Regenerate any long educational video as a short, summarized one

My previous project is finally done, I’m actively using it and enjoying it — which usually means it’s time for a new idea.

As usual, I started by analyzing my own workflow and realized that I spend an absurd amount of time learning through video courses. YouTube has summarizers, but reading a summary is not the same as actually watching and understanding a video.

That got me thinking:
What if there were a product that takes a long video, analyzes it, and generates a new, shorter educational video based on the core ideas?

Not just cuts.
Not just a text summary.
But an actual regenerated learning video.

I did some rough calculations using AWS (transcription, LLMs, video generation), and it looks pretty expensive for an end user. So now I’m wondering:

Are there any cheaper or smarter ways to build something like this?
Maybe alternative stacks, hybrid approaches, or tools I’m overlooking?

I’d honestly be the first and most loyal user of this product.

P.S. It would be amazing if you could control the “scientific depth” of the language:
– for kids
– for the general public
– fully scientific

I really want to watch molecular physics explained for preschoolers. 😅