A friend of mine is retired from pharmaceutical research. Where he did a lot of computer work. His wisdom is that the ease of patent granting should vary enormously by the type of invention being patented, and difficulty of being a credible would-be inventor in the field. And the thing patented must exist, and actually work. (So no "if we use some unobtainium..." patents.)

So - invent a new drug (for humans), where safety regulations guarantee that actual approval to sell the drug will cost many, many millions of dollars? Patentable. (Yes, there are other issues with modern drug companies & patents - Not. The. Point.) If somebody comes up with a working, practical fusion reactor? Very patentable - that can't have been cheap to do, and "thousands of experts have been trying for decades" is proof that this inventor really is something special.

Vs. software, abstract ideas, waterbeds, scanning to e-mail, a clever new way to peel potatoes, etc. - sorry, but patents should simply not exist for such stuff. You can only copyright your code, trademark your brand, etc.