amanvir_

Hey, I'm the creator of Guten!

Thank you, OP, for posting this, and thanks to the community for all your support!

To answer some common questions/comments/concerns:

- Totally agree with the sentiment regarding screens being a big problem in today's day and age. The main reason I wanted to make Guten was so that I could start my day off reading something on paper instead of staring at my phone. It also helps that you can't doomscroll on a receipt ;)

- I also love Little Printer - it seemed like such a cool product, but I unfortunately never had the chance to purchase one before it got discontinued. This is my attempt to bring back some of the functionalities in Little Printer that I'd find most useful in my day-to-day.

- BPA in thermal paper was a concern of mine as well, but I thankfully found some BPA-free thermal paper on Amazon!

Hey HN!

In my freshman year of high school, I built a simple, multiplayer buzzer website for my school’s quizzing club. That site ended up growing tremendously, and has since been used by hundreds of thousands of people. It's been 4 years since then, and I recently finished my first year at university.

Now, I've decided to build something even bigger: an audience engagement platform to augment your events and gatherings through live trivia and quizzing.

I've been working on azigy.com for the last few months, and I'm excited to finally be launching it!

Hey HN!

I'm a first-year college student who thought it'd be cool to leverage GPT-4 and OpenAI's TTS models to create an assistant that creates an educational "video" on the fly on any topic!

Currently, Sal is an early-stage prototype, but I think there's likely some utility in using LLMs and other AI tech to further the autodidactic side of education.

Let me know your thoughts on the demo!

I've had this experience with RAG systems as well, but I've been wondering if structuring unstructured text (like a novel) in the form of a knowledge graph may solve this problem.

Perhaps coupling a vector embedding search with a knowledge graph created from unstructured text could lead to more informed answers from LLMs?

It seems as though some companies and researchers are already experimenting with this idea: https://www.nebula-graph.io/posts/graph-RAG.

Hey HN!

I recently started my undergrad, and after a few weeks at university, I got inspired to make a tool to help others out there.

Introducing lekchur, a free tool that transcribes your college lectures (or really any form of speech) in real-time, right from your browser. It loads the whisper-tiny-en model locally using the Transformers.js library and performs all audio processing and transcription solely on the client side.

The output on its own may seem rather cumbersome to parse through, but I've been finding it useful to feed the final transcripts to LLMs like ChatGPT and Claude.ai to get concise notes and summaries.

Claude has a nice file upload tool, so I've developed a neat workflow where I can record, transcribe, and get summarized notes of lectures on my phone!

Here's a demo showcasing that workflow with lekchur on my iPhone 13: https://youtu.be/WQk86-9BCaI

For as long as I can remember, I've always been drawn to islands, especially the most remote ones on our planet. I've spent hours scanning Google maps for tiny specks of land and spent more hours researching them, often uncovering their rich histories in the process. These are some of the more obscure islands that I find interesting.

The “waterbed theory” is probably what you’re looking for: https://en.wikipedia.org/wiki/Waterbed_theory