juraj on Nostr: I did not expect this to be the case, but for coding shell script, I had much better ...

I did not expect this to be the case, but for coding shell script, I had much better success today with the new LLaMA-3.1-405B (through venice.ai) than through paid chatgpt. ChatGPT did not even get what I wanted, LLaMA wrote almost correct shell script in the first run.

In case you want to fix the metadata of media transmitted through Signal for importing to some photo app:

https://gist.github.com/jooray/032343f81885b4fdfd633e0ed4116526