The tech transfer is great for business. It's iffy for development. NRE is expensive. Especially if you really want to make substantial improvements. The US just spent $16.5B developing the Block IV for the F-35. They can do that, because the cost will be spread over hundreds to thousands of aircraft. The Gripen will never see that kind of investment as a program. Let alone talking about Canada doing so for a few dozen jets.
Not as hard as an aircraft that is but designed for those upgrades. The F-35 was designed with substantial margin. For example, oversized generators. And even still the F-35 had to undertake a massive effort to increase cooling for their onboard processors, which was the limfac on their radar.
A 4th gen aircraft that was physically not designed to have that margin will have fundamental limits in capacity for upgrades.
The USAF took delivery of their last F-16 in 2005.
They aren't planning on taking delivery of F-16s in the 2030s like we are. They are simply running down the fleet they have while slowly transitioning more tasks to unmanned systems. The head of the USAF has said that eventually he wants three unmanned aircraft for every manned aircraft.
Your suggestion is that we use more manned aircraft for tasks that will normally be done by CCAs or drones in 2050, for most other major powers.
Some of the discussion here is proof that even a lot of military folks don't fundamentally understand automation and where that will be 25 years from now. This ignorance and recency bias leeds to a perpetual obsession with cockpit seat counts.