My take..
1) Firmware is stored on the platters for economic reasons and ease of mfg. When you can knock fractions of a penny off of a production run of 10's of millions of the widget, well, the answer is clear. There is substantial savings. If you can do two or three pennies, WOW! There is now hundreds of thousands of dollars in savings, and that's just 1 product lineup.
2) I consider "firmware" to be fixed code, generally not user modifiable, small in size, and serving a highly specific function. A backbone function which is mostly transparent to the user. How it is stored is pretty much irrelevant.
3) In the years I've been "messing with computers" I've never had to ask a customer to upgrade or change disk drive firmware. Drive firmware is something the end user, even the technically inclined, should not be messing with. There are too many engineer wannabes anyway; they end up breaking things by tinkering. Part of the inaccessibility of the disk's firmware is also part of its stability. And security.
4) Regarding reliability. No HDD company is carefully engineering their products to last x amount of days and then fail. They would not be in business very long. Products may seem to fail more often today than yesteryear, but the reason is cost-cutting and pushing a design to the limit. Engineering and marketing are constantly fighting a battle about how fast and how far something with go. In the case of HDD, how much data can be stored reliably. Sometimes the outcome doesn't favor the consumer and you get a buggy product. A fact of life for all mass produced goods.
5) A disk with onboard hardware encryption is really no different than a standard non-encrypted configuration. In fact as time rolls on, all disks will be made this way. Cheaper than stocking 2 sets of parts for encrypted and non-encrypted models.
6) HDD mfg. will never ever maintain a database of firmware code and serial numbers for end users. This is not economically beneficial to the company. It also opens up huge liability issues.
7) Soft peripherals like sound and modems and network will only continue to gain in popularity. More and more functions will be offloaded to the CPU. And software will make it happen.
Everyone needs to understand that today's consumer electronics are not engineered to be played with by the end user. This is not the 1970's and the days of the Apple II. I remember they sold firmware updates in retail shopping malls back then. You could update the language resident in the computer, and even the software on the disk controller card. Thus gaining more commands in BASIC or getting more storage space on your floppy disk drive. The same forces that give us these super sophisticated cool toys also come a certain amount of "closed system" architecture. This isn't going to change anytime soon.
9) It is common psychology. When a product or "something" fails and becomes unfixable, we look for a replacement and totally focus on the weak spot that bit us in the ass the first time. We crawl into a rabbit hole with blinders on and totally ignore the big picture and other great features the replacement may offer.
10) Any high-density storage device, today, is complex enough to be treated as a black box. It either performs or it does not. There is no inbetween. Any sort of failure, you dispose of it and get a new one. It isn't worth the time or effort.
11) It is always important to make two copies of anything you wish to keep. In other words backup. If you don't you will learn a lesson one way or another. A properly prepared backup allows anyone to become their very own Sunday morning data recovery technician.