I think the solution is as you say -- make assignments that are hard to cheat on -- but also reduce the incentive to cheat. Just make it less important. High-stakes closed exams and emphasis on ranking students will certainly make cheating worse.
In one university, absolutely everything was assessed, every bit of lab work, every tutorial, the lot. Unfortunately, this means you're being assessed on lab work which might fail due to your inexperience, and it makes it unnecessarily stressful.
In another, the lab work and tutorials were almost never assessed. They were learning exercises, which you could try and fail at without any consequences. I found that you learned a lot more here, because you could try stuff out and fail without it affecting anything, and it was much less stressful.
In the latter university, you did three modules per term (~10 weeks) and you had a closed exam for each every term. New paper set every year. No coursework. No chance for cheating.
The exams were also not "high-stakes". Each one made up a small percentage of the course total, which increased through the three years of the course, so there was no problem with doing badly on one or more. It meant the term time was pretty much all about learning, with assessment right at the end. I'm very much in favour of this simplicity, and it makes academic misconduct almost impossible.
As someone who worked their ass off, graduated with first class honours and never cheated, rewarding cheating devalues the grades of honest students. It's absolutely unfair, and it needs properly policing. Good academics don't need to cheat.
There is huge pressure not to fail people, and it is incredibly difficult to actually fail--I know only of a small handful of instances. In one case, they basically didn't show up for six months on a 1 year masters and did zero work, and even then it took months of handwringing to actually kick them out. But it shouldn't require months of agonising if the rules are clear and consistently and fairly enforced. The problem is that they are in a thrall to rich overseas students' cash, and they don't want to rock the boat. They even have special dumbed down Masters courses which are strictly for overseas students only. An academic friend of mine who was a fellow PhD student teaches on one, and he wasn't happy at the standard of the course or the academic conduct of the students. If they were on a regular course they should mostly have failed. And there's also the issue of league tables--being lax in grading and punishing cheating gets you a higher ranking in the league tables, so there's a strong disincentive to have high standards.
At this point, I do think that universities need some sort of external intervention. If standards are to be upheld, academic misconduct must result in real consequences, i.e. being kicked out. Most universities have a session on academic misconduct for all new students, and they have to sign a statement saying they understand what academic misconduct is and that they understand that it is not permitted. That needs enforcing properly, no second chances. As for cheating on coursework, there's a simple answer to that: go back to closed examinations for all courses, and make the coursework component a minor part or remove it entirely. And structure the coursework to minimise cheating. For example, make every student's work unique so there is no possibility of using anyone else's work. As a scientist, lab work and written work were usually your own experiments and essays, so little chance for cheating.
The real test for the cheaters is when they later have to perform in the real world. There is usually a clear difference between motivated and competent people who are independent thinkers and doers and lazy chancers who struggle, and it's quickly apparent which category people fall into despite their paper qualifications and CVs. Whether you're in a university or a company, there are people you can rely on who know their stuff and can deliver, and the people who can't.