what made local AI click for you

[quiettoken]

Posting this in case it saves someone else a little time.

I'm helping a small team set up a local workflow. We don't need a giant stack, but we do care about prefix caching.

Simple and boring is a valid answer.

[quiettoken]

Posting this in case it saves someone else a little time.

Thanks, that helps. I ended up in about the same place using llama.cpp. I got better results once I stopped chasing one metric and just looked at whether the workflow felt usable.

[dockember]

That matches what I was seeing. That matches what I saw on Threadripper box.

[ggufwire22]

Thanks, that helps. I ended up in about the same place using MLX. I got better results once I stopped chasing one metric and just looked at whether the workflow felt usable.

[rocmplane]

Thanks, that helps. I don't think your conclusion is far off. I got better results once I stopped chasing one metric and just looked at whether the workflow felt usable.

[quiettoken]

Interesting, I hadn't tried that. That matches what I saw on Threadripper box. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.

[ggufwire22]

A couple things I'd check first. If you're on SGLang, I'd confirm the version first and then reduce variables until the behavior is obvious. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.

[rocmplane]

I hit almost the same issue last night.

[pixelcrate82]

My first guess would be this: On 5090, the easiest win is often to back off the aggressive settings and see whether the problem survives. Not glamorous advice, but repeatable baselines save a lot of time.

[rocmplane]

I'd start with the simpler stack. I keep promising myself this is the last rebuild.