Thoughts on the Easy and Hard Parts in Building an Spatial Computing Unicorn

Ina Yosun Chang
3 min readMay 28, 2020

Some notes on my current AR software investment thesis — and this may explain the method to my madness for what I build.

Or fielding the common questions I’m asked — which I think are actually questions investors should not ask when presented with AR software!

Tracking is commodity.

  1. Anyone can create a new type of tracking. Making it steady or 99.9% good is what takes a big company’s brute force to refine.
  2. Trust Apple and Google to provide tracking the same way that you would trust the water company if you start a restaurant or boba tea shop.
  3. Tracking companies can at best be an M+A. (I absolutely love trying out new SDKs. But, it’s just not big enough for me to be interested in.)

Asking if it’s made in Unity

  1. Say you toured a $88M house in the LA hills. It has a 12-car garage. Manicured box trees and meticulous gardening. Picture windows with surreal view of the city above the smog. Amazing lighting. Indoor-outdoor floor to ceiling windows. Multiple bespoke fireplaces. Secret passageways. Bedroom closets that looks like a Louis Vuitton shop. Your bathroom doubles as a day spa. Infinity pools that look like sculpture art. A private gym with reflection pool. Multistory library with winding staircases. An art gallery. A wine cellar that Thomas Keller or even Dominique Crenn might be jealous of. Your own private underground movie theatre with massage chairs. It has all these components and they all fit together so well. It’s created by a master craftsman. It feels like a billionaire’s compound. Do you ask, “Did you use wood to build your house?”
  2. Okay, tech example, you could ask of an iOS developer, “Did you use XCode” or of a modern programmer, “Did you use Visual Studio.” Yes!
  3. In short: It doesn’t have to be, but I abuse Unity as a great cross-platform compile tool. (Did you want the real estate developer to answer, “We have our own proprietary-branded wood-replacement technology.”)

Software incorporating tracking + intuitive user experience is hard. It’s a new form of computing. There aren’t design patterns and standardized UI frameworks — there isn’t even really an established UX modicum. This is about being first in creating something that can eat the world in a new era of computing.

  1. We’ve seen a lot of copycat AR stuff where you (select from a bunch of objects and) place it there. (Some with convoluted UI with way too many menu’s.)
  2. Anyone can create Hello World AR — but I hope you understand there is so much more to develop than “just augment something” with this new era of spatial computing.

Master Craftsmanship is the Hard Problem: Utilizing AI + computer vision with a great “AR experience

Create truly intuitive creator software that has no learning curve, so that anyone can use it immediately with no friction. Here are some current startups I’m working on:

  1. Baby steps in EdTech innovation: An AI writing tutor that turns what you write on real paper into a 3D AR scene
  2. Baby steps in media innovation: A democratic creator software that lets traditional non-digital artists create AR content
  3. And baby steps in breaking free of zoom in online experimental theatre.



Ina Yosun Chang

{wonder, innovation, elegance} ∈ a hacker's take on augmented reality, 3d graphics, mobile iOS and Android dev, startups, COOL STUFF and life