I wish this didn't force us to use pre-existing lego pieces by abstracting away the math. So so so many cool things could be built in the computer vision sphere

I wish this didn't force us to use pre-existing lego pieces by abstracting away the math. So so so many cool things could be built in the computer vision sphere if engineers weren't afraid to go a level deeper, ignoring all the opencv stuff, and just implemented the math themselves. There's no reason we can't get creative on a level further down. With all the clever math people that are enthusiastic about abstract algebra in security, you'd think there'd be some spillover of that math confidence into the CV world.