See Part 0 for the latest table of contents and sample code. New chapters will be added over time. Chapter 4: The Provider Trait and a Scripted Mock The previous chapter gave abcb a conversation shape. We now have Role, Message, and Session, and those types can leave memory as
See Part 0 for the latest table of contents and sample code. New chapters will be added over time. Chapter 3: Modeling Messages, Roles, and Sessions The previous chapter gave abcb a place to live. This chapter gives it the first thing an agent framework actually needs: a conversation. That
See Part 0 for the latest table of contents and sample code. New chapters will be added over time. Chapter 2: Setting Up the Workspace The first code chapter does not call a model. It does not parse a tool request. It does not know what an agent loop is
See Part 0 for the latest table of contents and sample code. New chapters will be added over time. Introduction: Building the Loop The word "agent" is already overloaded enough that it can hide more than it explains. Sometimes it means a chatbot with tools. Sometimes it means
Table of Contents * Chapter 0: Preface: Why I Am Building This * Chapter 1: Introduction: Building the Loop * Chapter 2: Setting Up the Workspace * Chapter 3: Modeling Messages, Roles, and Sessions * Chapter 4: The Provider Trait and a Scripted Mock Sample Source Code: * Github: dikoko/building-a-local-first-agent-framework-in-rust Preface: Why I Am Building
Last Tuesday night (5/19), I went to the Computer History Museum to hear Joanna Stern talk about her new book, I Am Not a Robot: My Year of Using AI to Do (Almost) Everything. This was one of those Bay Area events that felt almost too easy to attend.
Not long after K-Pop Fireworks Night at Oracle Park, the real K-pop moment came much closer to my home. BTS came to Palo Alto. Technically, the venue was Stanford Stadium, not Palo Alto, and technically this was part of the BTS World Tour Arirang. But from our side of town,
The Giants are one of my favorite recurring topics on this blog, even when the team itself is not giving me much help. This season has been hard to watch in the usual torturous way of 2020s Giants baseball. Still, Oracle Park keeps pulling me back. Yes, that's
Last month, I got an invitation to Stripe Sessions 2026, and I still do not know exactly why. Maybe it was because my title says Director of Engineering at Toss. Maybe it was because I have been spending too much money on AI tools, and many of those payments quietly
The next book club selection with Ms. Salle was the Earthsea Cycle by Ursula K. Le Guin. The first book I read in the club was The Dispossessed, and I became almost obsessed with the beauty of her prose. I even tried to imitate her writing style for a while.
In the previous post, I set up Gemma 4 locally with MLX and connected it to my usual local-model workflow. That post was mostly about the environment itself: getting the model running, exposing it through an OpenAI-compatible endpoint, and making it usable from tools like Open WebUI. Once that was
In my previous post, I explored what it looks like to run coding agents with local LLMs on an M3 Ultra Mac Studio. That write-up was mostly about the environment itself: treating a single machine as something closer to personal infrastructure, capable of hosting multiple models and running them continuously
