It is sometimes easy to explain why it doesn't work as expected if you just check which Blender version is used in a tutorial. The behaviour in the tutorial is from a much older version and only appearing if you used Connected Only (which you do not either in your screenshot) - but it was not intended behaviour. It was treated as a bug because it could lead to distortions and was removed, read more about it here: Proportional Editing: 'Connected Only' option can cause distortions depending on topology

If you mean the Adjust Last Operation panel where you can do things like change the number of vertices on a cylinder or the resolution of a UV sphere for example, I don't know if that was F6 back then but it's F9 for a long time now, maybe try this. Apart from that, it is not possible to bring up the settings anymore if you did any other action after adding the object. It only shows the parameters of whatever was the last action.

Of course only a single frame being rendered over night is probably not (solely) dependent on the animation itself. I had several times now where 4.2 (you said you updated) got stuck even in simple scenes, so maybe the current stable release is not yet so very stable after all.

It also depends on what you're rendering. When I render a smoke simulation and the smoke builds from frame to frame, there are more and more volumetrics to render which makes each frame slower. Or let's say you're approaching a planet displaced with Adaptive Subdivision. Sure, far away things have less resolution, but the closer you get the more the planet might fill the frame, i.e. more displacement has to be calculated. I don't say your animation is anything like this, but we know nothing about it and I just wanted to show there can be several single or combined reasons then.

And here specifically explaining among other things why your new material is not automatically showing on the object: blender.stackexchange.com/questions/292434/…

As the author of the answer you have linked to I have to ask for more information. Theoretically it works like I explained there, so when you say it is still not working correctly it would be helpful to know in which way it is not correct or we need to know your exact settings to find where they might differ to cause the issue. But while I'm typing this comment I see @JohnEason has provided a tutorial link. Maybe this already clears things up.

If you do not want to rotate the model, there is nothing you can do. The world orientation cannot be changed. And it would not make much sense either. A scene can have multiple objects, not just a single one. Everything based on world coordinates (not just objects, modifiers, constraints etc. as well) would change just because you wanted to avoid rotating one object.

@Lawton Did you see my edit? Looks like I did not save it earlier...

@Lawton I've added a paragraph at the end of my answer. With your normal map now I can see a difference compared to my normal map. But when you switch the Normal Map node from Tangent Space to World Space, this should be less noticeable since the refraction is not different anymore, for slight discrepancies see the edit in my answer.

Without your file it is hard to tell what is going on there, because "by default" a light source in 4.2's Eevee doesn't do this (at least not on my PC).

No, you cannot edit procedural not "really" existing faces manually. You can change modified faces by other modifiers, since they all create procedural effects... for example first increase the resolution of an object with a Subdivision Surface modifier and then deform this highres mesh with a Displace modifier. But manual editing is not possible. You could only keep a duplicate with unapplied modifiers.

For the choppy and unrealistic smoke when using force fields: hard to tell, this can depend on the domain settings, the flow object's settings (how many Sampling Substeps have you set for example), the force field settings...

You can do one of the following things: either as @Jakemoyo said, set the Cache > Frame Start value in the smoke domain to a negative value. How much is dependent on your scene, the resolution, the flow settings, probably Dissolve settings etc. or the other ethod would be, start simulation at frame 1, but after you have baked the cache, use a negative Offset value to skip the first part of the siulation. In both cases you will to simulate more than you will be able to use afterwards of course since it will be cropped.

Please do not write "SOLVED" in the title, accept the answer instead, that's what this function is for so people can filter the forum for unanswered questions.

I do not know what kind of 4.0 you have there, but I tested it on my PC with Blender 3.6.9, 4.0.2, 4.1.1 and 4.2, the behaviour is consistent in all versions. As soon as I plug the Shader to RGB node inbetween, the reflection disappears. The only reflections the Shader to RGB node shows is the environment (grey in your case or an HDRI in my case). And this still works in 4.2, so I cannot see any change.