Nate Jones released a video that I have been thinking about. It has a great insight and wanted to bring to the group. https://www.youtube.com/watch?v=UsCgEuIAclE Many of us are improving our AI skills here, learning and building things. As we build things for resumes or our own business, maybe we should think about the White boarding at the beginning of a project. I am good about scope for projects at work but not great at it for personal projects. The end goal changes and scope creep enters and the original intent is lost in the woods. We have smart people here with variety of experiences. We could use each other for such things. I have suggested on the discord, (In the premium chat) that we could meet there and have open discussions. show and tells, etc.. But a white board discussion on personal projects could be even more useful. For Example: I currently have a video game I am working on: I am close to the point of making public a GitHub repo for Porfolio purposes. No graphics, just a dashboard basic game mechanics and the Unique part. You can hold audiences with different factions, those are run by LLM and decisions are made in that chat and remembered for the next audience. I would love to have a discussion on the implementation, Graphics, How complex/simple should the system be? Where do I stop for the Portfolio presentation version? Possibly recording the session for future use. It would be good practice either way. Anyway, that was on my mind this morning and wanted to share. Get other people's thoughts. below will be a paragraph on the game. Polis — Rule the Unrulable A simulation of political conflict inside an ancient-Greek city-state — where rival factions scheme on their own, and you bargain with their leaders through a live LLM. You play the Prytanis — the city's presiding official (the Mayor role in the engine). You can't command the factions; you work them — endorse and condemn, levy taxes, broker deals, spend coin and favor — and you can hold live, LLM-driven negotiation audiences with any faction leader, who bargains in character. Agreed terms are parsed into structured deals that feed back into the simulation, and the faction remembers how the conversation went.

First, congrats to James — well-earned. 🏆 This isn't about placement, and it's nothing against anyone who placed. I just want to get better and understand the bar I'm building against. I shipped Achiote for Week 6 — a food-memory researcher for immigrant families. Give it a sound-alike, a smell, a color of a dish nobody wrote down, and instead of guessing a recipe it investigates: treats the memory as evidence, weighs sources, keeps the unknowns visible, and ends with one small, cheap taste to test the memory before any full recipe. - 📁 Folder: github.com/simongonzalezdc/achiote-food-memory-researcher - 🌐 Live, open demo: achiote.kyanitelabs.tech Two honest asks: 1. @Jake Van Clief — a couple weeks back you mentioned building a scoring rubric and sharing it so we could see exactly what you weigh. Is that something you can post? It'd help all of us calibrate, not just me. 2.If you got a chance to look at Achiote — where does it fall short? What would you change about the folder or the framing to make it land as a researcher? Straight answers welcome. That's it. Thanks for running these great learning experience.