On the "discovery" of techs: it would be great if there were a way to influence or guide the discovery effort, by either having a choice of areas to focus on, or having an event message come up, or both.
In either case, the prerequisite to be able to "focus" on a specific area should be that you've discovered a "problem" you need a solution to, or maybe you've encountered something the enemy has but you don't.
E.g.: you encounter really heavy tanks enough times, your military then suggests you really need something to counter them. Then you could (either directly via the message, and/or manually) choose a "focus area" for the discovery effort. Which could be something as vague as "anti-armor tech". Then, the next tech to be discovered would be something to help against armour: for the mil res council that could be new types of weapon tech; for the model designers, anti-tank guns, RPGs, or heavy tanks for yourself; while for the operations council it could be OOBs that take tanks into account more.
Something similar could occur for the other research council, e.g. discovering better ways to protect from the environment, radiation, etc. - if those things are a threat that you've encountered.
Naturally the general pre-requisite rules would still have to apply - so an option to focus on a specific area probably shouldn't even appear if it's not possible to discover that particular tech/model/formation yet. So, that limits the usefulness a little bit of something like this, since what we can discover is already (somewhat) limited by what's already available; however, this kind of focusing should, in theory, make much more sense than just getting something random from whatever techs you currently have potential access to.
While there have of course been times where someone had a brilliant idea for something that came out of nowhere or was discovered by accident, in general research (especially military research) is far more focused in real life - even coming up with ideas isn't completely random, it almost always results from trying to find a solution to a real-world problem.
The only problem with this approach I can see is that it might make things TOO predictable. A possible solution for that would be that "focusing" on a particular problem shouldn't 100% guarantee that a potential solution is the next discovery, but instead it could just significantly increase the likelihood.