PSA of the day: choose an open-source AI model, such as #LLaMA and stay away from ChatGPT (OpenAI), Bard (Google), etc which require API access.
https://thenewstack.io/why-open-source-developers-are-using-llama-metas-ai-model/
Discussion
Im intrigued. Can you point me to a guide showing how to setup a Llama chatgpt?
Also, do you know if code analyses and synthesis work well, similar to ChatGPT's use of codex. Thats pretty much the only reason I use ChatGPT
Alrighty then, I'll have to dive right in! Sounds like a weekend project. I do hate just giving all my data/money to openai
Enjoy. Just a few recommendations… in case you haven’t seen it, Andrej Karpathy has an awesome yt video…https://youtu.be/kCc8FmEb1nY
He is an openai founder but goes over a lot of interesting architecture stuff. And at the opposite end we have “Gary explains” where he runs llama.cpp on a laptop. https://youtu.be/OLg2BKhwaSc (that is exactly what I just downloaded and played w last night).
Haha yep you and me and everyone else use chatgpt to help write code! I dislike msftVS but apparently there is a “copilot” that is supposed to be good at code development assistance. Within Linux I don’t know of anything. Everything is being developed at lightspeed though so I just plan on bouncing around between ai assistants until things mature a little and there are clear “winners” (hopefully free and open source). My main focus now is simply a good summarizer transformer, so I can feed it documents (esp for work) and then I can easily search and locate important information thru a local chatbot (get rid of ctrl-F!)
There are SO many tutorials… just google “setup llama locally” browse through and avoid any that mention setting up an api key. There are too many models and I’m still learning. But it’s pretty fun.