Can current AI produce a mathematically constrained knit design that the image actually matches?
Mysterious_Green_544 · reddit · 2026-07-21
The author is trying to use AI to adapt a knitting pattern to leftover yarn, with mathematically valid stripe sizing and visuals that match the calculation.
What keeps going wrong
- The model often forgets the constraint that the design must be mathematically feasible.
- Image generators produce sweater visuals that look striped in a vague sense, but ignore the exact stripe widths and layout.
- Failures include stripes changing width, fading, stopping midway, or differing between front and back.
What they want
They are asking whether the right workflow is to:
- generate SVG/vector drawings first,
- use CAD or knitting-specific software,
- split the task across multiple AI tools,
- or switch to a different model entirely.
The core question is whether current image-generation models can produce an accurate, constrained design where the rendering really reflects the computed solution.
More from Apps
- Proof-of-concept flags 1,800-review Etsy shop reselling $6 Alibaba necklaces for $75 — HankYeomans · 2026-07-21
- ARISE study tested 45 AI clinical tools in 1,100 consult cases — HealthcareAIGuy · 2026-07-21
- Google Quietly Launches Gemini 3.6 Flash: Cheaper, Stronger, and Agentic-Focused — OwariDa · 2026-07-21
- Whop launches a CLI so businesses can be run programmatically from the terminal — eptwts · 2026-07-21
- Business chatbots should add RAG only when answers depend on company-specific documents — recro69 · 2026-07-21
- HandoffAI says its H1 model can automate construction takeoffs like a human estimator — ycombinator · 2026-07-21