Checking in: Resetting

Since I last posted... I took a small break. I got a discord bot working. All it does is barf responses in a discord channel from OpenAI's API - nothing magical. The only neat thing is that it takes the whole chat as context - so it can engage with multiple people at once. It's kind of clunky since it's designed to interact 1:1 - so if you all say goodbye in the thread it addresses each person individually. It's like talking to someone who's sole interest is having conversations one on one. I'm going to work on implementing MCP's. Initially when anthropic shared the paper I tested it. I think now that more of the community at large has adopted the standard it'd be worth getting into again. Not as, roll your own as it once was. I'll keep you posted. 

Day 1: Progress

Little to no progress so far. Realistically, I expected today to be slower at work and for the path forward to feel clearer. Instead, I’m stuck—still anxious about making architectural decisions at this level. Years of troubleshooting a product built on shaky foundations have made me skeptical of every option in front of me.

Like crushing a spider… there’s a brief moment of relief, only to watch in horror as thousands of its offspring emerge from its corpse. Each line of code feels the same—progress made, ground claimed—yet in my mind, I see a thousand potential bugs, limitations, and design flaws spiraling into existence. I know this is for me, but the scars remain. I didn’t build that product, didn’t make those bad calls. But I knew those developers, those project managers. They had good intentions. Even the best-laid plans can derail.

Here’s my attempt at silencing the noise and pressing forward.

Currently Listening To: Borderline // Tame Impala

Day 1: Touch Base

I was able to get some code roughed out yesterday. I'm not entirely happy with where things currently sit... like I said it's really quite rough. My goal is to get to a version 1 that crosses off all the requirements that I set forth in on Day 0. Right now it's too crude to commit to my repo. I'll keep working this afternoon after I get some work meetings out of the way. 

Current Trajectory: Day 0

I'll share a detailed breakdown of the overall vision soon, but right now my priority is successfully navigating day one of restarting this project. At this point I've put in more than enough... 6 months of prep work, and I want to ensure I'm not getting lost in the big picture. Today we start. The goal is building from a strong foundation by tackling small, manageable pieces first.

Current Objectives:

• Unified Entry Point: A single, streamlined file to effortlessly launch and halt the entire system.

• Dynamic Timing/Metronome: A centralized process managing timing, where system state updates occur based on defined cycles.

• Persistent State: Maintain stability with at least one variable securely saved to survive unexpected shutdowns.

• Function Registry: Automatically populate eligible actions by pulling from a dedicated 'functions' directory.

• Multithreaded Action Queue: Enable concurrent execution of queued actions, independent of system timing cycles.

• Example/Test Functionality: Create a simple, verifiable function to ensure foundational components are working correctly.

Over Thinking

 As I'm getting prepared to start drafting out my plan ~ I can't help but watch as my mind runs away from me. Thinking about far off edge cases and how I might handle them. I keep comparing myself to enterprise solutions or worrying that what I'm working on will be irrelevant in a short time or even before I finish. 

I need to ignore the noise. Stop listening to the 'AI' tech news. It's time to put my head down and get something basic launched. I have the tools and the path to any information that I could need... Nothing's stopping me but me at this point. 

It's time for me to get out of my head and out of my way. One step at a time. 

Decision: Language - Python

 I'm going to write this as a python framework. For the most part I do my programming in Python. I like that if I need to I could push down into C++ or, god forbid, C. 

However, there are many other reasons Python seems like the path forward. 

1. Most of the AI/ML programing the cool kids are doing is happening in Python Notebooks ~ so it should be an easy fit for those developers. 

2. It's super easy to read and write in Python

3. Python continues to be well supported; however, with how popular it is, if the general community were to leave python someone will make a tool to port code to the new cool language.