These revisions are definitely considerably better than those you'd submitted before, and Tofu will be along in a bit to mark this challenge as complete.

Since you were having trouble understanding why and how they're better, I did want to take a moment to explain that myself. As I've shown here, the boxes from your initial set have some key issues - two to be specific, though they're related to one another.

The issue comes down primarily to the fact that you were trying to keep the edges of your boxes parallel when drawn on the page, effectively placing the vanishing points for those lines at infinity (as discussed in Lesson 1). The second, related issue is that in attempting to keep them parallel like this, you would fairly often end up with lines that diverge slightly as they move away from the viewer.

The divergence is less of a concern - rather, it's the intent to keep those lines parallel to one another which is incorrect. A vanishing point is not placed anywhere - rather, its position is determined by the orientation of the 3D form it's meant to help represent. We do not arbitrarily choose to put our vanishing points at infinity, but there is a specific circumstance in which a vanishing point will go to infinity - that is when the set of edges it governs run perpendicular to the viewer's angle of sight. Basically, when those edges run right across their field of view, not slanting towards or away from the viewer through the depth of the scene.

As soon as the lines angle themselves even slightly such that they start coming closer or moving farther away from the viewer through the depth of that scene, that vanishing point must become concrete, and as the given edges rotate more to come towards the viewer (rather than moving across their field of view), the vanishing point's proximity will increase and the rate of convergence of its lines will also become more dramatic.

All that said, the box challenge is all about rotating boxes freely and randomly in 3D space, making it quite unlikely that we'd ever run into a situation where a box would align so perfectly along any of its axes, that one of its vanishing points would go to infinity. If we did end up in such a perfect situation, however, we could only ever have a box align two of its sets of edges perpendicularly to the viewer's angle of sight. That third set of edges will always end up coming straight at the viewer.

Now, the reason your newer boxes are vastly better is that they don't make this mistake. You've got more conscious, purposeful convergences to all your sets of lines. You may feel that they're worse, simply because this forces you to contend with many, many cases where those lines in fact do not converge consistently - but that's what this exercise is all about. Having students try to achieve more consistent convergences, and then analyze their results to see where those convergences were off.

You can - and should - play with convergences that are both dramatic and shallow when practicing this exercise in the future, of course. Just make sure that even in the cases where the convergences are very gradual, that they are intentionally present.

Anyway, as I mentioned before, Tofu will be around later to mark this challenge as complete. You will of course be incorporating this exercise into part of your regular warmup routine, as with all the exercises you've encountered thus far - so yes, it is very much a matter of mileage, but now that I can confirm you're headed in the right direction you will be in a better position to improve on your own.

Edit: Oh, in regards to the rushing thing - don't set deadlines or expectations for yourself as you move through the course. Students will sometimes feel it necessary to assign themselves an expected point at which an exercise or lesson should be completed, but this can only encourage us to rush, be it in the execution of our marks, or in how thoroughly we go through the instructions.

It will always come back to simply giving each task as much time as it requires to be done to the best of your current ability. Each lesson is not the same - some are more demanding and more complex than others, and therefore they will ask more of you. All you can do is commit to sitting down to work on the material regularly, and to be willing to break up individual drawings across multiple sittings as needed (this will be more of a concern as you move forwards).

The 2 week cooldown is not a deadline. It is quite the opposite - there to force students to slow down, and not to try and complete the work as quickly as they can. So, in that sense, pushing more towards a month is a good call, but still a little misguided since you're still laying down some sort of expectation to meet, rather than allowing the work to take as long as it needs.