Self-taught homelabber. Local AI enthusiast
I've been tinkering with code when the ZX Spectrum showed up at our house when I was five. Decades later I'm still the one who gets called when someone's PC dies or their dad needs a new build.
A year ago I turned that hobby into a proper homelab. Twelve months in I've got a multi-device LAN of older but very capable tech. There's a roadside dumped PC in the mix, plus a few CEX bargain bin components from before the memory crisis hit. Proxmox, Home Assistant, my own domain, secure remote access. The whole thing runs as a complete Docker stack.
Half the fun is breaking it. Sometimes it's 1 step forward and 2 steps back, but failure is the only way I learn. I document the broken parts so the next person can skip them.
The bit I'm most proud of: I spent the first six months building it all by hand. No AI agents, no Copilot autocomplete, no shortcuts. Just YouTube tutorials, GitHub issues, and strangers in forums who didn't owe me anything. I wanted to understand the stack, not just inherit it.
Around the six-month mark, I started letting AI into the workflow. I've gone a bit mad with it: agentic systems, local LLMs, self-hosted inference on my own GPUs. I'm building toward digital sovereignty. Cutting the cord from a monthly AI subscription the same way I cut the cord from Netflix.
There's a tension I want to name openly. I'm wary of what generative AI is doing to creative industries, and I won't use it to replace human craft. But I don't think the answer is to reject the tool outright. The open-source community is so close to proving you can run serious, capable models on your own hardware, under your own control. We don't have to choose between capability and conscience.
The public conversation about generative AI is loud and polarized. My take is simpler: it's nuanced. You can use this technology ethically, you just have to choose to. It's a groundbreaking tool, but it's still just a tool. What you build with it is the part that matters.
DMs are open if you want to talk shop about homelabs, self-hosted inference, or how to keep your hands on the steering wheel while the world argues about who should be driving.