That's 4 years runway in my country, living a comfortable life. I'm certain I could build an app that makes more within that period of time. Or even if it's not a profitable app, it would be substantial research into... something.
There's an untapped customer market of Solo RP. Basically people who take something like D&D and become the player and the GM, but most don't enjoy playing both roles. It's not necessarily D&D; there's hundreds of tabletop rules that people just want to experiment with but can't find interested friends. Video games don't really work because they're railroaded and don't give as much agency.
There's some basic tools like Mythic GM emulator, which is basically a set of rules to make things surprising. This is something that can be done 10x better with just an app, and done 100x better with an app + GPT-3.
From that, you can branch into a new genre of gaming based around AI.
Some will bring up AI Dungeon, but the downside there is that AI is terrible at maintaining state. The game is good at simulating a date between you and your possessed spear. But 15 mins later, it will forget about your demonic spear boyfriend. Humans are great at maintaining these states, and making the call whether this should be a big part of the game they're playing or not. AI Dungeon is poor at conflict resolution - you stab the dragon once with your magic spear and there's maybe a 50% chance of instantly killing it, because it's not tracking injury or difference in strength, or even plausibility. Humans and RP systems handle this much better.
Can something like this make over $200k? I'd bet on it. AI Dungeon certainly did.