Alan Pun

Our latest launch video for @VirajAero ! Congrats team

BTS of our viral, sold out launch video for @Luofei_Chen and @okihomeai 🙌🙌

Snippets from our launch video for @nettle, who just raised $4.8M to build the AI workspace for loss control. We told a story of how innovation has always been a part of the history of insurance, and how Nettle is carrying on that legacy into the future. It was a pleasure working with you Jack Miller and team, thank you for trusting us!

Snippets of our launch video for @okihomeai, your Personal AI Computer. Results: 1.1M views on X. Sold out First batch. Became a much larger convo on what's coming next in the world of personal AI. Thanks @Luofei_Chen for trusting us again!

Our latest launch video for Oki Home, congrats to the team!

Every launch video has two audiences that want opposite things. Investors want to see the future, to feel "this is going to be huge." Buyers want to know what problem you solve, why it's better than other solutions and if they can trust you. The video can serve both. Open with big picture then switch to product explainer. Investors and buyers both see what they want.

For a great launch video, hook is the most important (obviously) so a lot of launch videos open with a lot of controversy. "You can control people using AI." But controversy can backlash. You want to make a big claim without having TechCrunch articles like 'This YC company claims AI can control people.' Go till you cleverly make this claim without people being angry about it.

Only 90 seconds to sell Artificial Societies to a cold audience. Here's how we structured it: 1. The grand narrative hook: The product uses AI simulations to predict how many likes your LinkedIn post will get. But we dug deeper and thought to ourselves: What is the big picture here? People always wanted to predict the future. Tarot cards, magicians, or palm readers. All of it was invented so we can feel some control over what happens next. Artificial Societies does the same thing, with a thousand AI simulations instead. 2. The product Now the explainer. Almost every product comes down to step one, step two, step three. 1. Upload a draft of your message 2. Hit simulate 3. 30 seconds later you know how it performs But which features to show matters more than the steps. Ask yourself what was impossible before or very hard to do so? Predicting the future, that was Artificial Societies edge. 3. Proof Big claims invite questions. This section exists to address criticism. I always ask our clients what are people skeptical about with the product and always add a section to answer it. The video got 300 new business inquiries.

I had a convo with a Silicon Valley founder about the 996 culture. They adopted it from the Chinese, but if only they knew how inefficient most Chinese companies are.

Most of you know me as the 'launch video guy' for tech startups. But it wasn't always what I did... I spent my 20s trying lots of stuff: Architecture, coding, an e-commerce business, websites. I was throwing spaghetti at the wall and seeing what stuck, but I always knew I will do something creative. So in 2020, at 30, in the middle of COVID, I quit my architectural job and went into filmmaking. I started small, learning how camera, lighting, editing, etc works. Then I moved into motion design. And because I'm a secretly a tech nerd (I was reading Kevin Kelly books even back in university) who happens to get on with tech founders, I niched into tech, which was the real turning point for me. And now I run a end to end video production company doing viral launches for screenpipe | YC S26, Voltair, Brumby (Formerly GrazeMate), Artificial Societies and 50+ more. Me being the 'launch video' guy is genuinely a 14+ year story.

Two ways to make a launch video: A founder on camera. Nobody on camera. Both work. It just depends if anyone knows the founder yet. No recognition means you have to lend the product your face, because there's nothing else to borrow trust from. Once you have recognition, you can step out of the frame and let the product carry it.

If a client is surprised by the first animation draft, you messed up. Get the green light on the script first. Then the storyboard (with references next to it). By the time you're showing the draft, they know what to expect. Surprise is fun in film, but in client work is extra revision round that you could have avoided.

We've made 3 launch videos for Mesmer (YC P25) since May 2025. Another one is on the way. Each one supported a change in their product. Startups move fast. Insights and positioning change. You don't need to launch every year. You can do it every time your product pivots.

There is a trick I use from Hollywood in every launch video we make: Play a little montage before diving into the video. People have terrible attention spans, this keeps them hooked. Example from our launch video for @voltairlabs.

What Office of Tomorrow does: End to end video production. What Office of Tomorrow doesn't do: Distribution. In a traditional ad agency, distribution was always a separate department. Today there are separate companies who never touch the creative and spend their whole day setting up influencer deals or buying Meta ads. I looked at that world and had zero interest in learning how it works. I'd rather focus on what I am most interested in. The downside is there, though. We don't have full control over how the work performs. But most of our clients are YC-backed, so they use YC as a distribution layer. And the ones running videos on their own company accounts or personal brands sort it out themselves. I want the studio to be known for one thing and one thing only: We're the studio that make complicated products make sense to people.

One idea from architecture has shaped how I think about launch videos. "Form follows Feeling" Before you design anything, decide how you want people to feel. Then work backwards. The visuals, pacing, music and copy should all reinforce that feeling.

My first ever 3D motion design project back in 2022 was for Meta. The only issue? I’d only been learning 3D animation for a month. For the next eight weeks, I worked 12-hour days, weekends included, searching Youtube for tutorials to learn while I worked. The project turned out great and is how I started working with all the tech companies like I do now. Another time, a crypto company in Hong Kong hired me to shoot a commercial that required quite elaborate lighting. I had never set up a light in my life. I also said yes, and spent most of the budget hiring an experienced cinematographer, and that project also went well. Both times, I was completely out of my depth. Both times, I said yes and figured it out, even if it meant working overtime. Especially early in your creative career, that’s how you grow. As they say: Jump out of the plane. Build the parachute on the way down.