middle game, possibly the opening problems


Fundamentals come first: develop your pieces and get your king to safety. You are not going to win if you have no pieces on the board to attack with, and neither if your king is vulnerable in the open for your opponent to attack. Are you trading just to trade, or does it make your position better or your opponents position worse? How can your opponent punish your (short and sweet) plan? Are you leaving behind a weakness?
Play slower time formats while you are learning, rapid would be far more instructive than blitz.
Pieces out, king safe, analyze games and understand why the move was a mistake, play a lot of games vs. humans. Chess.com has a lessons feature I think, and there are also lots of free resources for beginners on youtube.

well, if you did not make mistakes......you would be a grandmaster. So...you are going to make mistakes. Every move you make might be a mistake in the computers opinion. Try playinga top rated engine and see how long it takes to thrash you. People make more realistic errors than a computer bot that's programmed to make errors. I find the computer doesn't like sacrifices, calls them blunders, even when it leads to a win. Sacrificing pieces you won't need to open up attack lines works fine.....at lower chess levels.
My experience of playing the computer was that it wasn't very realistic. For example it would play a good opening but make terrible blunders in the middlegame. People are more consistent, they dont play a 1400 opening and 400 middlegame. It was easy to win, I would offer a big sacrifice, the computer would ignore the free material and leave me with a great position. This was a year ago on a 1200 rating. It may have improved.

As a fellow beginner, I find the game report to be overwhelming at times. When I first joined chess.com, I basically only reviewed my blunders. Maybe also 1 or 2 mistakes, but inaccuracies were definitely too nuanced for my limited knowledge level. With more experience, however, I began gradually diving deeper into the report. Regarding your specific question, it seems like you may be developing plans that are easy to defend and counter. You may not have noticed the flaw while playing against another beginner, but a better player would have made you pay and the report will grade you as if your opponent is a GM. Having said all that, I know how it feels to be proud of a tough win only to find out that I actually played like trash and basically just got lucky my opponent was worse!

Because humans make natural errors, just the way you also would. Computers on the other hand blunder on purpose, which makes for an unnatural experience in contrast to how actual human vs human games would play out. When analysing a game vs. a human opponent, you can learn from your own mistakes but also from theirs. This both during and after the game.
The engine will play decent chess and then just give away pieces knowingly, for no reason. Now compare this to humans, which make oversights on accident. Even if the mistake was made because of a faulty plan or idea, looking at that plan can help you understand WHY the error came to be, and thus improve your own chess.

Regarding the computer vs human opponent question, their styles of play are so different that if your goal is to win against human players in the long run, why not practice against them now? (As others have said, the computer seems to mix credible play with random nutty blunders, and playing very safe while watching for and capitalizing on blunders is not always as productive against human players.)