We're an early-stage edtech startup. Our goal: use AI to build high-quality, computer-animated versions of every STEM course in the world, in every language, and bring them together on a platform like Coursera, but built for the AI age, with deep LLM integration. We want that literally. Every course. Every language.
This project simply couldn't exist without generative AI. We're two co-founders, and we don't speak five of the languages we just published in. There may be mistakes, but these courses wouldn't exist in those languages otherwise.
We produce lectures fast with AI. But once a lecture is done, adding another language is even faster, it's O(1). So we just made this in six languages. We could keep going.
These videos aren't static MP4s, they're open for improvement. If you have ideas, suggestions, or find mistakes, please open an issue on GitHub: https://github.com/academa-dev/multivariable-calculus
Would love your feedback. Happy to answer questions.
Join the waitlist for upcoming courses at https://academa.ai.
Tech stack: We forked 3Blue1Brown's Manim, synced TTS with animations using OpenAI Whisper, and used Inworld AI for text-to-speech.
I assume you could do a topological sorting of the content to ensure prerequisites are covered, but I'm curious on how automated this could be.
Looks super cool either way!
We actually made the course ourselves, following James Stewart's Calculus and our own approach to teaching these topics (we're two PhD students). We wrote everything in English first, then translated it into other languages.
AI accelerated our workflow, we prompted the visuals and explanations we wanted, then polished the results ourselves. So we're confident in the teaching quality and correctness of these videos, they're not pure AI artifacts :)