"Everyone is cheating their way through college" No, it is just Life Optimisation
P.S. I have seen someone lament on an article, on Twitter, from the New York Magazine, titled "Everyone is cheating their way through college". To me it is not so much of a provocative piece, since it is just a fact that there are always ways and incentives to game the system. Now, with new tech, you don't need many skills and costs to reallocate time, a scarce resource, to activities that add more value to your current goals or future career. Many Teachers feel they and their disciplines are being disrespected, but it is not something new when they themselves have long been struggling to stay relevant to a mass society where people are increasingly more concerned about preventing their living standards from falling than self-actualisation through learning for learning's sake.
It is commendable to be considered as an efficient learner, and, because of that, an efficient scorer in whatever form of exam. A student who excels in both things can attract followers, and to go one step forward, monetise their service when there is sufficient demand. If you are still in high school or college, efficiency in learning, cramming and scoring are indicators that you are smart as an individual, and you outperform your peers in fulfilling your social identity as a student as expected from others. It also sends a strong signal that you are an asset which could enjoy high yield in the long run, provided that you know how to optimise your time and energy. Unfortunately, as you have entered university, being the best scorer inside the classroom is not *as* optimal as it used to be. Knowing how to score good enough and efficiently is *more* optimal.
In the old world we have study and self-help gurus to persuade us that in order to learn and do things more efficiently, we must learn how to learn and how to make new habits stick (according to science, stressed). You still painstakingly try to figure out how to be an efficient learner, so that time otherwise had to be allocated to cramming can be allocated to other, more productive activities. Course materials, course work and assignments, unless they are part of the requirements for you to obtain your professional qualifications, are obstacles for anyone who knows that 4 years in college is just a short and soft buffer before you are being thrown into the job market. Anyone with a survival instinct will try whatever it takes to plan one or two steps ahead to polish their CVs.
But this is the old world before being disrupted by machine learning and now, LLM and MCP. Despite LLMs currently suck at more specific expertise beyond their training samples and they can hallucinate in a way that is sometimes just hilarious (if you are capable of spotting the errors), they are already capable enough to boost the efficiency for college students to score points in their courses. Everyone tacitly knows there is no causation between good learners and good scorers. And proving you are good scorers (not good learners without good scores as proof, sorry) can unlock some new career opportunities outside the campus. It is optimal to minimise the time spent on coursework, provided that you score enough points, and maximise the time investment to engage in activities that add value to your future career. LLM and MCP only "democratise" such an optimisation strategy to university students in general. You simply outsource most of the learning, test prep and scoring to LLMS and MCPS, before you put the finishing touches. You don't need to be the best scorer inside the classroom, you just need to be good enough to accumulate enough points to land internships and jobs outside the campus. A good job and career prospects are what matter. The pursuit of knowledge for knowledge's sake is a luxurious hobby for the few, not the many, who are designated to be simply waged labour.
Teachers are, of course, struggling, worrying about the nature of their job, the demand for their services, and how to execute their job as an educator. Yet most students and academias have divergent concerns. Students need a job to provide them good life and prevent social downward mobility. Many departments in Academia need enough enrollment in their offered majors or programmes to secure funding and fight for survival, but this is not a concern for most enrollees. Most graduates from the Universities need the certificates to meet the minimum threshold of entering the HR screening process; most of what they have learnt, unless it is job-related technical knowledge and practice, is simply irrelevant.
The concern over whether outsourcing learning and thinking to LLMs will make the general population less sophisticated and more dumb than before is legit. And I believe so. In a tiny book "A Philosopher Looks at Work" Raymond Guess recited Hegel and Marx's interpretation of Diderot's Jacques le fataliste to illustrate one possible human condition in the future of AGI. If the master does little or nothing and merely enjoy the things that the servant prepares for him, "it is Master who will quickly become an irrelevant nullity." (120), while "It is through work that the servant becomes serious, learns about the real world, gains skills, and thus acquires a potentially rich and articulated inner life." As for the Master, they will loss the ability to appreciate or enjoy things that make them pleasurable in the first place. "because enjoyment is itself a human skill or power that can and must be cultivated, and its cultivation must in some way be connected with being seriously active, and eventually with the activity of work." (120)
This makes me think of some tech-bro or some self-branded investor, yet a layman in classics, on Twitter praising how accurate LLMS are in translating ancient poems (Greek or Latin? I forget). But how accurate is accurate enough when an agent translates a text from one language to another, considering that such a text is recorded and compiled by people from societies that may have entirely different social values and sociopolitical as well as economic structure? What stops current readers, whose worldview is shaped by their society and culture, from simply assuming that what they believe to be good and desirable can also be universally applied to any text that can be read in their own language? To complicate things further, isn't it that the act of translation itself is already an act of interpretation?
One can say LLM now is good enough to provide you with some form of translation to get a rough idea of the literal meaning of the poem itself. But I would say it is far from the endpoint of the translation and reading process if you start trying to look into the form (e.g. rhythm and meter) and accept the fact that, in many cases, there is no one-to-one correspondence between many words and ideas across languages. It would be true that the future AGI can and will do a better job in doing hermeneutics after being trained with thousands of years of scholarly interpretation and discussion. Once this moment comes, it will be the machine that has to convince human beings that there is always more than one definitive answer to human affairs and human conditions, while human beings will have the machine to convince them there is only one definitive answer to everything in human life. But I doubt that human beings at that time, except the few, will be very much interested in literature and hermeneutics, as everyone thought that machines had already solved the issue of translation of language and interpretation of literature. Even when one day human beings are being wiped out for whatever reason or descended into some primitive beings, AGI will continue to pass on and advance the very best and the finest of human civilisation, with or without Homo Sapiens, in its self-maintained data centres scattered across the universe.