i should make games for older supercomputers. older in the way that they are no longer the bleeding edge of computing hardware and would therefore have a knowledge base on how to use them to work from, but also much more powerful than anything the consumer market is about to see in the near future. (this would obviously be a pre-singularity idea, as the “near future” will become way too short a timeframe)

sure the games wont have a large userbase, but they could break the way for pushing consoles to their limits earlier in their release cycle. the same reason why the games at the end of the release cycle can make use of the hardware the best… if you start using the hardware for gaming way before the release cycle even begins… you’re already a pro when it hits.