Happy Sunday and welcome to InsideAI.  I'm Rob May, a Partner at PJC investing in AI and Robotics.  This week I've written some notes about why the Unity IPO is good for AI startups.

Let's start with the most popular articles of the week:

TikTok will reportedly become an independent, U.S.-based company, with a majority stake held by its current parent company ByteDance. ByteDance will retain control of TikTok’s AI algorithm, in accordance with the Chinese government’s new restrictions on AI and other technology exports. Oracle will be given a minority stake in the new company and will be responsible for processing U.S. users’ data, which will be stored in the U.S.  Last week, TikTok revealed details about the algorithm and data practices that have been key assets at making the company viable. The algorithm helps recommend the next videos based on machine learning techniques that decide what content users are most likely to click on.

As Nvidia prepares its $40B takeover of semiconductor company Arm Holding from SoftBank, the Chinese government is concerned that the company’s technology could be politically weaponized against Chinese firms, and may try to block the deal. 95% of chips designed in China use Arm’s technology.  The deal - considered one of the largest tech acquisitions ever - faces heightened scrutiny due to its magnitude and significance. The $40b purchase will require regulatory approval from the U.S., U.K., EU, and China.

An autonomous ship is about to begin trials before it sails across the Atlantic Ocean using only AI and solar energy. The Mayflower is scheduled to make an appearance on Wednesday before it officially makes the voyage in April 2021, following the same path that the Pilgrims did in 1620. It will use cameras, radars, IBM Watson, and computer vision technology to navigate. Crews will begin trialing the ship off England's south coast over the next several months. The actual 3,000-mile voyage will take 12 days rather than the two months it took the Pilgrims.

This commentary is also published here.

When you want to train a machine learning model to discriminate between pictures of dogs and cats, you can show the model thousands of examples per minute because it’s all digital. If you want to train a robotic arm, using a machine learning model, to set a table, you can’t do it nearly that fast. Every iteration is governed by the laws of physics in the real world, and so having a robot do that tens of thousands of times may take way too long.

One way around this is synthetic data. If you create a virtual world that is similar in its physics to the real world, but where you can do thousands of iterations very quickly, you can augment your real world training set and get better performance. And one way to create a virtual world is by using a gaming engine like Unity. There are a bunch of other benefits to synthetic data that I won’t go into here, except to say that Unity has embraced the use case.

The more important point for AI startups is that this is just beginning. There are a few trends in Unity’s favor that are happening in AI that will drive Unity to grow and provide opportunities for startups to build on top of Unity.

1. As rendering things in gaming engines improves, the use cases will improve along with it. Synthetic data becomes more valuable.
2. As use cases for Unity grow, the businesses you can build by creating things for synthetic worlds are going to increase. One way to think about it is, it was hard to create a social game before Facebook. Once Facebook grew big enough, leveraging Facebook to build a social game became a thing. That will happen with Unity and synthetic data/media of all types.
3. Unity could be the basis of the metaverse. If you think about a virtual world where systems connect, bots perform tasks, and humans can pop into and out of, it can go way beyond Second Life. It could get to the point that it seems and feels almost real. Unity could be the layer that drives a lot of that.

So as a venture capitalist, I'm excited for the Unity IPO.  When a platform gets this large, it means you can probably build a $100M business on top of it.  That means lots of innovation opportunity for the entrepreneurs who are paying attention. 

Thanks for reading.

@robmay