I have a question with regard to updating the various recipes. I’m working at EPFL in Switzerland, at C4DT, and we’re currently looking into sovereign solutions for web services.
I put together an example at https://cloud.c4dt.org with various recipes. I really appreciate all the work which has been done here. The tradeoff between usability / flexibility of coop-cloud is unique, from what I saw. However, a lot of recipes are old - jitsi (> 1year), nextcloud is a bit behind (trying to get involved there), and others.
I’m a programmer since 1981, but I really appreciate what LLMs can do. So a lot of the code I’m writing nowadays comes from LLMs. But I know that some people think that LLMs are devil’s work (well, they are…), and don’t want to touch them. So my question is this:
would you appreciate me getting more involved over the coming months in updating various recipes by carefully proposing PRs done by LLMs, and tested on our server? Or is that a no-go for you?
I cannot promise I’ll hang around for long. We’re currently looking at various ways of deploying things, and I couldn’t convince everybody yet with the good foundation of coop-cloud But having all these people writing docker-compose files in their corner is definitely sub-optimal. So if something like coop-cloud can become a collection of up-to-date and easy-to-deploy repository of docker-compose files, that would be awesome!
In Autonomic, we’ve actually been playing with a recipe maintenance and upgrades systems that is LLM assisted.
Here is the output of one aspect of it:
However, in Autonomic, we recognise the wider context around LLMs and other forms of AI as a as an attempt to fundamentally reshape the relationship between labour and capital in many sectors of the economy so we remain highly critical of the companies behind these technologies whilst investigating their uses.
Often as the open source/co-op community, we can only do so much. I’m hopefully going to be visiting the Barcelona Super Computer this year which is a publicly owned resource that is being used to train public models.
Thanks - very interesting. I don’t understand how Autonomic and Coopcloud is linked - are you a mirror of coopcloud?
1 Like
kawaiipunk
(KawaiiPunk - Autonomic Co-operative / Tech Care Co-op)
5
We’re a tech worker co-operative. We were the founder member of the Co-op Cloud federation which democratically govern the Co-op Cloud project. We use Co-op Cloud to deploy and maintain our infrastructure (when appropriate).
I’ve increased your trust level, you should be able to post a link now. Sorry we had to lock things down due to AI spam bots…
Incidently there is a thread on development an AI policy for Ansible I started here:
I’m still boycotting AI but realise that for things like writing Ansible playbooks a local LLM is totally viable and is something I’ll have to get around to setting up at some point in the future.
But having all these people writing docker-compose files in their corner is definitely sub-optimal. So if something like coop-cloud can become a collection of up-to-date and easy-to-deploy repository of docker-compose files, that would be awesome!
