I have made an OS in CUDA

i have made an os in cuda The whole stack, network (tcp/ip), graphics, UI, agents, storage is in CUDA (kernel in raw PTX). The entire inference and agent stack is also self contained within video memory (& GPU storage), and every single thing by default supports massive parallelisation by default. There is a shared vectorisation (embedding) pipeline in synchronity with the filesystem I have built to be native to the GPU. The tokeniser, and all other related elements necessary to run local inference are also written in CUDA. The interaction between you, and this, goes through an interpretation layer that the system itself is non-dependent on. Containerisation is also done, inside of this system, via the same methods It is called AOTX-1 (Ahead of Time eXecutive One) and is by design ahead of any user display, perception or hardware, the process of transiting data to your actual PC is via a labouring process that writes it using the (now unused) CPU. You are finally seeing this, because it works. don't want to show and tell yet, but here's my debug panel, this is just kernel information and ticks (the sound is telling)