Sure, there is quite a lot of example code for the 16F84, but there's also enough to get started on more recent PICs, like the 16F628 or 16F88...
your argument about it still being in production makes it seem like you don't realize the way companies like that operate. If their only market was hobbyists, it probably would be obsolete, but they have to deal with companies as well. the F84 was popular, and thus it probably ended up in a lot of commercial products, some of which are probably still in production or active use. Even for something as mundane as switching from one PIC to another, the cost of changing a product that is in production or active use can be very high, so it's easier for a company using them to simply keep doing so, and then phase them out the next time that they are doing technology upgrades for other reasons. I worked at a semiconductor manufacturer (where you would expect to see a lot of fancy technology) and they were still using some PAINFULLY obsolete software and hardware in important parts of their plant for just that type of reason (running OS/2 on 486 computers, or worse). Microchip then raises the prices sharply, so if companies really want to use them, it makes it worth their while to keep manufacturing them, and at the same time making them much less appealing for new designs... in addition to the other hints they drop, such as "not for new designs" and the strong public sentiment online that people should stop using it. Of all people, a hobbyist should be the most deterred from using it... there's no certainty that they will continue manufacturing it for long, it's way overpriced, and (compared to other, more modern PICs) it's really a piece of junk.
If you are really SO scared about getting started that you don't dare attempt it with anything but the "ford model T" of the PIC world, then get yourself one or two of them, take the example code you are so attached to, get a "hello world" program running, and then move it over to another PIC, such as the 16F628. Then you'll be starting with code you have tested and verified working, and only have to deal with the changes due to migration. And once you have successfully migrated to the newer PIC, take your 16F84's, throw them out the window, wave goodbye, and never look back.
As a professor of mine once said, "if you can flash an LED you can change the world". As most of us know, often by far the hardest part of getting started with PICs is successfully setting up the circuit and programming the PIC. Even having rock-solid example code won't help if you can't do that successfully, and getting that done is even easier on the newer PICs than on the F84, since you don't need an external oscillator on many of them. All you need is one "hello world" program that works on a newer PIC, and it can be just as good a starting point as a hundred 16F84 programs. read your books or websites, make changes to your 'hello world' code, and if things break you can backtrack to the last revision where things still worked. there's really not much reason you should need dozens upon dozens of example programs for anything but reference, in which case there's no real reason to actually use the F84 itself, instead of just taking ideas from the example code written for it
And I think that beginners can enjoy building their first circuits even if they use an old uC.
