Lmao no they didn't, they bet the farm on cuda for scientific work, which goes hand in hand with consumer graphics market (essentially the same hardware). They

Lmao no they didn't, they bet the farm on cuda for scientific work, which goes hand in hand with consumer graphics market (essentially the same hardware). They've just pivoted to AI semi recently (considering their long history) using the same hardware as always, for graphics/scientific/AI workloads. Sure they tweak vram amount/bandwidth/support additional precisions but it's all really same old same old. Just with more transistors every 2 years.