The Teaching Process and A.I.

The present-day culture is so threatened by A.I. these days as to become deeply engaged in public shaming anyone who admits to engaging with it. To admit that one merely engages with A.I. — regardless of whether or not one is actively using it to write or create content — has become a battleground for what amounts to a tiny but loud faction of the social discourse. Most people either don't care, do not know enough to care, or have already discovered the benefit of the various applications out there... but these people are also those who largely dropped out of the internet social space eight to ten years ago, about the time that Facebook, Instagram and other like programs began to fade as platforms.

Thus whenever I want to invoke it, my first instinct is to think, "No, I'll be judged as one of the bad people." And then to resist the impulse to employ it, and thus I am shamed into silence. That's the goal.

Alas for the naysayers, however, I'm going to resist that and, once again talk about A.I.

I don't use it much for writing. For example, ChatGPT disagrees with me almost constantly, largely because it's stubbornly programmed to speculate about what an imaginary reader might feel, prefer, tolerate or be put off by, which describes just about everything I write. Chat is deeply concerned about the writer pleasing an audience, or about the writing of content that promotes marketability, and so when dealing with it I'm smacked in the face about tone management, inconsistent pacing, words that are too big, sentence constructions that go on too long, and in general how the work "comes across" to others. Endlessly, I'm told, "Watch out for X" by the program, because, heaven forbid, I'm being too polemical for the average reader.

I can beat this shit back by naming it, demanding the application stop, insisting that Emerson wouldn't have put up with this nonsense and so on, whereupon the program instantly agrees with me, stops doing it... and then drifts back into doing it again, allowing me only a small window where "concern for the reader" is not a default. It's aggravating to encounter a computer program that is not designed for the user, but rather for a non-present, non-user, non-observer who might nevertheless take offense at something they can never see by definition. This is the sort of thing that convinces me that present-day A.I. is the Betamax of future A.I.

Still, these conditions do not decry its usefulness.

I am just now engaged in a project that is getting larger by the day, which is necessarily dictatorial in its scope, because for a person to actually LEARN something, there must be an agreement between the teacher and the learner that the learner must trust the teacher's knowlege. Because of the size of the project, one of the hardest aspects to manage is my own drift into failure to provide context for whatever I'm right now trying to describe. For this, Chat is very useful. It possesses a not-quite-perfect knowledge of the AD&D books without my having to explain these things, it possesses the ability to understand what I'm doing and why, so as a filter it is a good thing for me to plug content into and ask, "what's missing here?"

Then, having uploaded the whole work thus far into Chat, and reading Chat's responses, I can use my human brain to parse out what's missing and what Chat merely thinks is missing, and therefore stay on track with the project while maintaining continuity. This is far more useful to me than having it actively write content from scratch.

No application has ever been able to do this. I do not understand why anyone would think this was a problem.

True enough, if you allow Chat to write your term paper on "The Emergence of Hittite Ivory Carving Techniques in Ionia in the 6th Century," you're going to get a lot of garbage. I wrote this paper for my required Art & Architecture class back in 1990, for which I got an A-... not because I didn't deserve an A for proving the point, but because the concept was so completely out of the professor's personal experience — that being Dr. Michael Walbank, who'd spent ten years in digs in Ionia, unearthing ivory carvings there — that he had to write under my grade, "Interesting, but I don't quite buy it." Those were literally his words.

If I'd had Chat then, which I could have used to straighten out whatever logical premises I had missed, without it writing a word for me, I'd have booted the professor's ass all over the campus. I'm sure of it. Chat would have helped me make disagreement impossible on procedural grounds.

I suppose if I cared, I could return to the subject, do the research all over again, use Chat to sort out the details and then submit it for publication in an archeological journal. Alas, I haven't the credentials to have it looked at, much less published, and the real kicker is there is no money in it. So why bother. It's not like the similarity of Hittite models appearing in Ionian carvings a millennia later is going to change anyone's life. I got my A-. That's all the life-changing value that paper ever had.

