Goose, open code, zed all powered by ollama running locally. I feel like a renaissance man!

nostr:nevent1qqsyuqd8l4hzzzcn3jxlsj7lfvfxcjy376w468mazllq99vsq0dgypqpr9mhxue69uhhq7tjv9kkjepwve5kzar2v9nzucm0d592wen9

Reply to this note

Please Login to reply.

Discussion

How does it compare to Claude?

It’s getting closer! The open models are catching up to frontier proprietary models and at the same time getting smaller and more powerful at the same time - look at gpt-oss:20b & NVIDIA nemotron-3-nano:30b & devstral-small-2 they are excellent examples

Okayyyy looks like I’ll checkout what you mentioned in the OP and have a go!

πŸ‘ high spirits

Any tutorials out there on these tools? I downloaded ollama on my start9 and never did touched it

Try zed for editing, opencode for the AI stuff and nanoGPT for your API

Even better if you want good models with some privacy ;)

Also which model in ollama?

nostr:nevent1qqstge2wvhdpzq0dm5pe9nee03pn987y8kxhkstey80n93kvg2k9a9cpr4mhxue69uhkummnw3ez6ur4vgh8wetvd3hhyer9wghxuet5qgsgzu4eypfy0h07nxmcxvs8stgrztarqksen7etaz37v437yz60pcsrqsqqqqqpydehyy

I have so many local models but I really just want a bunch of mini LLMs agents? that can "read" a book that I can then talk to as if it was the author then another and another each one ingesting a single text then they fight to advise me!

What hardware are you running?

M4 Mac mini 64gb - it is solid hardware for small local AI but not unattainable levels of performance for the avg computer user

How would you compare goose with opencode on strength and weaknesses basis?