Rachel How

Apart from my main desk, I added a plain IKEA desk on wheels so I can roll it between my living room and my home office, depending on where I'm filming. I've simplified everything down to lightweight, portable gear, so setting up a scene is much quicker and easier.

A simple, blank notebook and a pen. Nothing beats it for thinking loosely and connecting ideas without being distracted by software.

I'm not going to pretend I've solved this. From design, vibe coding apps, making videos, to researching ideas, I enjoy all of them.

So when I hop between different kinds of work, I go by energy levels and whatever is keeping me excited that week. I don't try to move everything a little each day. I'll give a week mostly to scripting, then the next mostly to filming and editing. The weeks I try to do all of it at once are the weeks I get nothing done and feel terrible about it.

Mobbin’s YouTube channel grew from 200 to 30,000 in 6 months. That looks like a lucky break, but I’ve been making YouTube videos since 2019 and I’ve made plenty of mistakes along the way.

I used to make the video first and leave the title and thumbnail until the very last step, right before uploading. Often it flopped because the packaging wasn’t attractive enough.

Now I start by asking "would a stranger click this over anything else on their YouTube homepage?" Title and thumbnail come first. Ideally the video idea is something unique to the channel, something people can’t find elsewhere.

Things 3, Figma, Aside Browser, DaVinci Resolve, Notion, Claude, Cursor, Notion Calendar, Apple Notes.

Every YouTube video starts with mocking up the title and thumbnail in Figma. I place my thumbnail candidates alongside other videos' thumbnails on the canvas, so I can gauge whether mine stands out and whether the title and thumbnail work together at a glance.

From there, it's research, script, shoot, edit. I spend most of my time on research and editing. Then, the process ends where it began: another round of designing title and thumbnail variations. I'll explore several side by side on one page, upload the strongest ones to YouTube, and run A/B tests to see which performs better.

Definitely vibe coding with Claude or Cursor. Building apps is so fast now that I can stop worrying about technical limitations and just build the ideas in my head! I love this era we live in.

Whenever I'm building an app with AI tools, I still design it in Figma first. There's a lot of talk about skipping straight to prompting, but I genuinely believe the canvas is not dead.

I always encourage people to design on a canvas before they build anything, whether that's Figma or literally a piece of paper. The thinking happens on the canvas.

I storyboard parts of my videos in Figma, frame by frame. Sometimes as slides, sometimes as full illustrated sequences. A little hack I use: Instead of animating each frame in an editing tool, I wire them up in Figma's prototype mode, hit record, and plug that screen recording straight into DaVinci Resolve as finished footage. It saves a lot of time and lets me stay in the tools I'm most familiar with.