Update README.md

README.md

@@ -35,18 +35,21 @@ and by the super-resolution registry code from the [Spandrel](https://github.com
 - Organized hyperparameter management
 
 
-> <details> <summary>Why not `diffusion`, `transformer`, `sgm`, `ldm`, `text`, hf.co folder-structure, (brand-specific trade word)</summary>
+> <details> <summary>Why not use `diffusion`/`sgm`, `ldm`/`text`/hf.co folder-structure/brand-specific trade word/preprint paper/development house/algorithm</summary>
 >      
 > - Exact frameworks (SGM/LDM/RectifiedFlow) includes too few
 > - Diffusion/Transformer are too broad, share and overlap resources
 > - Multimodal models complicate content terms (Text/Image/Vision/etc)
-> - HF.CO names do all of this, inconsistent across folders/files
+> - HF.CO names do all of this & become inconsistent across folders/files
+> - Development credit often shared (ex RunwayML with Stable Diffusion)
+> - Paper heredity would be a neat tree, but it complicates retrieval
+> - Algorithms (esp application) are less common knowledge, vague, ~~and I'm too smooth-brain.~~
 > - Impartiality
 > </details>
 
 > <details><summary>Why `unet`, `dit`, `lora` over alternatives</summary>
 >    
-> - UNET/DiT are shared enough to be a genre but not too narrowly specific
+> - UNET/DiT/Transformer are shared enough to be genre-ish but not too narrowly specific
 > - Very similar technical process on this level
 > - Functional and efficient for random lookups
 > </details>