Will literature survive?

It would have been much harder for AI to simulate the features that make us animals rather than humans. Once mankind decided to confine expression to symbols, once the symbol of cattle was first carved, its fate was sealed, though it must have been hard to foresee back then. These days I am thinking what “life” can be left in various forms of “symbolic expression.”

Computer code: This is the easiest prey for AI, since the most important feature of code is that it runs; only the select few who maintain it care about how it’s written, consumers don’t.

Math: This is strangely in-between computer code and the next, literature. Perhaps the most significant bit is still that the math “runs,” i.e., that proofs are correct. However, consumers of math actually read the code. I remain mildly shocked that AI can solve millennium problems as if they were tic-tac-toe, but still can’t quite write a paper that I enjoy reading.

Literature: This doesn’t need to run, it’s all about the way the code is written. Can AI produce literature that I will enjoy reading before bed, like a Nabokov tale? If you only know Nabokov from Lolita, read his tales instead — by the way he used to say that Literature and Math share the same “combinatorial delight.” At the moment of this writing, the answer seems that AI can’t. It would be most interesting if this task remains hard for AI, that no matter how hard it reads all that’s out there it still can’t reshuffle it in a way that’s both new and stirs us inside the way great literature does. (Literature was perhaps the biggest passion of my parents and they succeeded in transmitting it to me despite all my computer stuff.)

Music: This is the field I know least about, but it’s another excellent example, especially because we can restrict ourselves to pure score-writing, which seems more comparable to the above.

Graphical art: This would be another one, but it’s somewhat less formalized than the other ones. (This isn’t necessarily a problem for AI, but it is for me as I prefer to think of expression within a very simple symbolic domain, as is the case for the above ones.)

Even more on a personal note, after about two months on the west coast it seems I have finally converged on a nice hiking loop of about an hour that I can enjoy every day. It enters a natural area and climbs up to a stunning view of the bay. There’s even a bench. I mourn the math problems I will not be pondering during it, but perhaps there will be something else, like all the mega-exciting new results, or this blog post.