Starting grad school is like being dropped into a jungle with a machete and being told "find something new". Maybe (video game-like) you have a supportive supervisor, and so you are given a crude map. If you have labmates or fellow students, you can fend off the predators together. A good funding source allows you to travel faster. Best of all, maybe you come across some Tilley-hatted explorer who is so excited about the jungle that they give you detailed directions. There are infinite paths through the jungle, but some are harder than others.
The other analogy was for how to be a good supervisor, which is like a parent teaching a child to ride a bike. The parent can push a child off, and say "peddle!". This will be followed by lots of crashes and scraped knees and maybe the odd close call with traffic. Maybe the child figures it out, and is a fearless cyclist. But they might give up on biking all together, too. Or, the parent can hold the handles the whole time and say "great work! you're riding a bike all by yourself!" The result is a confident little cyclist who will probably crash when they finally get the opportunity to ride without help. A good supervisor probably holds on at first, then graduates you to training wheels and then takes them off. There will still be a few crashes, but the result is a cyclist not afraid to go alone, and without too many cuts and bruises.
Or maybe you've heard better ones?