Every token vendor seems to be having issues, by proxy of hyperscaler capacity issues. Whether or not we get served up quants is opaque, and should be a finable offense for not delivering what you are charging for.
The dual spark is definitely nice when this happens, and even though I doubt I'll ever pay it back in terms of saved token costs, the freedom sure does feel great!
I suppose this is as good a plan as any.
I had to migrate away from Github Copilot in June, and I am not looking forward to doing the whole moving thing again.