The results are really good, now i want a system to create a scenes and somehow combine them together. Any thoughts ?

I'm truly mind-blown by how good these shorts are performing. This guy started it maybe a month ago, and already 200K subscribers. The idea is similar: have a consistent character over multiple videos using Gemini. And for a very specific topic, maybe something about the outdoors. If you're guys familiar with the Outdoor Boys YouTube channel. That's something I would love to replicate in shorts. Maybe give some tips on outdoor skills: how to start a fire and other basics. And I think these videos will perform very well. That's the legend: https://www.youtube.com/@OutdoorBoys Maybe do similar content to him.

I want to build a system that goes through a YouTube channel, extracts all the comments, and then generates some actionable insights. For example: 1. What videos to make next based on the user's comments and how many times they asked for it 2. Analyse the videos and tell me what is the ICP that we're targeting and what other videos might be interesting for them. 3. Give a scoring based on the comments on each video, whether it was a successful video or not, and what insights can be extracted from the comment section. I know other big YouTubers have created similar automations to do automatic analysis of different competitors, but for this system, it won't be like an analysis of competitors. It would be more just for individual insights. The ideal customers for this are somebody who is having much more success than me. They're having a lot of different comments, which I've had in some of the videos, and it was really difficult to remember what the people were asking for and what other videos they asked for. Look at the kind of emotions that people like it and stuff like that. I would like the AI to just have a look at these videos and then generate some ideas and some insights. YouTube did add this feature, but it's so bad that it just doesn't make any sense. I think looking at the comments and seeing what the users are asking for is a better way to generate these future ideas.

I want a system that works as an RSS feed or some sort of tracking different websites like the important ones (Google, AWS Courser, Anthropic, OpenAI, n8n) to stay ahead of the curve and always be notified when something new has happened. The AI or the system should be able to get this data and then with an LLM summarise it into maybe a maximum of three or four bullet points of the findings. Then gather these in a nice report per finding and the link, and send a personal email with this. The end product could be a simple landing page to subscribe to this newsletter and we provide these insights automatically to you. It could be maybe on a weekly basis. Example something i found out today and end up summarizing: Cursor 2.0 is out: https://cursor.com/blog/2-0 TLDR: - Composer Coding Model: A new frontier agentic coding model that is 4x faster and completes most tasks in under 30 seconds. - New Multi-Agent Interface: An interface built around agents and outcomes, allowing users to run many agents in parallel seamlessly. - Embedded Browser: The browser is now built into the editor, including full developer tools and agent controls. - Improved Code Review: Agents' changes across the entire codebase can now be viewed in a single, consolidated place.