Depth-aware light injection in TypeGPU
I got a 448x448 monocular depth model down to ~8 ms on my M4 Pro across ~250 dispatches, which is fast enough to use in realtime :D
Since the inference is written directly in TypeGPU, I can just feed the depth buffer straight into the lighting pass. It never has to leave the GPU or go through any extra synchronization/interop step
Inference, lighting and draw all go through the same command encoder.
@terraindustries raises additional $18M bringing our total seed to $52M. This round was done by existing investors @8vc, @SilentVentures, @svangel, Nova Global, Belief Capital, and Grant Gordon
We're also opening our first global office in London for our Ops and AI teams (1/2)
product interview sessions when customers are telling you all things nice *maybe because you’re compensating them for their time, so you have to keep probing past the politeness to know their real pain points
I’ve been thinking a lot about this:
thinking about your own thinking ~ an important metacognitive skill.
Cos you really do have to examine how & why you act the way you do to recognize the patterns shaping your actions & create progressive shifts in your life
I now value my long walks more than ever, especially for brainstorming.
Now that execution or sharing early prototypes with my team is faster; thinking deeply is the moat in identifying the right gaps to solve for