ren-n-stimpy
Yet I now have a PhD and a Best Paper award.
The key: specialize.
Study a sub-discipline so deeply that you know every paper's references' references. the reference set will collapse to a strongly-connected graph. jettison weak connections to get a clear picture, and, really, a narrative of the field, as in a survey paper.
just naturally, you'll start to be able to quote author affiliations, and their publication list in time-order.
once you do that, you'll find yourself hearing/reading comments, and immediately saying: oh, shut up, that has been done! that sounds like A, said by X, and B, said by Y, and even a bit like C, said by Z. and, then, voila, you'll see that no one has said D yet, even tho it is just obviously a variant on C, or whatever. and then you write it/prove it/code it. then you publish it. and all of a sudden, even though you've (say) never lexed or parsed or emitted or reduced... you start seeing yourself in other people's references as "D said by You".
Then you win.