Hbarkh

I'm nearly done studying for the GRE, this is a huge gap that needs to be filled. None of the test prep companies have the software know-how to build anything fun. I think your product can succeed. My 2 cents (take with a grain of salt) is that real learning is difficult, so your ideal learning algorithm will constantly be giving people problems they're bad at, and attacking their weaknesses. This could discourage people who are used to 'studying' as just doing busywork/rote problems. Either provide education on this piece, or mix in some easy stuff. Good luck I love your idea.

He's rapidly gaining followers on twitter and rumors range from he is an insider at openai, to he is a next generation model larping as a human. Anyone have insights? He's interesting to follow nonetheless.

Has anyone implemented this successfully? I'm 90% of the way there but I'm not sure I fully understand the node expansion process. Results are so-so. Improved for sure, but not solving my toy math problem...