What can a scientist learn (only) from patent literature/database rather than from research literature?

Question

I'm currently diving in a field I'm no expert in, mainly I have methodological expertise and a theoretical background, but not in the research object itself.

Now I'm seeing this is a field of applied phyiscs with really a lot of patents being published, still a fundamental science field. To me it's clear that in branches like AI or quantum computing research the Intel, IBM, Apples research groups are probably several years ahead in some research questions in comparison to academic research groups in expertise, depth and equipment and probably most of their secrets and knowledge is not published, rather kept secret and in the best case (for scientists) filed as a patent. Of course they collaborate with some academic groups intensively and also publish papers (last weeks quantum computation record of google unintentionally became public). But papers can only be published years after a patent is granted to my knowledge.

Though I'm no expert in patent research and databases and language. But as a researcher/postdoc can I risk to focus only on journal publications to inform myself and stay up to date or is not a necessity to also look up the recent patents as they are more up to date in state of the art and knowledge than papers? Often keywords don't help here as patents are written very different to papers. So also practically, patents are really not good to self-inform.

What strategies are there to not start research on something already solved by industry? Attending much more industry symposia? Skimming patents? Has industry in general a high/low incentive to share their secrets and questions/problems they try to solve? In my field a lot of governmental and industry money (several hundred millions) are flowing into this research branch in form of academic funding. How do you deal with this problem/opportunity to keep yourself informed? Reading the most cited patents like for papers and do citations have similar value at all in patent literature?

Answer 1

This may vary by field, but in my opinion most patents are not very useful for a researcher. The patents that I have read are all written by attorneys in attorney-speak. Trying to figure out what is the real technical content behind the attorney-speak is difficult if not impossible.

To give an example: I used to work at a large company. My company wanted to patent some of the work that I had done. My boss had me send my material to the attorneys, and they drafted the formal patent application based on my work. When I received the patent application back, I could not even understand what the patent was talking about, and it was something that I had literally invented. So good luck to anybody else trying to understand what it means.

Occasionally I have come across some good technical content in patents, but it's usually in the situation where the researchers have filed a patent first and then published a journal article immediately afterwards. They did the patent just to protect their intellectual property. In that case the patent does have some good info, but the corresponding journal article is even better.

Again, your mileage may vary. In some fields I guess this might be different.