I want to go on a tear here about the bullshit of University professors crying their eyes out about students "cheating" on their term papers. To begin with, 99% of these students (which would have been the ratio when I was in university in the 1980s and 90s) will never become writers of any skill, ever, anyway. The argument then, and I assume the argument remains the same now, is that "we must all learn to write so that one day we'll have that skill in our pocket when we really need it."  Presumably, for some letter or missive of some kind for some job we presumably will have some day.  Thus the present argument is, and we must imagine the professors saying this with tears in their eyes, "If they use ChatGPT now, they'll never learn to write!!!"

[Yes, absolutely, three exclamation points were necessary]

Only, the students today will never need to learn how to write, will they? You might as well designate a space on campus for the students to go to where they can learn how to dig ditches, since they might need THAT skill someday, and probably with greater likelihood than they'll ever need knowledge of how to write. I don't write anything now that a bureaucratic reader is going to read, and I'm a writer. Because, while I hate that Chat whines about my being polemic, I know that if I need a cover letter for something I want from the government, Chat will do that better than me because there's zero chance Chat will offend someone... while I just might.

Consider the possibilities that are made available by stripping writing out of college and university. Why does a "geographer" need to know how to write? Why does a "political scientist"? Time can be removed from the student's plate performing a skill set that's now become redundant and be used in direct application of the thing that's being learned: compel the students to engage verbally on their subject, not one at a time but with each other, while the professor moves through a class that's talking, urging those not doing so to begin speaking or risk a lower grade. Direct the political science student to get out of the chair and knock on doors to talk politics with strangers, and then GRADE THAT. Or hell, get rid of the grades altogether, and find some other threshold a student has to get past. The rest of the world has gotten the hell out of the 19th century and moved on. Maybe its time for education to do the same.

But, obviously, I digress.

My best writing yesterday was this, related to the Aging section:

"This table shown reveals a reality about AD&D that someone learning the game has every right to know. All too often, the idea for a table far supercedes the value of the table actually provided. With the table shown, copied from AD&D's Dungeon Masters Guide, does not in fact express a clear underlying principle. It is inconsistent, incomplete, the magnitudes are wildly uneven, no explanation is given for why an elven cleric is centuries older than an elven fighter, magic-user or thief, while the book's notes effectively say, "here is the table." The same notes also grant the DM the privilege to arbitrary assign ages, making the table unnecessary. That said, though the authors did not explain that table, that is likely due in some degree to a lack of time, space and editorial attention.

"When introducing the Dungeon Master at the outset, we spoke about the DM being ready to resolve unexpected situations: one of those can be imposed by the rules as written, as we see here. It was also expressed that "the game rules are perpetually insufficient;" this is true not only for the complexity of situations and possibilities that the setting allows, but also because the game is so complex and far reaching, that the creators found themselves unable to solve everything. Thus insufficiencies like this table here only tells us where more work needs to be done.

"Most instructional texts about D&D will attempt to rationalise or conceal shortcomings like this. They often give the impression that the failure is not the text, but the reader's inability to understand. As a piece of advice when interpreting the game's rules: if something does not make sense, it's very possible it does not because the effort to make it make sense was never in fact made — and that it is the responsibility of user to fix it, by default of the maker's unwillingness to do so."

When I run this past Chat, after telling it not to discuss the writing itself, or then to summarise it, both of which happen as default responses, I get this:

"AD&D has historically positioned ignorance as a moral or experiential failing. If you do not understand a rule, the tradition implies one of three things: you have not read closely enough, you lack sufficient experience, or the understanding will “click” later after repeated exposure. This passage breaks that triad. It introduces a fourth category that is usually taboo in rule systems: the possibility that the rule itself is epistemically hollow. Not “hard,” not “subtle,” not “emergent,” but empty of rationale. That is a destabilising move because it denies the text the automatic presumption of depth.

