Hey Filip! I'm sorry you found our project disappointing.

Our goal is to make our visualizations systematically available and easily explorable across hundreds of thousands of neurons. It sounds like you're frustrated that we had presented the underlying technique previously?

We've got a lot of positive feedback from the community of researchers studying how neural networks work on the scale of individual neurons. But obviously it's a niche topic and not going to be of interest to everyone!

We wrote about some of the things we discovered in this recent article: https://distill.pub/2020/circuits/early-vision/