I think from a higher level there are some interesting things, you're right that I should have gone into it a bit. One of those is that raising my ceiling of play, like challenging myself with super tough puzzles, was actually totally unproductive. Raising the floor of my play, however, was hugely beneficial. I imagine this generalizes somewhat to other pursuits.
The other takeaway I've had is how different procedural learning is from declarative learning. Among one of the weird features of it is I don't actually know how much something is helping until I get to the board and play some games afterward. There were exercises I thought were helping a lot in the moment, but had no effect on my play, and vice versa.
Also sometimes people that are good at something are the worst to ask for help. For example, if you ask "how can I stop blundering?" (one of the most common questions on forums), common advice (from good players), is to have a mental checklist before making a move: threats, captures, skewers, pins, etc. But nobody actually plays like this, running through some checklist before every move. You just slowly rewire your brain over 100s/1000s of games and puzzles. If my queen and king are lined up now, and there's a rook on the board, that's as obvious as a flashing light on the board saying "hey watch out for pins!", but there's no easy answer to tell a beginner that will get him to that state, so people try to convert that feeling they get into a manual approach, and it just doesn't work. I guess this is like the key feature of procedural learning, that it's resistant to verbal explanation, but that doesn't stop people from trying, so you just need to learn to take it with a huge grain of salt.