Just installed Ollama 9 days back. Just kicking the tires so far to get familiarized with it before doing real AI hobby work. Any advice would be nice.
Got it, openclaw, and Claude-code installed on my Threadripper 64 core 7985WX with 256GB's of ram and dual 5090's on Ubuntu. Currently have:
NAME ID SIZE MODIFIED
gemma4:31b 6316f0629137 19 GB 40 minutes ago
gemma4:26b 5571076f3d70 17 GB 45 minutes ago
nemotron3:33b-q8 74d89c84a443 36 GB 7 hours ago
granite4.1:30b-q8_0 0f7a2b54edab 30 GB 7 hours ago
qwen3-coder-next:q8_0 3f68e12b44ee 84 GB 8 days ago
qwen3-coder-next:latest ca06e9e4087c 51 GB 9 days ago
qwen3.6:35b 07d35212591f 23 GB 9 days ago
gemma4:latest c6eb396dbd59 9.6 GB 9 days ago
I keep reinstalling everything to make sure I'm not leaving anything out before I make a snapshot of the whole env so that I can use this both as a subject of research and for research. I don't want remembered things or USER customization to make things not reproducible.
What are the essential tools/skills/plugins/... for doing AI research and code development?
Once I get this like I want I'll start hammering it with AI experiments. Right now I'm looking at whether I can use my openai pay-per-use account(gpt-5.4) as a open in an emergency fall back if my local models can't figure something out after some number of tries.
I've been ripping off free usage from chatgpt, gemini, and claude.ai for a long time now.
While I understand my local models can't compete with them and ability to automate things in a feed back loop interests me.