Hey everyone! I really like how websites like Brilliant try to teach in an interactive way. But it's like $160 (or more depending on the region) a year so I've been making a free one.

It's called OpenAha. right now there's 140+ courses and 1000+ lessons in math, physics and chemistry. no account or ads.

A few to try if you only have a minute:

untangle a graph: https://openaha.org/courses/graphs/planar/?step=1
draw the house in one stroke: https://openaha.org/courses/graphs/konigsberg/?step=3
program a little robot: https://openaha.org/courses/coding/sequence/?step=6
Two pendulums passing energy back and forth: https://openaha.org/courses/waves/coupled/?step=3

This project is mostly implemented with claude: I planned every course, then wrote the lessons with AI. so far only about one third of the pages have been carefully reviewed by a person. That's why I'd really like people to poke at it. If a lesson is confusing or sth's wrong, make a PR or hit the flag button in the lessons. Also feel free to contribute on github to make this even better.

code: https://github.com/open-aha/open-aha.github.io

Edit, to clear something up: OpenAha isn't affiliated with Brilliant and doesn't use any of their content. I don't even have a Brilliant account. Brilliant just uses the style of learning I liked (try things first, explanation after). every lesson was written for this project, and we're improving it as an open source community.

And yes it's not finished, there will be mistakes. That's what the flag button on every step is for, every report gets fixed in public on github

you are viewing a single comment's thread
view the rest of the comments

[–] Telorand@reddthat.com 6 points 1 day ago

$13/mo can be a lot for people, that's true. There's certainly a subset where they simply don't have that extra. I suspect that subset is fewer than people might imagine, but I honestly would rather people have accessible education than not.

For the second part, I think it's a hard ask for experts to contribute in that way, but hey, people spend their hobby time working on Linux and other OSS, so who knows?

I also have ethical issues with the use of Claude itself (as well as other tools/platforms provided by billionaires), but I think that's beyond the scope of your original questions.

There are other ML models made from scratch by universities and experts which aren't beholden to shareholders (and might not be accessible by the public, yet), so if you are determined to continue using AI in this way, keep an eye out for potentially better alternatives.

permalink

fedilink

source

parent