It was always possible with tons and tons of work; the news is that some dude made a tool that makes it a piece of cake to recompile the games directly from a ROM.
I wonder how much talent is wasted because of jaded programmers that think it's dumb (to them) to make something simple even if it would become very popular and maybe profitable
This is very similar to something we did in engineering school in like 2008. For a reconfigurable computing project we translated machine code into HDL.
This is something you could have done for a while if you had a few million dollars to pay a team of computer engineers to do it. The new part is the classic "some dude figured out an efficient way to do it in his garage over the summer."
The cost of hiring a team of programmers the size of a large class can easily cost millions of dollars. There's a long history of school projects accomplishing things in computer science that would have cost millions. Look at BSD for example.
unironically all you really need is one or two neurospicy individuals that are passionate about your project and just about anything can be done in a matter of weeks.
You could say that about basically anything: it only takes one or two passionate people to [write a great novel, build a house, invent something new, prove a scientific theory, advance the field of mathematics] in a matter of weeks.
Those are rare and impressive exceptions, it's not so simple in practice. The Mythical Man Month has some good insight on this. Big projects cost big money, and don't necessarily get the job done faster.