You can hydrolyse glucose in one step with the correct protein and release a lot of energy. You could actually be far more efficient doing this than breaking down the steps into chunks.
It's true that ATP >< ADP is an easily reversible reaction with a discrete, small amount of energy that should be manageable, but why ATP is used is almost an impossible question to answer; it's like asking why the exact structure of DNA is exactly so; and not subtly different. We can examine the structure of these molecules and rationalise why their structure is good and helps their functionality, but knowing exactly why life evolved to use these molecules and not others is a question you could spend your whole life thinking about. Attaining that level of knowledge and understanding of biochemistry would mean being knowledgable enough to design molecules that on a molecular level can be better than proteins, DNA and RNA for performing tasks. Humans are at the very least centuries away from such a feat.
An important point to mention in ATP is that the glucose-ATP cycle is very important outside of glucose and ATP also. A lot of the intermediary products are used in various other biochemical cycles and can be useful to the body in different ways. Contrast this with a design where you just take glucose and immediately smash it into energy (which is really only good for one thing - energy production). The system has evolved around this state and helps carefully monitor a number of things in the cell.
It's hard to appreciate at A-Level where you are just taught cycles independently of one another, but really biochemistry in cells is a massive mish-mash of various cycles that all feed into each other and depend on each other to function properly.
Re; 20 amino acids.
This is a really great question because it's easy to get a handle on why degeneracy is important in the context of how frequently DNA gets damaged or mis-replicates. By looking at the patterns, a smart A-Level student could probably guess what the most common mutations or mis-replication events are in the cell. I could easily see an Oxbridge interview question stemming from this idea.
Chemically speaking, 20 amino acids might be a bit of a misnomer anyway. It is debatable whether or not 20 is the true number, when most animals cannot even produce tryptophan on their own. Likewise, it is very common to chemically alter amino acid groups in the final protein design.
If we take a reductionist view and think "if a chemical group exists on R, and this chemical group cannot be massive and needs to serve a reasonably different function from the other R groups (amino acids), how many simple examples of this are there for a hydrocarbon based (but with some source of N, P, S) lifeforms?" The 20 amino acids actually cover a very good range of this multidimensional issue (hydrophobic vs hydrophilic, large vs small, flexibility, length, oxidising or reducing power).
Of course, more examples are possible, but we need to remember that cells exist in an aqueous environment and there is a certain limitation to how "dangerous" you can let your basic building blocks be to avoid poisoning yourself. It took a very, very, very long time for life to even start using Oxygen, because of how destructively powerful Oxygen chemistry can be (it is obviously a powerful oxidising agent but also mismanagement of its structure can cause free radicals).