pre

no no you did that one to yourself. I'm just doing my job

That's two words. Accounting is gonna be pissed.

Actually not Fibonacci but it woulda been cool if it was necessary, just increasing po2 I think will reducing patterning.

If multisampling isn't allowed for some reason, then it might be interesting to try an increasing Fibonacci? I haven't thought this through much but, starting small localizes against patterns, but the min width of the max value will grow such that I believe it must always converge.

I feel like a 1+2 distance multisample combines the approaches to reach a middle ground. You'd converge twice as fast, but avoid patterning with the near sample.

This feels like a gaussian blur or similar GPU operation, so there might be something inspiring there too.

Xbox, eat berries, he's a fair enemy

Will a heatsink and fan be enough?

hey siri: what if this ware is firm for more than 4 hours?