zackashen

I think this is a fair take. We are constantly iterating on the videos and the writing to make sure that the quality is high. We would love to hear more about what disappointed you. While some users have enjoyed and gotten value out of the videos, you definitely aren’t alone in your opinion. The product is early and we feel that with continuous iteration, feedback and model improvement, that more and more people will fall into the prior category. If you are open to providing more detailed feedback it would be really helpful for us. Shoot us an email at [email protected]

Thanks! We might end up doing longer videos, but we just found that there was a happy-medium for our attention span with the videos and making them bite sized kept us the most engaged. However, this decision was made before we made some significant improvements to our video generation workflow and so we definitely have been talking about whether longer videos could make more sense now. Our goal is to just deliver the content in whatever format helps people learn and stay engaged the best, so if that ends up being longer videos or some variety of short-form and long form than that works for us.

That's fair. It is certainly something we will add if people want it.

Not currently, but we are going to add some features in the future to allow users to report courses that have content that needs to be reviewed for false hoods or any other reasons. In the meantime, if you do see this feel free to email us and we can address it.

This is a totally skippable part of the onboarding. Sorry if this is off putting. I had just really enjoyed this part of other apps like Substack where I can see when people I know join and follow them. Totally understand if this is something you want to skip or don't find it appealing.

Yea I think this is true also.

That's totally possible, but I think from personal experience we haven't really enjoyed those tools because they struggle to form a structured curriculum that has videos and assessments. We love using guided learning or a chatbot to teach one of f concepts, but for things that we enjoy passively learning about it's just not really that engaging. I think there are plenty of people who teach themselves things for fun and and want to have structured courses to learn from. I'd be curious if you still felt this way after trying it out.

Courses are publicly available. When you generate a course anyone on the app can enroll in it.

We use pretty comprehensive web searches to find examples and real worked problems to use in courses and base the curriculum and roadmaps off of publicly available syllabi, so from our experience the content has been accurate. We haven't seen anything inaccurate but that could certainly be possible, just like learning via the internet or from any other source.

Yea we generate the course icons with a quick image generation model so total accuracy in representations is not guaranteed. But the actual courses themselves use research to construct the outlines and videos. Feel free to email us ([email protected]) if you end up taking that course and have any other thoughts.

We've seen some of the people on the app using the social features, but yea I think on some apps it doesn't get used. I think if no one ends up using these features we'll just remove them.

Yes we have the ability to share courses via a url. No worries. Appreciate the feedback.

Sorry about that. We wanted to implement more social features on the app to let users learn with their friends which we figured letting users find friends from their contacts/phone number. Understand that isn't for everyone and we might make some changes to make those social features opt-out.

Both workflows follow the same broad pattern. First, break the video into a set of scenes, then plan the scene, write the code, render it, and iteratively refine the result.

For Remotion, the AI determines what resources the scene will need, whether that be images, generated assets, or prebuilt visual components. This part doesn't happen in Manim because its more for math based visualizations.

Re costs, both workflows have a series of presets that make the implementation less expensive and more consistent. The reduction in price for the Terra model series has been particularly helpful for us.