Software Patents Are Bullshit

Patents are enshrined in the U.S. Constitution for a single purpose: to promote the progress of science and useful arts.” Alas, when it comes to software, it is difficult to imagine a system worse at this than the current one.

Everyone knows this–and most people will readily admit it–but no one, with the possible eventual exception of the Supreme Court, is willing to do anything about it. No one is sufficiently incentivized to fix a system which hurts everyone a little while helping a handful of trolls and corporations a lot. Even people and organizations who would benefit from radical reform have been so traumatized by the status quo they now have an irrational fear of change.

And so: “Two days ago, Bank of America was granted a patent on [nmap].” Patent war goes nuclear: Microsoft, Apple-owned ‘Rockstar’ sues Google.” “Google sues to protect Android device makers from Apple-backed patent hell.”

Oh, it goes on and on and on. “Apple Wins Big Against Samsung In Court.” “Google Awarded Patent For Free Rides To Advertisers’ Locations.” “Court: Google infringed patents, must pay 1.36 percent of AdWords revenue.” “Uber-troll Intellectual Ventures faces Motorola in first patent trial.”

Etcetera, etcetera, etcetera, etcetera. And for better or worse — mostly worse — the US patent system influences and infects its counterparts around the world.

It’s true there’s been some progress of late. The US House passed the Innovation Act, which “is designed to target the patent troll problem.” Better than nothing, I suppose, if it passes the Senate; but who are we kidding here? The problems with software patents do not begin and end with patent trolls. Software patents are fundamentally harmful from the word go.

As Matthew Yglesias puts it in Slate:

The problem is that we’re granting far too many patents, tying up vast swathes of industry in litigation and negotiation rather than innovation … In the thriving digital sector, patents don’t work at all, and more patents are likely to cause more litigation rather than more innovation … In almost no sector of the economy do politicians talk about their desire to promote more monopolies and less competition, but that’s exactly what the recent round of patent reform is all about. America […] needs more thinking about how to roll back harmful monopolization without doing too much damage to the sectors where the system works.

Does society benefit from the existence of software patents? Hell, no. In their absence, is there a risk of vitally important software secrets being kept secret for generations by insular guilds? Hint: I’m a software engineer by trade and it was hard for me to type that last sentence without laughing aloud. I suppose it’s possible that there may be some kind of darkly brilliant genius software innovation that no other team of developers in the world could ever stumble upon. But that seems pretty goddamn unlikely to me — especially when contrasted with the enormously negative costs of the software-patent status quo.

It’s sometimes argued that patents protect startups from larger companies, but in today’s software world, that protection is unnecessary. Startups can and very frequently do succeed without the benefit of patents, simply by executing better and faster than the incumbents. As Mark Zuckerberg put it on Facebook’s tenth anniversary:

When I reflect on the last 10 years, one question I ask myself is: why were we the ones to build this? We were just students. We had way fewer resources than big companies. If they had focused on this problem, they could have done it.

The only answer I can think of is: we just cared more.

Patent law is a disastrous affliction which hamstrings the entire software industry. It is essentially legal DRM, and as I once said of DRM, “it reminds me of the great Ryszard Kapuściński‘s depiction of the Soviet economy:”

One can assume that a significant portion of the Soviet metallurgical industry is devoted to producing barbed wire … For the matter does not end with the wiring of borders! How many thousands of kilometres of wire were used to fence in the gulag archipelago? … If one were to multiply all this by the number of years the Soviet government has been in existence, it would be easy to see why, in the shops of Smolensk or Omsk, one can buy neither a hoe nor a hammer, never mind a knife or spoon: such things could simply not be produced, since the necessary raw materials were used up in the manufacture of barbed wire.

How much energy is wasted, how many millions of dollars are squandered, how much innovation is stifled because of software patents? The total costs are immense. And the benefits? They honestly seem nonexistent to me.

We could mitigate this damage by giving software patents a shorter lifespan, say 5 or 7 years instead of 20. A useful way to think about this is the Tabarrok curve:


It genuinely seems to me, though, that with software patents, that curve only bends downwards, and while five- or seven-year patents would certainly be much better than the status quo, zero would be better yet.

But unless the Supreme Court steps in to cut this Gordian knot with an unexpectedly enlightened sword, I can say with considerable confidence that nobody will do anything about them. The system has evolved into a status quo wherein nobody has any incentive to do the right thing. That problem, alas, is enormously larger than mere patents — and commensurately more difficult to fix.