Their side
When Windows Speech Recognition still makes sense
There is a narrow case and it is worth stating honestly. If you are on Windows 10, need voice control of the interface rather than just dictation, cannot install software, and cannot use anything that needs a connection, WSR is what is available. It is free, it is already there, and it runs locally. Being old does not mean being useless.
It also does something we do not: it drives the interface. WSR can open menus, switch windows and click by voice. We only turn speech into text. If hands-free control of the computer is the requirement, WSR is closer to the right shape than we are, even in its current state.
And the voice-training step, dated as it feels, genuinely helps on distinctive accents. Modern tools mostly abandoned per-user adaptation, and for some people that was a loss.