Things that can't be mass produced as easily: Music, sound effects, 3D models, 2D art, etc...
But of the four things just mentioned, 3D models and 2D art must heavily fit the visual style of the game. Visual studios that don't match are more noticeable than audio styles that don't match, in my opinion (though both are noticeable if extremely clashing).
Further, 2D sprite art, 3D models, and music is very noticeable if multiple games are using "each other's" resources - especially different developers using the same resources. I think it's a bit of a distraction to the immersion of the game get that, "Wait a moment, I've heard this song before... what other game have I heard it in?" feeling. Same with many types of art.
While there is room for some stock 2D art, 3D models, and music to pad out a game, I feel like games must have enough original material that the market won't shift to a stock-art only approach... but with textures I feel like the market will become stock-art only. For things like 3D models and music, I think the market will become more contractor-oriented. People who can understand a project and deliver customized material for it.
3D models *could* become stock art, if certain things happened. If models and the equipment/clothing they wear get separated out, with standardized formats, with standardized skin texture maps, and standardized rigging, so people just have several human models of different builds (short guy, fat guy, tall female, tall male, etc... all already rigged), and then buy hair, hats, skin textures, weapons, etc... for an assortment of different people using the same formats all already ready to snap into place with the ID'd bones of the model skeletons, and buy stock animations fitting their theme that "just works" with those 3D model skeletons, I think alot of 3D modelling could potentially become mass-produced for humanoid figures and stock furniture, vehicles, animals, weapons, rocks, and other objects. But there will still be some need for customized contractor work for things like aliens (aliens in one game are uniquely different in shape, texture, and animation then aliens in another game) and other things.
With this kind of art-related market, I feel like for each type of resource-art (as opposed to art that are finished products like games, movies, music albums, etc...), the market will inevitably be split into mass-production 90% dominated by two or three companies (the automobile industry or food industry today or steel production in the 1900s) or will be split into tens of thousands of contractors for hire creating customized work (plumbers, car mechanics, appliance repairmen, etc...). If someone is going to try to make a living in one of these markets, choose a market that requires custom work and thus is harder to mass-produce for. These are the ones that will most likely become contractor-oriented markets, and contractors usually make good money and work decent hours.
Cabinets are mass produced - one model of cabinet is the same in ten thousand peoples' houses. Kitchens are not (usually) mass produced, kitchens vary too widely in shape, so the people installing the mass-produced cabinets are independent contractors.
The key to becoming a good and successful contractor depends on: How easy you are to work with, how good a job you do, how quick a job you do, whether you actually show up in time, and how capable you are to find solutions to customers' individual and unique needs.
(Again, this is my own perspective - not a professional's analysis of the video game market)
As for making quick easy cash on the side, if it's really quick and easy, others will find it quickly, and then it'll become quick, easy, and crowded (like textures), and it'll become more of a marketing game (Who can get noticed the most?), and the difficulty will increase, and the prices will decrease. Quick and easy money doesn't stay quick or easy for very long, unless you are one of the first to find it.