There are many language-learning apps, but almost none that focus on improving conversational Hindi for kids.

Made this web app for my nephew, based in Singapore, after watching him struggle to find anyone to practice Hindi with outside of family calls (since most of his friends are Chinese). The idea is to have a 24x7 partner to speak with Hindi and make it fun. This can complement the formal Hindi classes that most kids of Indian diaspora parents take.

My nephew started using this and is enjoying it!

Link: https://www.hindispeakingtutor.in/

Looking to hear feedback from the community!

this is interesting but anyone has tried to create a small scale version of their own search engine using only their bookmark links. I am surprised why that's not more common. If someone has 5k+ bookmark links and wants only results from those links given a query, it should be straightforward.

YaCy.net does something like this AFAIR

Yes I think having an LLM generate such OnlyConnect style questions (with the right prompting) should solve for the problem this benchmark seems to have of LLMs, most likely, being trained on past years OnlyConnect questions.

Thanks for views. I can understand that such a tool might not be relevant for someone with your reading style. However, this is no manipulative algorithm. It is just a tool to help you decide your next read based on your own reading history.

I would assume the argument could be made that "can be run locally", "w2v can be trained faster than other models" and "doesn't cost per-call to use" to be some arguments for why someone would try this.

Very good resource on building from the basics in a concise manner.

I want it to be more like a proof of concept. Building a scalable and fully usable long term will require a lot of investment which I don't think would be the right thing to do until I have some validation from users. Does that help?