1. Demand
The problem isn't just that fighters are manned. It's also the concept of aircraft designed primarily for air-to-air combat that's getting less relevant.
After Vietnam and Yom Kippur (1970s), wars have seen a decline in symmetrical air-to-air combat. In the Falklands (1982), most Argentinian fighters skipped air-to-air armament. Most of what happened later was interception missions - multirole fighters shooting down light bombers (1990s).
That both last major air wars happened in the 1970s is no coincidence. By 1980, computer tech had enabled surface-to-air missiles to pursue maneuvering targets; SAM kills have been going up and air-to-air going down ever since.
Even when two powers with large, modern air forces clashed recently, they didn't engage in epic air superiority battles. They have set up theater-wide air defenses, killing anything that flies deep inside, and use their aircraft sparingly, close to the frontline. Both do deep strikes on one another, but they use missiles and suicide drones instead of interdiction missions. The odds of coming back from those are too low to invest in reusable aircraft.
Does this mean fighters are dead? Not yet. If two superpowers clash over an island chain, distant enough from either of them, they might not have the same ability to "castle" behind air defenses. It's just that this scenario is a small subset of possibilities, and most fighting happens over land.
2. Substitutes
When one needs to penetrate layered theater air defenses, the loss rate for fighters can get as high as 30-50%. It was such in WWII, it gets back there with S-400 or Patriot present. This means a reusable manned platform isn't perfect for deep penetration.
Drones are harder to spot, flying low and slow. Missiles can fly hypersonic (manned fighters aren't effective there), making them harder to kill. Neither incur PR costs or give the opponent prisoners to interrogate or trade. This advantage is shared by remote-controlled fighters. But the other two - the ability to evade interception, or to easily replace losses - aren't. Fighters are slow to replace not just because of the complexity, but also because they're built to be flown, not stockpiled like missiles.
Stealth is the one counter manned fighters still have to extend their life. But it's not perfect, it's expensive to achieve and maintain, there are counter-stealth technologies and methods. Theater air defenses are being designed to exploit these weaknesses. There's still some cat-and-mouse to be played, but expendable drones and missiles are a safer bet.
3. Complexity
For a remotely piloted combat drone, or UCAV, copying the design of a jet fighter isn't optimal. Fighters pay a lot in weight, fuel consumption, and maintenance costs for supersonic flight, high thrust:weight, and 9g maneuvers, all unnecessary for a drone.
Practical UCAV are built as simple and efficient subsonic stealthy platforms, giving them range without refueling. Starting with the MQ-1, they could carry a fire-and-forget missile for air-to-air. It's not as good as a fighter, but the drones are much cheaper, and with good enough missiles, they can hit bombers or deter fighters.
Converting old fighters to drones can be cheap at first, but their maintenance and overhead aren't. Air wings comprise 30-50 men per fighter, which won't go down much for the same plane, just remotely piloted. That's why QF-16s are used as disposable targets and MQ-1/MQ-9s as reusable combat vehicles, not the other way around. Drones are built for low maintenance.
4. Has it really passed?
Keep in mind that Elon Musk is in the tech business, which does give him a bias for unmanned systems. He builds them, after all. But as spectacular as military tech evolution has been, nothing has replaced "boots on the ground" for the long game yet.
Manned fighters are "eyes in the air". They give the pilot better view of their surroundings, better feel for the aircraft, skin in the game. They'll remain relevant for patrolling non-hostile skies and managing unclear rules of engagement.
But as to how far they'll advance beyond F-35, FC-31 and Su-75 - that's less clear. I expect one more generation for sure. The real question is if there will be another generation of heavy fighters (F-22, J-20, Su-57 successors, with certain risks), or if it will all be light fighters and drone controllers from now on.
Post-Cold War, light fighter programs have been a lot more successful than heavy fighter ones. They provide "eyes in the air" for less, and heavy ordnance can be outsourced to drones now. If I had to guess, I'd expect China to follow up on the J-20, but Russia needs drones more than it needs fighters, and while the US is developing the PCA, it's seeing much less focus than unmanned platforms. India and the EU have been going all-light from the start.
Summary: the role of fighters has been partially replaced by drones, but all major players are still pursuing manned fighter programs, mostly light ones. A human pilot is, with few exceptions, a known quantity; AI systems and remote links are not.
