6.4K
BAIf you can build a coding project like this, you’re not competing for jobs… you’re competing for leverage.
Right now, AI can handle basic spatial tasks. Depth estimation. Simple relationships. Object detection.
But it cannot robustly reason about spatial manipulation, 2D and 3D structure, or mental rotation.
That gap is massive.
AI still struggles to truly understand and interact with the physical world. Which means whoever helps close that gap becomes extremely valuable.
If you can:
• Build even a small proof of concept for spatial reasoning
• Model geometry as first-class primitives
• Show real-world object reasoning or environment design
You won’t look like a junior engineer.
You’ll look like someone thinking at a foundation-model level.
This isn’t a “CRUD app” project.
This is a “define the next AI category” type project.
It’s not something one developer fully solves alone.
But even an MVP that pushes this direction will separate you from 99% of candidates.
I put together the exact resources I’d study to attempt this.
Comment “SPATIAL” and I’ll send it to you free.
@bashi_fuirkashi
