The short answer is that nothing is wrong with most pond filters, pumps and other items, it is just they are not perfect.
When a new car comes to market, it is typically the result of 5 years of development including millions of miles of testing both on the road and in the test lab and even then – it is rare that a new model does not see significant revisions after a year or so, once real world feedback and issues are addressed. At this point, it is unlikely that a particular model will sell well without significant upgrades or enhancements every 3 or 4 years, simply because the technology moves on.
Making a car has several advantages over a piece of pond equipment most of which are to do with scale. When you produce 50,000 or 100,000 units you can invest in solutions that allow very low cost of production and also, as you develop the specific elements there is enough feedback to ensure that weaknesses are clearly identified and improvements are produced.
Pond equipment, by comparison is typically the clever idea of one enthusiast, developed over a period to a point where it seems to be better than alternatives, then 3 or 4 more are made and tried on other ponds, adjustments are made based on feedback and then the idea is turned into a production model using whatever technology is most appropriate for low volume / limited production runs. Any product that sells 500 items is a big winner in this marketplace.
Thus fibreglass was the production method of choice for most filters until Evolution Aqua came along with stainless steel and injection moulded units (see the Nexus 200 in use on my temporary pond back when the original pond build was underway). Recently and more pertinent to this website, some producers have started using CNC technology which apart from reducing production costs, has the huge advantage that it allows the benefits of large scale production while allowing production runs of 1 item (as each unit can be specifically customised for the particular user need).
It is this mentality that allows Tesla to differentiate themselves from other car users as when they come out with a new feature, you don't need to buy a new car - you simply download the capability (or order the component) and hey presto, your car is upgraded!
So, if the typical pond filter has had limited development and testing (especially given that ponds are far more different from each other than roads) and it is a series of compromises based on affordability and materials, it stands to reason that it is the ideal thing to consider hacking.....