Everything Takes 10x Longer Than You Think

Something I constantly remind myself of is that everything takes 10x longer than you think it will.

Any time I think of a “quick” project I can build, I realize shortly after that it’s going to take far more effort than I was prepared for.

There’s always details you don’t can’t about until you start.

The more important piece, at least in the context of building products, is that nothing great can be built in a weekend, let alone a few months.

It’s the constant reminder that to move fast, you must first move slow.

Great things take time to build, so your choices are limited.

This isn’t a bad thing, it’s just crucial to understand.

It is not supposed to be easy, nor should it happen quickly.

It’s about setting the correct expectations.

I’m launching my first side project in Q2, and one of the criteria I’ve come up with is the idea must be something I’m willing to work on for at least a year.

This is because you won’t get any indicators that what you’re doing is working for a long time.

Knowing this, it’s become increasingly important for me that the projects I decide to work on are ones that truly excite me.

The idea is to work on something that you deeply care about and would be doing regardless of whether it “succeeds”.

In the end, those that take their time to build something great, no matter how long it takes, are the ones that will win (or so I believe).

What I’ve been building

I’ve had two side projects I’ve been working on this quarter, one of which being a web app boilerplate. It’s meant to use for all of my projects and not spend time rebuilding the same core pieces (like authentication and billing).

This is the last screen of the onboarding flow (after you’ve chosen a subscription plan, created a team etc.). You’d be surprised how much effort something as simple as this takes, but the amount of edge cases, especially when dealing with payments is enormous.

This is another page I’ve spent a ton of time on but it’s now close to being finished. I’m building the dashboard to be team/workspace-based, where you can invite members and share certain data with them. Inviting a team member and having them onboard was a pain to implement, but I’ve just about sorted it out.

Things I’ve found interesting

links to valuable stuff I thought was worth sharing

Zen Browser - I’ve shared Zen before but with how much they improve it withe very update I felt I had to again. I’m thinking of creating a video on YouTube talking about my setup with the browser and why I like it so much.

Tailwind v4 - I use tailwind for all of my projects to write CSS and I love it. v4 just released recently and I’m excited to spend some time learning it. The new website looks dope.

WorkOS - For implementing things like authentication, I’ve been using WorkOS. They provide a nice API to work with that removes some of the headaches dealing with social logins. They’re also focused a ton on adding enterprise features so I figured it’d be good to have those ready to drop in. Their support has also been very helpful.

Not Every AI Problem is a Data Problem - I’ve not read through the entirety of this paper but it’s an interesting read about scaling LLMs and other AI systems. The TLDR is that simply adding more (and specifically synthetic) data does not scale well. Models heavily rely on quality data, so simply creating tons of it and hoping it’s good isn’t enough.

Have a great week.

Cole

P.S. If you found this letter helpful, please consider sharing it with a friend :)