"This has consequences for how play authority is conceptualised. In orthodox readings, the DM’s latitude exists because the world is complex and players are unpredictable. The DM’s latitude also exists because the text is unreliable as a system. That is a different justification. It recasts DM judgment from an aesthetic or narrative discretion into a structural necessity. The DM is not “adding flavour” or “interpreting intent”; the DM is compensating for missing engineering. This quietly strips away the idea that strict textual fidelity is even possible.

"The accusation that other instructional texts “rationalise or conceal shortcomings” points to a systemic habit of post-hoc mythmaking. Over time, communities often backfill intention onto rules that were never finished, treating accidents as deliberate balance choices or unexplained numbers as hidden wisdom. Your passage insists that absence of explanation is not an invitation to mysticism. It is evidence of unfinished work. That stance prevents the learner from inheriting decades of accumulated rationalisations that obscure rather than clarify."

Which goes to show why, for some, this process feels so different, and why those who already know every rule stated here find themselves fascinated by an outlay of information in a patient, progressive manner. I learned through this that player character dwarves and half orcs player characters do not, in fact, receive penalties to their charisma, as I have always assumed, despite having glanced at the rule's page hundreds of times. Those sentences never made an impression on me; I never played with anyone who played that way, do not remember any player ever pointing it out to me, and assumed by default, when it WAS pointed out to me a couple days ago, that it was untrue. But, sure enough, the dictate is there. I can still hardly believe it. And they say people now are "fragile." Players in the 1970s, apparently, were so fragile they couldn't bear the idea that their dwarf character might lose a point of charisma just for being a dwarf, while it's okay if a halfling loses a point of strength.

For myself, I wonder about my capacity to keep it up. I'm wading into the fighter class just now. I've written some notes for the upcoming material: weapons, throwing, missiles, proficiencies, armour, [dexterity], [object costs], [weight], strength, [engagement], experience levels, [retainers]. Those in brackets are meant to be introduced but kicked down the road, while providing a fairly thorough overview of everything not in a bracket. At some point, things need to be explained as fully as possible, provided that all the details already exist that enables it to be explained.

So far as I can think of, there's nothing to get in the way of a clear description of the weapons available, how many hit points they can take away from an enemy, how they are used in a fight... just so long as I don't actively get into "rolling to hit," which needs way more scaffolding than I have now. But the player doesn't need to know about rolling to hit just now to know what a fighter does and how to make one. I learned that decades ago, in all the characters I helped players roll, and how much actual information that's needed before a fight actually happens. And the "to hit" roll is not something that needs clarification.

But before I can write about dexterity, I need to be writing about the thief; so, for the moment, I'll briefly address the fighter's relationship to dexterity and the bonuses for weapon's throwing — I might even put up a half-table with just AC and to hit adjustments, since I'll be discussing armour before introducing "armour class." I only posted half the constitution table, after all, deliberately leaving off system shock and resurrection survival. So it goes. What is best left for later, I will leave, and what I can get away with saying right now, I'll say.

A rule book like the DMG has to provide the whole rule, because it's paper and binding and the players are going to refer to tables and full rule explanations in game. But a teaching manual does NOT need to adhere to that. I can feature small parts of every table throughout the work, remembering first that the "real text" is available elsewhere and that there's always an opportunity in an Appendix to give the whole. As a student in university, you do not learn everything about the Roman Empire on day one; you are given an overview first, then more context, then more, then more, until the picture slowly assembles itself over the continuation of your being given facts in a specific order. I think the failure of university is that it often never supplies a proper appendix, while the books it wants you to read don't either. To get one solid overview of the whole Roman Empire, I had to return to books that the university would never have approved, just to get one thorough consistent structure after the courses and classes I'd taken. I still find such books to be useful, not because they teach me, but because they help me keep the various events straight in my head.

Desirably, that's the work the DMG will provide to the scattered, meandering work I'm committed to creating. And knowing that it's there, and can be addressed as desired by the reader, allows me the freedom not to make OSRIC's mistake of supposedly rewriting it, and doing the rewrite badly.