[P] Open-source Neural Search framework to implement semantic search & multimedia search. Just released 2.0, seeking your feedback.

I heard your feedback on 1.0 release post on my project Jina, many people were keen to use Jina for multimedia search because that's where use of Neural Networks makes significant difference. So I focused on that part and I was able to transform it from 1.0 to 2.0 within 3 months.

Last post on 1.0 release to give you some idea what this project is about

Actually, I should say - "'we' made this", because there were more than 155 contributors who did it, not just me. The primary changes we made

• We saw MachineLearning beginners struggle in using Jina 1.0, so we separated the codebase where Machine Learning expertise is required(jina-hub) and the one which MachineLearning beginners can use(the jina core). Now ML beginners don't need to worry about jina-hub and can use jina hub packages directly to implement ML specific tasks without the need to understand advanced ML concepts. While advanced ML users can create their own jina-hub packages.
• We cut down a lots of abstractions to make it easy to use for beginners
• Made python APIs more intuitive to use
• Improved performance(3.6x faster on startup)

Here's Jina 2.0 and here's Jina 1.0. I seek feedback from people who are looking at this project for the first time, as well as people who have tried their hands before but had some challenges in using it. Few questions, I'm seeking answers to

1. Do you feel that we have reduced complexity by a lot of margin?
2. How easy it is to use for a beginner now?
3. What questions are still unanswered?