Nov 22
Like probably everyone who is remotely connected to mathematics, I read the four page remark.pdf by Alpoge, and like most people who have read that, I do not really understand much about it. What I can be sure is that, even embodied with the most powerful model in the world, I will not be able to prompt anything meaning regarding the Riemann hypothesis. As a separate timeline, I concluded (or rather, abolished) my Castile campaign in EU5 after winning the Italian wars. At the end of the gameplay, I felt quite bored with micro-managing so many markets, to the point that I was thinking about vibe coding a mod that mimics my style of gameplay, or even vibe-code a game for myself. Both leads to my (temporary) thinking of AI and ourselves.
If we do not completely exclude ourselves ---- I mean human beings as a civilization or species ---- from the universe (I do believe the alternative exists, as I am a gestalt player after all), it appears to me that AI does two things: makes some previously intractable problems solvable, and processes many solvable problems remarkably fast. I think the former is, from a scientific and technological perspective, strictly beneficial for human beings. When I was a high school student, I was quite worried that the ecology and climate of the planent will be destroyed. A decade later, I learnt the Ozone hole is recovering and (most) organic garbages can be completely burned into common molecules. There is no reason that terraforming and Dyson spheres are out-of-reach, though they may come in entirely different forms.
The latter is, in principle, as powerful as the former, but in practice, possibly harmful. And I believe it is the organization of our civilization, rather than the technology itself, that is responsible for the potential harms. For decades, academia holds the tradition that most (if not all) credits go to the person who first discovers/solves something. By saying credits, I mean very real things like whether someone can get paid for food and housing. In the past, the system works in two ways: for the established researchers who have grand ideas but limited time, the system encourages them to push the frontier problems, while disencouraging the participation in solvable problems that still eat their time. The latter usually becomes the first few papers by young students/postdocs, some of them will become established years later and make fundamental contributions to the fields. However, the advances of AI almost remove the ``limited time'' constraint, as one can solve the solvable problems in a remarkable fast pace, depriving many researchers of the potential credits they deserve. Similar things can happen (and is happening) in economy, as one can produce (material) products remarkably fast and cheap, denying the human workforce the access to jobs (and importantly, the associated pay).
We have been dreaming of a world with unprecedented material abundance, however, should that actually arrives (along with proof, papers, but not necessarily science, abundance), our economy and society are completely unprepared for that. As the majority will not be allocated credits for their intellectual efforts and physical work, it is essentially a dumping, of not only material products but also intellectual properties, from few individuals in possession of the large computing powers and already learnt knowledge, towards the majority of our civilization, which, according to my first economy course, may lead to the destruction of goods production and intellectual training of the majority. If one believes the majority deserve the rights of living ---- for the reasons of humanity, or genetic diversity, or intellectual reserve, or simply role-playing, whatever ---- this is entirely avoidable. As a civilization in a whole, there should be some ways of allocating credits to prevent the mass poverty and illiteracy, something our civilization has fought so hard against. In fact, the organization and diplomacy of our civilization is so advanced, such that even international minimal corporate tax treaties are possible. If the whole civilization wishes some institutional changes, there is no reason that can not happen. The real question are, who belongs to and also thinks him/herself belongs to the majority? what is the relation between the majority and the whole?
At last, there is a fat, cute, proud, ginger cat on the way from the MTR station to my rental flat, and it appears she never doubts her role as the master.