How was that AI layering at the beginning done or conceptualization or writing of the code, debugging? Give us an example.

So recently at In The Pockets, I think we will also talk about later, but we did something with Gaussian Splats.

And there, for example, I started building the API step by step.

I told the AI how I thought it should be.

Then it made a plan for me.

I looked at the plan.

I made some adjustments.

But do you see, let's say, the Lens ecosystem leaning towards a utility, leaning towards the gaming? What's your take?