My best guess as to how to move forward: pick a high-level language --- Ruby, Python, or Javascript (Node), but not Rust, Go, Clojure, or Haskell. Find a straightforward project that interests you and learn how to build it. If your project takes the form of a Rails or Django application, you'll have an especially easy time self-teaching yourself.
I would earnestly try to get this done before considering any kind of boot camp or paid class.
People will be quick to point out that you're only N% of the way to being a software developer by getting to the point where you can bang out a Django or Flask application, and they're right, but you're 80% further down the path of learning how to learn software development, which is the important thing to optimize for. The trick after getting that far is to keep ratcheting up your ambitions, so that the next project you do is harder and a better learning experience than the last.
There are some edx online courses for free and plenty of programming books to get you a little experience to see if you even like it. Bootcamps might be able to get you a job and they actually focus on more of the practical building of software, but a computer science degree still looks better to many employers from what I've heard. Having an advanced degree already might help fill that void though. If you haven't programmed before, you may not actually like it. I'd recommend doing some small to medium sized projects first to also get a feel for it.
If you're writing business applications, you'll want to know something like Java (although they both have Java in the name they are completely different languages), C#, or maybe Python depending on what you're targeting. At some point you'll be dealing with databases and will have to learn SQL. A lot of business programming is what they call "full-stack" in that there is a web portion (often in JavaScript), a server portion (often in Java although there are many options here) which represents the "back-end" and a database which stores the data.
Doing embedded development on microcontrollers? C or C++ are both good.
Data Analysis? R, Python, or Julia and SQL will be a lifesaver.
Scientific programming? Python, Julia, Matlab/Octave/Scilab, Fortran, and C++ are all helpful.
In short, there is a lot of tools, languages, and frameworks out there. Don't let that freak you out though. You'll learn as you go by starting in one area and probably picking up pieces of other areas as you go.
And even if you like it and find it fascinating, programming for your own pleasure is a very different experience than programming 8+ hours a day with a team of people trying to meet an impossible deadline under pressure from management. After years or decades doing it, the excitement starts to wear off.
Also, the culture and work conditions vary a lot between employers. E.g., open office vs. enclosed workspaces; work-life balance vs. spending all your waking hours at the office or partying with your co-workers; contempt for "old" people or people with differing political views vs. a diverse group of people who can get along with each other; etc.
Before investing a lot of time and money into this, you may want to spend some time talking to people who've spent years doing it for a living.
My recommendation would be to check out CS50: https://www.edx.org/cs50
This site can fill in the rest: https://teachyourselfcs.com/
Incidentally, concert pianists are probably also not concert pianists, in the setting you want this comparison to apply. Yes, if you want to play with a major orchestra or win a classical grammy, you probably need to have been playing since you were 4 years old. But virtually no-one in our field operates at that level (despite a large cohort of developers who think they do); we are, almost all of us, constrained more by our environments than our potential. If the market for concert pianists was isomorphic for that of software engineering, a late-career switch to concert pianisting would also be reasonable to contemplate.
No we really wouldn’t. What I’d dearly love to be true is lots & lots of people to be given the opportunity to try the basics of writing software & the more complex stuff to be taught more plainly.
Even at the outer edges of the industry complexity spectrum we gatekeep way too much.