Since the US patent office was founded in 1790, patents have transformed both qualitatively and quantitatively, both in the nature of the patents themselves and how they are used. There are much more detailed works on this, like the excellent The Patent Crisis and How the Courts Can Fix It so I’m going to touch on the highlights enough to hopefully convince you that patents have become an increasingly worse benchmark for innovation, value, or really anything besides demand for patents over time.
The number of patents has exploded over time. The number of US patents issued per year has increased by four orders of magnitude since the patent office started - from 33 in 1791 to 354,430 in 2019 (this is two orders of magnitude more than the US population growth.) The number didn’t go above 1000 until 1854, and didn’t go about 100,000 until 1970 (Expanded list of WTF happened in 1971 and afterwards 🤔)
Things that did not increase by four orders of magnitude: the size of the PTO or the amount of time it takes to process a patent (though the latter has increased significantly, which creates its own set of problems.) As a result, each patent examiner has to plow through patents as fast as possible, which combined with the legal mandate to default accept patents unless there is a strong reason not to means that more BS patents get through the system.
The number of patents also reflects how their use has changed over time. In the past, their use was primarily what most people think of - a way for inventors to get paid for their invention. Modern patents have taken on many more roles: tools to exclude competitors from entering an entire market, defensive measures against other patent-based lawsuits, and an internal performance metric for companies like IBM. All of these uses incentivize people to patent anything they can think of, decreasing patent quality.
Anecdotally, interactions at a patent-obsessed company go something like “So I was thinking it might be possible to …” “You should patent that!” “I build this piece a little differently…” “Did you file a patent?” “I got dressed in a different order today” “PATENT IT.” Even though there weren’t even monetary bonuses associated with patents like there are at some companies, the attitude alone had an incentive-warping effect.
Transformations in the qualitative nature of patents over time has also increased their BS factor. Until the 20th century patents were primarily concerned with mechanical inventions (although the first US patent was a chemical process - improving potash creation). This meant both that it was conceivable that a single patent clerk could understand most patents and that it was possible to create aworking model of the invention (which feels like a natural check on BS).
The explosion of types of technology in the 20th century meant that it’s much harder to guarantee that the person examining any given patent has deep expertise in the area. Patents went from purely mechanical inventions to chemical processes, to electronics, to biology, to software and combinations between all of them. There are many specializations within each of these areas, so while someone trained in biology might grok the basics of a patent immediately, it would take them a few days to understand it deeply. And we’ve established that they don’t have that time because of the ever-increasing number of patent evaluations, meaning more questionable patents slip through the cracks.
The incredible diversity and specialization of modern technology also means that patents no longer provide fulfillment the platonic ideal of being a complete, replicable description of a technology.^1 This incompleteness, combined with the fact that you no longer need to have built the technology to get the patent creates an asymmetry that incentivizes people to patent half-baked ideas that can then cut off further exploration in that area even if it isn’t a solved problem.
^1: A similar situation is true in scientific papers.