🤖llms

Why I’m Not Worried About My AI Dependency

This is a solid take on relying on AI for design or product needs.

Enough AI copilots! We need AI HUDs

Some good food-for-thought and examples of non-chatbot AI UIs…

Building a web search engine from scratch in two months with 3 billion neural embeddings

This is a REALLY impressive deep dive into making your own search engine. Wow.

Building a Personal AI Factory (July 2025 snapshot)

Pretty elaborate workflow, but always interesting to see what’s working well for people with coding agents.

Add Recent History Section to Popup by Copilot · Pull Request #2 · aluhrs13/JankExt · GitHub

Really impressed. I’ve been dabbling with GitHub Copilot a bit to just get a feel for it. This new “assign a task” feature is impressive. This is a PR where you can look at the whole prompt and the result. Not other interaction from me. I’ll give it more a detailed review, but it’s fully functional and seems to be what I want.

Building TMT Mirror Visualization with LLM: A Step-by-Step Journey

Love seeing how other people approach prompting. A good step-by-step for anyone who wants to work on their skills.

A cheat sheet for why using ChatGPT is not bad for the environment

Probably won’t blindly trust some of this, but seems like a reasonable counter to one of the more common anti-LLM arguments.

GitHub - leafeye/lunchmoney-mcp-server

I’ve been using LunchMoney for budgeting and it’s a great tool. Haven’t really had a chance to use MCP much, so I’ll definitely need to go play with this.

Understanding the recent criticism of the Chatbot Arena

Another day another LLM drama. Anyone using LLMs for serious purposes needs to make sure they have evals representative of their real use-cases when comparing new models. Benchmarks and leaderboards mean nothing.

GitHub - microsoft/markitdown: Python tool for converting files and office documents to Markdown.

Saving this become I’m sure I’ll want it at some point for LLM stuff.

You're using GPT-4o image generation wrong

The /llms.txt file

Works on My Machine

Everything I've learned so far about running local LLMs

As my friends and family’s “what is AI?” person, this overview of running local LLMs is great.

Your AI Product Needs Evals

I keep coming back to this post when I play with AI. Such a great summary of what’s needed to work effectively.

Embeddings are underrated#

Creating a LLM-as-a-Judge That Drives Business Results

A more recent post about LLM evaluation, and a bit deeper into using experts to support development.

Thinking Like an AI

Great summary of how LLMs work from Ethan Mollick

What is a Transformer?

I gave a presentation about AI basics a few weeks ago and wish I had this available at the time.

Compare PDFs

Simon has some of the best AI knowledge today, and I love that he’s sharing how he prompts and approaches some of the problems he’s solving.

AI’s $600B Question

The economics of the current AI boom are staggering.

picopilot

Great practical little demo of LLMs in action and a basic VS Code extension.

Clustering ideas with local ML/AI models

A solid intro to a good use-case for vector embeddings and LLMs using entirely local AI models.