Selfhosted

for coding tasks you need web search and RAG. It's not the size of the model that matters, since even the largest models find solutions online.

I have it roleplay scenarios with me and sometimes I verbally abuse it for fun.

As cool and neato as I find AI to be, I haven't really found a good use case for it in the selfhosting/homelabbing arena. Most of my equipment is ancient and lacking the GPU necessary to drive that bus.

Learning/practice, and any use that feeds in sensitive data you want to keep on-prem.

Unless you’re set to retire within the next 5 years, the best reason is to keep your resume up to date with some hands-on experience. With the way they’re trying to shove AI into every possible application, there will be few (if any) industries untouched. If you don’t start now, you’re going to be playing catch up in a few years.

I installed Llama. I've not found any use for it. I mean, I've asked it for a recipe because recipe websites suck, but that's about it.

Sorry, I am just gonne dump you some links from my bookmarks that were related and interesting to read, cause I am traveling and have to get up in a minute, but I've been interested in this topic for a while. All of the links discuss at least some usecases. For some reason microsoft is really into tiny models and made big breakthroughs there.

https://reddit.com/r/LocalLLaMA/comments/1cdrw7p/what_are_the_potential_uses_of_small_less_than_3b/

https://github.com/microsoft/BitNet

https://www.microsoft.com/en-us/research/blog/phi-2-the-surprising-power-of-small-language-models/

https://news.microsoft.com/source/features/ai/the-phi-3-small-language-models-with-big-potential/

https://techcommunity.microsoft.com/blog/aiplatformblog/introducing-phi-4-microsoft%E2%80%99s-newest-small-language-model-specializing-in-comple/4357090

It'll work for quick bash scripts and one-off things like that. But there's not usually enough context window unless you're using a 24G GPU or such.

7b is the smallest I've found useful. I'd try a smaller quant before going lower, if I had super small vram.

Converting free text to standardized forms such as json

Have you tried RAG? I believe that they are actually pretty good for searching and compiling content from RAG.

So in theory you could have it connect to all of you local documents and use it for quick questions. Or maybe connected to your signal/whatsapp/sms chat history to ask questions about past conversations

I've integrated mine into Home Assistant, which makes it easier to use their voice commands.

I haven't done a ton with it yet besides set it up, though, since I'm still getting proxmox configured on my gaming rig.

I've used smollm2:135m for projects in DBeaver building larger queries. The box it runs on is Intel HD 530 graphics with an old i5-6500T processor. Doesn't seem to really stress the CPU.

UPDATE: I apologize to the downvoter for not masochistically wanting to build a 1000 line bulk insert statement by hand.

I've run a few models that I could on my GPU. I don't think the smaller models are really good enough. They can do stuff, sure, but to get anything out of it, I think you need the larger models.

They can be used for basic things, though. There are coder specific models you can look at. Deepseek and qwen coder are some popular ones

Currently I've been using a local AI (a couple different kinds) to first - take the audio from a Twitch stream; so that I have context about the conversation, convert it to text, and then use a second AI; an LLM fed the first AIs translation + twitch chat and store 'facts' about specific users so that they can be referenced quickly for a streamer who has ADHD in order to be more personable.

That way, the guy can ask User X how their mothers surgery went. Or he can remember that User K has a birthday coming up. Or remember that User G's son just got a PS5 for Christmas, and wants a specific game.

It allows him to be more personable because he has issues remembering details about his users. It's still kind of a big alpha test at the moment, because we don't know the best way to display the 'data', but it functions as an aid.