Weekly Notes 01/2025

Uma met her another set of cousins this week/end. We visited the local temple to see the elephant with her on Saturday. Uma wanted to pet the elephant :) On Monday, we went to the beach to see the sunset. Uma loves water but has yet to figure out the ocean and waves to enjoy the beach thoroughly. Uma also learned a lot of words this week. Kids learn from kids more than adults.

• 💻 Workwise, it's a slow week. I have taken a few days off to spend time with family before returning to Bengaluru this weekend.
• 🎦 We watched Rifle Club. I liked it. It's a home invasion/action genre movie.
• ✍ I wrote my year-end post capturing the highlights of 2024. I created a gallery post of featured images from Weekly Notes/2024. It's a pictorial representation of the year.
• 🕸 I got to attend the Wiki Conference Kerala 2024 like last year. It was fun meeting folks. They are generally kind and collaborative folks. Hopefully, I get to do more than just participate.
• 🕸 Anand (my friend and a fellow Infyblogger) has been writing weekly notes for a year now. He uses it to gather what he has learned over the week.
• 🧑‍🏫 I am also working on refreshing MAD1 and MAD2. I hope to make it more interesting than what it is now.
• 🤖 I have been using Alphine.js and HTMX in personal projects. They go well with Flask or Bottle micro libraries. I love small libraries, so I remember them well and can debug easily. I often use an LLM to code, but knowing the libraries improves my code reviews. A good code review is necessary if the project has to be alive for a decade or more. I also have a simple RAG built by formulaic.app using my favorite stack's documentation and some tech notes. I still use Zeal Offline Documentation Browser for technical documentation, but an additional search and answer layer (RAG) using locally running LLMs would make it more useful.
• 🤖 I use an old machine (LC230) as my daily driver. My access to computing superpower is through a network (local or internet). I use Open WebUI to chat with LLMs. It kind of gives me one interface to access the LLMs across networks. I run the WebUI locally on my machine; that way, I can access previous chats (history) even if I don't have (offline) access to LLMs. It's working well for now. Of course you can also build your RAG inside OpenWebUI.