Let's take the intent of patents at face value, ignoring the perverse incentives and broken review process that may jeopardize their utility even in the private sector. Patents claim to benefit the public by providing incentive to (a) invest in research that may be easily "scooped" by competitors and (b) disseminate information about inventions so that others can benefit. I do not find either compelling in the case of universities. The vast majority of research at universities occurs regardless of possible royalties because it is externally-funded and researchers compete for grant funding and prestige. Publishing a significant result first garners citations and prestige (and likely helps fund future proposals), directly benefitting both the researcher and their institution. Similarly, researchers publish important innovations because publications are literally the currency of academia.
You're assuming too much about academia and the people in it. We're not all automatons driven by the need to write papers. Plenty of people (eg grad students) are willing to commercialise research work given the opportunity.
Also, please don't assume that whatever can get published in a Nature paper is ready for mainstream use. It can take years of further R&D (that's not paper-worthy) to actually create something for the market. Dis-incentivising that process would be detrimental in the long term.
Note: my comments relate to science-based patents, not software ones (I don't really understand those).
Edit: If the problem is with the Patent system, that's where a solution must be found. Identify and deal with the actual problem, not the symptoms.
Grad students are primarily motivated to graduate so that they can be treated well and paid commensurate with their expertise. That is accomplished via the usual academic channels. Grad students that want to found companies are better off doing that. Yes, there is a huge gap between research papers and marketable products, but universities usually are not funding that conversion. Spinoff companies might, but universities want to get IP rights before the spinoff happens so that they can get royalties. Where is the risky investment in research that patents are supposed to be making possible?
> universities usually are not funding that conversion
Yes, they are. Universities are essentially a combination of a hackerspace with a startup accelerator, they have just grown to the point that you don't recognize them as such. The services they provide are essential to starting companies requiring more infrastructure than a laptop computer. They provide the multi-million dollar cleanrooms required to do nanoscale fabrication, the multi-million dollar spectrometers and chromatography setups required for chemical and biological engineering, seats on even the most expensive mathematical and physical simulation software, and mechanical prototyping facilities. They house a collection of experts, technicians, and cheap skilled labor (grad students) the likes of which cannot be found anywhere else or bought for less than a small fortune. They have industry connections that span many different market sectors and they have respect from society at large that gives them special dispensation to perform certain kinds of risky experiments (medical trials, chemical synthesis with reagents that could be re-purposed to produce drugs/explosives/nerve gas, and nuclear reactions). They even provide small stipends for their tinkerers.
These things cost money, yet many of them are advantageous or essential for starting a company, depending on the field (do you really think you can develop a next-gen neural interface, MEMS sensor, or drug delivery mechanism in your garage?). The arrangement is simple: a spun-out company can repurpose research done using these facilities if they promise to give the university IP revenue (the specifics of which are negotiated on a case-by-case basis). The nice thing about using patents to carry out this process is that they can be negotiated retroactively. Startup accelerators require equity up front, but that's not a workable arrangement if you won't know how commercializable your product will be until after you have spent your research budget.
I would hate to see this side of patent law disappear. Such "reformation" might just succeed in turning the university into the sterile ivory tower of arcane, pointless tinkering you seem to think it is. We need to fix patent law, not eliminate it.
You're still assuming way too much (and also over-generalising). Grad students leave their studies all the time, sometimes to start companies based on their work (some of my friends have done this). Others may leave academia and join the workforce and sometimes grad school gives them a pay/status bump and sometimes not (irrespective of whether or not they finished - which indicates 'usual academic channels' don't always matter).
I've realised from your final question that you're conflating two different issues. University research likely isn't driven by the desire to make patentable stuff. Looking back earlier in the thread, I think that's the point you're making and I do agree [1]. However, you're mixing it up with the separate discussion of whether (and how) the University should participate in the commercial exploitation of research work, e.g. when someone does want to patent something and make money from it.
On that point, it's worth comparing with what happens in industry. Employment contracts can state that anything you end up inventing during your service belongs to your employer [2]. Therefore, is it unreasonable that University, as an employer, would try to assert the same? I'm not making a judgement on this, rather pointing out a view that may be held by others (I suspect it would be an unpopular view in most of my circles and I'm not sure how I feel about it). In any case, that seems to be your main question and I'm sure it's a contentious one for any academic institution (it was at mine).
[1] Unless it's funded by pharma companies. I've had friends who can't reveal details of their work until the work is complete and the patent applications are in and the papers have been published (in that order). In all cases, they were partially, or wholly, funded by big pharma companies.
> Where is the risky investment in research that patents are supposed to be making possible?
Focusing on the "risky investment" part, the investment is not just monetary: I spent a full year of my grad research barking up the wrong tree. All I got out of it was a few pages to pad my thesis with.
Just like publications, as useful as it is to future researchers, you can't make a thesis out of "we tried this, it doesn't work", and a huge chunk of research work is exactly that.
why do we believe that commercialization of scientific product requires patents? Let's take the extreme case of pharmaceuticals, which is hard "like hardware" and requires a ton and half of R&D... Salk and Sabin released their polio vaccines without patent, and it was commericalized just fine, even a product which was from the get go intended to self-expire as a result of eradication. Neither of them died in penury.
You are assuming everyone is working at a well-funded research institutions. There are smaller universities where research grants are necessary and patents are also a way to keep someone for promotion. When you go to some scientific /engineering conference, you'd find many people holding one or more patents. This is almost a competition and a ranking system. The more patents you have, the prouder you feel. Anyone can conduct a research, with or without research grant, but patent is not. When you have a patent along with a published paper, people look at this product very different - this is good - otherwise someone else would have already done it.
