Everybody’s doing it. I guess I need to do an AI, too. In my home, I have a few different tools that use generative AI and LLMs. I talk to my Home Assistant Voice Preview voice assistants which leverage a self-hosted Ollama running llama3.2. I use OpenWebUI, Tabby for as an experimental coding assistant. I use DeepInfra for larger models that don’t fit on my own GPU.

However, my problem is that each program supports different providers and models. Some support OpenAI style APIs to any provider, some only support Ollama APIs. If I wanted to forward my Home Assistant queries to DeepInfra, it wasn’t easy to do because there wasn’t an integration. If I wanted to change the model that Tabby uses between different models, I had to redeploy the service.

Digital advertising is everywhere nowadays. However, they are actually a giant risk to privacy and now, safety. To be successful, digital advertising depends on showing you highly targeted advertisements, which ultimately incentivizes them to build up profiles about you via your browsing history, search queries, location, demographics, and even behavioral patterns. More data about you means they can find ads that you’re more likely to be influenced by.

RKE1 (Rancher Kubernetes Engine 1) was Rancher’s first way of automatically deploying Kubernetes to a cluster. Think of it like minikube or on-prem EKS or K3s. Three years ago, it was marked as end of life (EoL) with the last release being July 2025. They have no migration guide and their strategy is just rebuild the cluster. I have 3 nodes a bunch of services running in Kubernetes. I don’t want to take everything down and rebuild it all. Let’s rebuild it while the plane is in the air.

Last year, I setup a Christmas lights show at my house. I started with some basic light sequences just to learn. I wrote a post on the basics. This year, I upped the ante and added more lights and starting making sequences linked to music.

I have one light controller running Falcon Controller/FPP, a Kulp K8-B controller. How do I get sound out? I looked at options for getting sound out to a speaker. For my first pass, I decided to push it to a Sonos Move speaker since I had one.

This is a continuation of my previous post where I talked about the challenges of using serverless/Function as a Service (FaaS) compute systems for ETL (Extract, Transform, Load) jobs. It sat in my drafts folder for a long time, so I just decided to publish it as is.

I used to work at AWS, and predictably we used a lot of AWS cloud services. In many cases, when an engineer looks for a service platform, they’ll often go directly to AWS Lambda because “it’s Serverless” with the justification that it’s simple and alternatives are too complex and not worth considering.