They encourage people to want to. Few do it, and they have to at least buy one to start.
Try lead shot in the resin, after a thin nonconductive coating on the traces/leads.
A few dollars vs. in some cases a $5000 product.
I find that tough to believe. RAM was expensive. Drill bits were cheap.
While that's a tough statement to refute, that doesn't mean making it tough for copyists is bad business.
If you make it a hell of a good challenge, you will do a fair business specifically to crackers, whackers and so forth who never would have bought them for the value of the item itself. In addition, each cracker will have to permanently deface and make unremerchantable the test unit if you do a good job.
I can think of a number of relatively mundane products in specialized niches that have not been cracked that have been available for years. They are profitable products.
Philbrick made the solid state operational amplifier module that was an insanely profitable little bastard for years and no one else ever successfully cracked it until it became a moot point via the op amp IC being cheaply available.
There are a cornucopia of little plug in "power enhancers" for cars and truck and agricultural diesel engines, all of which work on the principle of altering the analog or digital inputs from sensors on the engine to the ECU. Most are designed for easy removal so when the engine blows up the customer can remove them and say "Who, Me?" to the dealer. Most are constructed to be tough to reverse engineer. The makers are not stupid.
Ironically, a very successful aftermarket product for the Amiga, which DID, like several aftermarket Amiga products, make the builder a fair profit, was the notorious Emplant. With the Emplant and the right magic code plugs your Amiga was a Macintosh....and an Atari....and a Sun workstation....and a Next box. It was very cunningly designed and was a good example of a seemingly straightforward product that wasn't, because of a stellar bit of misdirection that has AFAIK never come out yet.