There is no substitute for solving problems.
Some anecdotal evidence for this:
https://www.dwarkeshpatel.com/p/dario-amodei#details
> Dario Amodei: We have generally found that if we hire someone who is a Physics PhD or something, that they can learn ML and contribute just very quickly in most cases.
In math/physics, it often won't. Solving lots of problems serves two particular purposes: To really solidify the concepts in your mind so you won't forget, and ensuring you learn the techniques and not just the knowledge.
For the former, you may find yourself in the position where you find yourself way over your head, and won't know where to start. You usually will not have a single gap, but many. You'll find yourself realizing you'll need to look up material from several textbooks to regain the knowledge you've lost. Once you begin that process, you'll pick up one of your old textbooks and while the physics knowledge may be absorbed, you'll realize you've forgotten much of the math needed to solve such problems. In the unlikely event you'll retain enough to follow the textbook, it is very unlikely you'll know the techniques well enough to solve the real world problem.
And your colleagues will. You'll be alone, and you'll drop out of that group. With physics/math, there often are hard boundaries in these groups. Those who meet the bar are in. Those who don't drop out, because it really sucks being the only person in the group who is struggling with what everyone else considers as basic.
SW engineering has a much more gradual change in skills amongst people, and usually the problems most people work on are fairly learnable in a short amount of time.
Math and its applications are a contact sport. You don’t truly appreciate it until you try to use it yourself.
I'm going to ruminate on this.
But you will spend the rest of your life arguing with people who insist that there must be a ("quick and dirty") substitute and you're responsible for finding it.