These are the raw, unedited notes I wrote while shaping this article. They are closer to a brain dump than a prompt library: my thinking, philosophy, corrections, and direction. The ideas are mine. AI helped turn them into a smoother, more readable essay.

Prompt 1
In the ghx project when was the last time we wrote an article
Prompt 2
Read the article and understand what work has been accomplished since then what ambiguity are we sitting in right now. What new vision are we moving towards to, what calculated bets are we making with ghx and SAF? I want to write a new engineering article in the ghx series
Prompt 3
Where is the pain lessons learned so far and hard challenges that we are banging our heads on? Why dont we have adoption yet still? Like i myself dont use the sidecar ghx yet its slow and not fully proven, and still havent proven why this is much better than sinply triggering a subagent with claude or smth like that
Prompt 4
Read the skills inside the gkoreli blog repo for writing engineering articles and writing style and start iterating on the article in the ghx series
Prompt 5
We have been building ghx heavily recently do we even have logs or analytics or soft failures or hard failures logged? Do we know how we are using ghx besides running the evals? Like are we even using it for building the ghx product itself? If not why not? I feel like this needs to be properly investigated and if the ghx product injury exists and we are bleeding out of it just call it out in the article, lets be fully transparent like over these days how much did we reach for it… opensource is full of great ideas and references why are we not looking into it whenever we are building architecture or need to look up some docs or reference implementatuons or looking for inspiration or conpetung ideas?
Prompt 6
Im using Hermes for engineering last couple days mostly i use my personal laptop and work laptops for engineering and i use ghx more or less heavily but i barely use the ghx sidecar tho, i was just curious like yesterday and today we spent a lot of effort in engineering like heavy efforts and curious if we reached for ghx, or do we even have any of this information available
Prompt 7
Cross validate more ground your understanding with work, with vision tenets, problems and cross reference open source but dont trust blindly ground yourself with understanding cross references