WOAH! Hardware LLM Acceleration. This is so much better than my current setup in terms of performance, but I really like my keyboard on the GPD Win 4 (2025)

one day it crashed for the last time. I remember when it was only because I came home to a HDD making sounds like a demon giving birth. The replacement was a mac and never looked back.

I could, but only because I've done one in javascript before (for those pebble watches). Give it that as context to know how to pull the API... sqlite mcp for storage... seems doable

I couldn't find a good enough one so I vibe coded one:

This README was the last time I tried to let AI be "funny".
github.com/chisleu/load...

Did you find any models that worked well for you? I've found the new devstral-small-2507 to be useful for things. It's hard for me not to compare everything to Gemini 2.5 pro and Claude Sonnet 4.0 though.

I'm not a fan of VSCode but Cline was so featured that it's all I use now.

So they have "Made with AI" and "Made by AI" icons you can put on your site to distinguish between the two.
I thought about including it on my personal blog, but I edit the final version and there is zero profit motive. Maybe that makes a difference.

congrats dude

caches are good and storage is cheap ;)

!!

I don't use k8s directly because it's so damn complicated, but AWS ECS is really straight forward. I can get auto scaling API containers in a blink with Claude and terraform. But it's only really useful for http/stateless IMHO