A Pledge About Robot Words, and My Next Chapter
I am making a pledge right now. I vow to be transparent about the use of AI in all of the work that I do. I really don’t care if it can produce fast results with low effort, I’m not reading that crap if you post it. No one wants to read robot words.
I have implemented an #Authorship badge system on the site. Every post gets an authorship badge, either Human Only (written 100% by me), or AI Assisted. For AI-assisted posts, the scope of the assistance is limited to: outline creation, rewording some sentences, and light editing. I never have and never will publish AI slop.
On a related note, this site is changing. So far, It has been a “proof of work” blog, documenting my 6-month journey of gaining deep AI tools knowledge and building processes, workflows, tools, and products. During this time, I iterated and developed my “specbuilt” approach, while using it to build tools and products, many of which I use daily in my own workflow.
Now that my specbuilt approach has matured, it is ready to be applied to products I put out into the world, taking on client projects, and helping guide clients through using this new technology well.
To close, here is a quick list of the projects that I completed before starting this blog, closing the “proof of work” circle that led to my specbuilt process:
| Date | Project Name | Description | Techs Used |
|---|---|---|---|
| 08.05.25 | wpfxaml-lesson | Reference implementation and UI guide for WPF/XAML patterns and common practices | C#/.NET, WPF XAML |
| 08.07.25 | mediabeast-auth-github | GitHub OAuth authentication implementation for MediaBeast | Next.js, GitHub OAuth |
| 08.26.25 | image-cleaner | Image processing / cleanup utility | Python |
| 08.28.25 | spinebeast | Proof of concept for identifying movies from DVD/Blu-ray shelf photos using spine detection and OCR | Python, FastAPI, OpenCV, Tesseract OCR, React |
| 08.29.25 | mediatreats | Proof of concept phone/web app for scanning UPCs and sourcing shelves of physical media | Node.js/Express, PostgreSQL, HTML/CSS/JavaScript |
| 09.02.25 | movie-hover-letterboxd | Chrome extension showing movie release years via tooltip with right click, then pull up Letterboxd review | JavaScript, OMDB API |
| 09.08.25 | zustand-demo | Research/educational React project demonstrating Zustand state management with modern patterns | React, Zustand, TypeScript, Tailwind CSS |
| 09.15.25 | photo-uploader | Windows system tray app that auto-monitors a folder and uploads photos to a Supabase database | Electron, TypeScript, Supabase |
| 9.15 - 11.15 | eBay listing tools | Worked on eBay listing tools and used them daily. Took a full month to focus on running $10k/month MediaBeast eBay physical media store business. Punted when confirmed inability to scale due to sourcing constraints. | — |
| 10.20.25 | upc-scan-listener | GUI app for listening to UPC barcode scanner input with fuzzy matching and rapid lookup | Python, PyQt5, pynput |
| 11.13.25 | bat-to-exe | Convert batch files to silent standalone executables | Python |
| 11.18.25 | p2p-movie-collection-tracker | Movie tracker using PouchDB for offline-first peer-to-peer data sync | SvelteKit, PouchDB, TypeScript, Vite |
| 11.19.25 | shelfgrade-phone | Flutter mobile app for ShelfGrade collection tracker | Flutter |
| 11.21.25 | shelfgrade-component-playground | Testing playground for “Discogs for movies” — ScrapingDog and UPC APIs | — |
| 12.07.25 | scraper-testing | Short-lived experiments with scraping various movie info, but none viable. Scraping is shady — wanted no part of it. | Python |
| 12.11.25 | titledate-to-lbx | Tkinter desktop app integrating TMDB API lookups with Letterboxd — duplicate detection, backups, database management | Python, Tkinter, TMDB API, SQLite |
| 12.15.25 | weektodo | Take existing MIT-licensed open source project and modify it to add multiple user profiles | Vue 3, Electron 25, Bootstrap 5, Vuex, SASS, Docker |
…and that’s about it.