Has anyone tried Claude AI Desktop? I am particularly interested in troubleshooting for leap 16.0 but curious about the overall experience.
I would be wary of giving a hallucinating LLM access to your PC. 
For troubleshooting, you could copy-paste the error into the chatbot’s web GUI. 
2 Likes
Thank you
Interested in why you say hallucinating LLM. I understand LLM is meant as Large Language Model and I also understand not giving access to the PC. That is why I ask the for experts opinions. I also reference chatbot opinions.
( Large Language Model : In technology and generative AI , an LLM is an advanced artificial intelligence system trained on vast amounts of text data to understand, generate, and process human language.)
AI does often hallucinate, Claude included. Part of the cause is the amount and quality of info provided in the prompt, but it’s not all. It can f.e. come up with wrong results for Tumbleweed when one calls that “openSUSE Tumbleweed”, since that will include the well documented Leap info, where Leap is very different in some aspects.
In ~half of the cases where I use AI, I still need to tinker with the outcome, or correct the AI chatbot, where it gladly comes with an entirely different outcome. That said, it has gotten better and better and has been helpful to me.
What I’d rather see ( and AFAIK is being worked on by some community members ), is a generic openSUSE LLM, maybe a per distro, that comes with clickable instructions and explanations to teach the user what is being done.
OFF Topic @moderators Can somebody maybe move this to Open Chat ?
2 Likes
Good article in this light:
It’s a “harness”.
What’s more, various studies have suggested that model performance (and cost) varies significantly with the harness used, due to different design choices. For example, the Pi coding agent relies on a minimal system prompt of about 200 tokens. Claude Code by comparison uses a system prompt of around 10,000 tokens (or did until last month when Anthropic trimmed the system prompt by about 80 percent). The same model will produce different results with different harnesses.
1 Like