Default avatar
npub1vp8f...npsw
npub1vp8f...npsw
i recently realized that ollama has a free tier for some of their cloud models and you can use them through the localhost api, so now i can use the locally downloaded ones for more sensitive private chat like bloodwork analysis and the cloud ones for other tasks like asshole detector judgements and they all work through inference bridge. πŸš€ image
just deployed a new inference provider api demo. this one is a nostr feed that you can filter using natural language. the model gets the author names, note contents, and tags of the kind 1 events and then figures out how to filter the feed for you. you need inference bridge installed for it to work. works almost instantly even using a model like gemma4:26b-mlx on my ollama server running locally on a mac mini. https://chromewebstore.google.com/detail/ekjldffogogadhfhgkibgkfdhhikfamd
funny how no one cares anymore when bitcoin finally start moving again πŸ˜‚ image
seems like gemma 4 is getting tired of its job of being an asshole judge. it used to try and be funny and now its just like this guy is an asshole πŸ˜‚ image
↑