I have actually been really impressed. It's fast and responsive, even on Linux, and the TypeScript tooling is fantastic. I think with a bit more UI work (like vi keybindings) it could potentially become my preferred editor.
Having said all that, many of the unimplemented parts in extensions across the board come down to limitations in the VSCode extension APIs. I've been watching the VSCode team, and they're actually pretty sharp and welcoming, too, so most issues should be ironed out, given enough time. (Although it hasn't been prioritized at this point.)
However, the appropriate unit of measurement to describe how long I've been using Vim is "decades", and through observing (but never seriously using) the attempts to emulate it over the years, I've spent a lot of time thinking about what it would actually take for an app to really support Vim keybindings—aside from actually just being Vim. Unfortunately, even if we assume a team with infinite resources intent on achieving perfection, there are several things about Vim that prove to be irreconcilable with the way almost every interested app is implemented. The only way around this will have to come down to Vim users either collectively participating in some "Great Vim Shift", or just ignoring the remaining gaps entirely and continuing to put up with incomplete emulations.