I would never advise such a thing. Jump ship as soon as you can!
I recommended a virtual machine because the cost of screwing up is negligible. You could wipe your hard drive and just run Ubuntu, but if you screw something up badly enough that you need to reinstall, you’re out of a working computer for that time. Plus, getting people comfortable with Linux is a lot like getting them to exercise or eat healthy - start slow, keep the expectations low, and integrate the new with the life they already have. Dramatic changes are usually a shortcut to failure. I’ve seen it a hundred times before, so my suggestions tend to be conservative.
The goal underlying my suggestion is for you to feel like you can get things done on Ubuntu rather than diving into a bunch of arcane commands. If you think you’re ready for the DevOps stuff, then by all means spin up an Apache or NginX server and start routing requests! Get a firewall going, create an SSL cert, or install Wordpress. By “Take things slow…”, I was thinking you’d spend a day or two just playing with all that Ubuntu has to offer and then you’d naturally progress to the more complicated stuff soon after. I’m not in charge, though. You can do what you want!
Ok, so if you’re gearing up for the backend stuff and you want to start on the right foot, @belcurv’s guide will get you going with SSH. A common task with remote servers is accessing them via the command line using SSH, then editing files. You can do the same with a VM, just minimize your VM window and connect from your Windows host. Once you can do that, get comfortable with the command line text editors
nano. Navigate directories and learn how permissions work in the Linux file system.
rscync are hugely useful tools. I would say these skills make up 90% of the DevOps stuff that I have to do for my personal servers, and I consider them the foundation of everything else I could ever do.
I’m definitely not the final word here. Just my suggestions.