We need a satellital #Nostr relay 🛰️

I have successfully run the #LLaMA LLM (from Facebook) locally on my GNU/Linux laptop using llama.cpp and I have achieved better results and performance than what I expected. It was the 7B model and I was also able to run the 13B model. You can expect much better results with larger models!

It is exciting to see progress towards the possibility of a self-hosted #ChatGPT that runs locally.

https://github.com/ggerganov/llama.cpp

https://www.ndss-symposium.org/ndss-paper/improving-signals-sealed-sender/ (2021)

Huh

Just created a badge for #Nerostr users! https://nostrgram.co has a great implementation of Nostr #badges. Take a look there!

https://badges.page/b/naddr1qqxxuetjdaehgu3dw4ek2uszyz275r3fznx5kqsd6agkyquq4umxma356hcxw23snr4fwm7t94uljqcyqqq82wgmejkf8

I made a #Telegram bot that allows you to interact with OpenAI API at full. It's pretty easy to self-host! With it, you can:

- Generate images with Dall-E
- Chat with ChatGPT
- Transcribe audio messages to text with Whisper

https://codeberg.org/pluja/openai-telegram-bot

#OpenAI Is Now Everything It Promised Not to Be: Corporate, Closed-Source, and For-Profit

https://www.vice.com/en/article/5d3naz/openai-is-now-everything-it-promised-not-to-be-corporate-closed-source-and-for-profit

Web-Whisper, the frontend for #OpenAI #Whisper I created, now can use the OpenAI Whisper #API (that was introduced just today!). You can now choose to run whisper 100% locally, or leverage the workload to OpenAI's servers (which are pretty fast) sacrificing privacy as your files will be sent to a 3rd party for processing.

https://codeberg.org/pluja/web-whisper