The first is hearing: telling that the vocal has 400 Hz too many, noticing the compressor is pumping. The second is deciding what to do about what you heard. Tutorials teach tools, which is a third and separate thing — which is how you can accumulate years of videos without resolving a simple mix.
This is the front door: 3 introductory scenarios that show how the format works — a mixing problem is posed, you choose what to do, and the explanation comes after — and then hand you the rest of the platform. It takes a couple of minutes and there's nothing to install.
The ear is the slow part and the one you can't skip; judgement is the other half, and it's trained by deciding. Exercises for both are spread across the platform, and each links to the guide explaining its reasoning — because practising without understanding what you're practising turns getting it right into luck.
Neither. The lab and much of the site work without signing up; an account only exists to carry your progress across devices.
Not to start. Everything runs in the browser and decent headphones are enough. What does help is having a project of your own to apply things to.
If you diagnose well and what fails is ear precision, skip this and go to EQ Surgeon or into Mission Mode at whichever world fits. The lab is the door, not a requirement.
A couple of minutes: it's 3 scenarios and they exist to teach the format, not to assess you. What comes after is what takes time.
If you don't know where to start, the mix skills test tells you your weakest area in eight questions. The ordered path for the ear is in the ear training guide. The first concept worth having straight before touching anything is the level you work at, and that's in the gain staging guide, and the full guide index is at Guides.