A hopefully useful thought on software patents

A while back I saw a post which pointed to https://defendinnovation.org/. It reminded me of a thought I had on software patents that I hadn't seen expressed anywhere else. I decided to post it here to get your feedback.

I had always had some sort of intuited judgment regarding software patents, and my judgement was that they were wrong. That they were unjustifiably granted. But during an argument about this with a co-worker, I couldn't quite articulate a compelling argument as to why. They just were. I had trotted out a lot of the sorts of arguments that we all make, but they just didn't feel compelling.

I thought about it and slowly realized that I actually didn't have a problem with software patents, per se. It was that software patents were an especially egregious example of an underlying problem I had with patents in general. The problem was this:

We, as a society, pay the same price for different goods regardless of value.

The price we pay to incentivize an inventor to disclose their invention is with decades of protection from competition. For a utility patent, that's 20 years from the date of the patent grant. And we always pay the same price, regardless of the underlying value of the invention to our society. So whether you invent a new method of swinging in a swing or a new drug to fight cancer, we pay the same price to know how it works.

And I think that's one of the key perspectives that often gets overlooked. We are paying for knowledge. The inventor is supposed to have figured something out that even skilled practitioners in the art hadn't thought of. We are paying for the secret sauce, so to speak. I think sofware "inventions", for many of us skilled in the art, seem so obviously non-insightful that we can't believe society has paid for such "knowledge" with a patent. Many of us realize that we don't even need to read most software patents to learn how to do what it explains. We're paying for "I was here first", rather than "I finally solved the problem no one else could". Yet we pay the same price regardless. That's why the Defend Innovation site wants to limit software patent terms to 5 years. Patent protection should be proportional to the value of the invention. But I believe that should be true for any invention.

That's when I realized I wasn't against software patents at all. I was against paying the same price for items of different value. Given that revelation (in my own mind), I would actually be fine with a software patent if its contents were the only way to learn how to do something that was exceedingly difficult (many had tried, all had failed). But that applies to all patents (at least utility). So I wouldn't agree with the 5 year term for software, what I preferred was for all patent terms to reflect their value. Most software patents would probably not get a patent term long enough to cover the application process (1-click). Some might deserve 30 years of protection (general AI, etc). The same goes for non-software. It's conceivable that a patent might actually be worth 50 years of protection (a small battery that weighed an ounce but provided 100 years of current, etc).

For me, then, this was really the essense of, and the strongest form of, the argument against software patents. Since I was so familiar with software it was much more obvious that we had all over paid.

Thoughts?