep me quite busy and also meet my revenue goal for the year. Then I have a tentative agreement lined up for August, pushing my availability into September.

I also still have another proposal out for work that might start very soon, but we'll see if that happens. I'm now in a position where I don't need any work, so I'm not hunting things down super hard.

Last Byte

• Vercel released an AI SDK.
• Sanity CMS.