New Qubes AI template proposal

[corporateblush]

I think we’ve reached the point where yours and mine further debate won’t change either of our minds, and I respect that. But there is one last thing I’d like to clarify for anyone reading this in the future:

My position: ‘I don’t need a stock LLM template myself, but others who do (want to) use AI might benefit from having it available as an option.’

Your position: ‘I don’t need a starting point offered to me, so there is no value proposition to weigh.’

The difference matters. You’re treating your personal preference as if it settles the question for all the users. I’m saying that my own needs shouldn’t define what’s best for users who might genuinely benefit from this proposal.

We can agree to disagree on whether Qubes should ship with a stock AI template. But at least we’ve been honest about where both you and me are coming from.

[FranklyFlawless]

Fair point, but I am also thinking about the Qubes OS team taking on your proposal seriously and maintaining it indefinitely. If your proposal was a community template instead, then I have no issue, anyone can create, maintain, and abandon it anytime.

[queen]

@blitz

Even for users that do use AI, I really don’t think that they would want an LLM installed by default, as they should have the right to choose which LLM to use.

My opinion as an LLM/AI user: Absolutely right. I don’t want a pre-built LLM/AI.
I build the LLM/AI exactly the way I need it for each project.

The Qubes team already has enough on its plate. AI is a much bigger project.

Added 2026-09-03T18:20:00Z:

@FranklyFlawless

Fair point, but I am also thinking about the Qubes OS team taking on your proposal seriously and maintaining it indefinitely. If your proposal was a community template instead, then I have no issue, anyone can create, maintain, and abandon it anytime.

The community template is good. I agree with that, too.

Or community guides like: Secure AI Inference with Qubes OS: A GPU Passthrough & Ollama Guide

[queen]

@corporateblush

Local LLMs are inevitable. Users will go to places where they don’t have to be experts in order to set it up and start to use it. And offering such a template would not be the same as an offering of a “Windows” template as a trick to attract new users

I only know of two user groups:

• Those who are regular end users will install the ChatGPT app or the Claude Cowork app. They don’t need an AI/LLM template.
• The others are those who work intensively with AI/LLM. They don’t need a template because they build their own.

[corporateblush]

Thanks for a tons of likes @queen.

@queen

I only know of two user groups:

• Those who are regular end users will install the ChatGPT app or the Claude Cowork app. They don’t need an AI/LLM template.
• The others are those who work intensively with AI/LLM. They don’t need a template because they build their own.

Yes I know. That is because people are often acting like ducks flying with their butts ahead. They know where they are coming from, but don’t know where they are heading to.
I am the opposite. My eyes are ahead . I see where this goes, not what was and is.

I can guarantee you will start to notice them soon, but that would be “coming from” perspective, too late then.

That was enough from me. I am off the forum. I have a LLM to train. See you in a year or two to necro-bump you on this topic. My bet: some of you will have “legit” excuses for current “nays”.

Thanks for the company. I wasn’t always sweet, but never malicious in my intentions. In my culture, we are always direct, so you always know where we are at with each other.

Cheers