I’m becoming more wary of big tech these days.

Slowly, I’m trying to migrate away from certain software in favor of open source. I was tempted to switch to Linux, but for now, I’ll keep using Windows.

I came across a video explaining how to host your own private, secure LLM models. Woah—neat. I gave it a go and, within minutes, I had my own private DeepSeek-powered model running on my laptop.

My main reason for doing this is to test how AI might help me with my finances. The more details I share, the better the answers—at least in theory. Since those details are personal, I’d rather not use a closed, big-tech-owned model like ChatGPT.

I’ve barely tested my private LLM (which I use with LM Studio). It’s definitely not as quick, and it guzzles GPU power—my fans go into hyper-mode whenever I prompt it. But no worries; for my particular use case, I think it’ll be good enough.

P.S. I recently re-designed this site using ChatGPT. Still can’t believe that I can generate working code with so much ease.