AI in Critical Thinking and the Weirdness it Creates

Welcome to episode nine of the not brothers podcast, where today we're going to talk about AI and critical thinking, the ethics involved with, that, the weirdness that it can create if you have overuse patterns in AI ⁓ and how it impacts things like academia, work settings, ⁓ any of those things that are involved. So Ryan, we were kind of talking before the show about, know, there there's a lot of danger with AI. There's a lot of power, a lot of good. But there's certainly some danger associated with it. And we were talking specifically about, ⁓ my kids go to a school ⁓ and they have pretty strict parameters and boundaries around lots of things anyway, but they introduced an AI policy this year, which is pretty strict and robust. ⁓ And that AI policy is basically a look, if you, if you get caught using AI for anything associated with this, you're going to be expelled. And they had two seniors just this week that got caught ⁓ using. AI for their senior thesis, which is a year long project. They literally have the entire year to do this project. And somehow some way, I don't know the specifics, but they, you know, they got caught using AI and were expelled essentially four weeks before their graduation date, which really sucks. Right. So like, what's, what's the balance there? Like, let's, let's, let's game this out and say, like, what, how, first off, how they, would they even catch them ⁓ doing this? ⁓ Is it even something that's, ⁓ Reliable. Do we have reliable tools to be able to do that? So like walk me through that.

Yeah ⁓

a lot.

You know, my reaction now is kind of the same as it was then. I'm like, I'd love to know what they used to validate the claim or the issue. But, you know, I guess before diving into that, like I think it's probably important to kind of back up and talk about one of the things that you mentioned, which is the overuse patterns, right? I think this kind of touches on one of those potentials that exist within. I mean, it exists everywhere, but AI is like really, really blowing it out of proportion right now or highlighting it for a lot of people, which is like, you know, we're subscribers of like StrengthsFinder and that kind of thing. And the idea is like you find your strengths and you tailor things towards your strengths and you understand your weaknesses and you don't try to fix them necessarily. You just try to, you are just aware of them. And there's also this thing that exists, which is an overuse pattern, right? If you're somebody like John, who's incredibly analytical and ⁓ has learner as one of his top traits. ⁓ That's great. But if you only learn forever and never actually execute anything, that's an overuse pattern. And that now takes your biggest strength and turns it into a negative. And I think with AI, one of the things that we're seeing that's kind of interesting is like it's incredibly powerful and incredibly fast at doing a lot of things, which encourages a lot of usage. I think we talked before on ⁓ previous episode about like, you know, how in some ways they're, the prompts are tuned to be very fast, to always give you an answer. Basically they're dopamine machines, right? We're trying to give you as many dopamine hits as possible. So you keep prompting because the more you prompt, the more money that, the more dependent you become on AI and the more money the ⁓ AI creators make in the end, right? So it's this hamster wheel that we want you to stay on. ⁓ But. That can create within people and within corporations, overuse patterns where now you find yourself using AI for things that you shouldn't be. Or relying on AI for things that it's not dependable for, or that you shouldn't be, or just trusting it in general. You and I have this conversation all the time when somebody within our team is like, I had Claude come up with...

I don't know, a list of competitors or a list of customers or like any data, any data set. And my first question is always the same. Like where the fuck did it get that data from? And in some cases, like ⁓ we've dug into, you know, where it obtained the data from and it did find the legitimate source and it did obtain legitimate data. And other times it just makes ⁓ stuff up, just completely makes things up. ⁓ So this becomes the danger, right? If you're... If you're saying that somebody used, you're making, first off, if you're expelling a kid, saying they're using AI, you better for damn sure make sure that you're right. And the question is, how are you? Because, you know, when AI first kind of hit the scene, this was a big, this was a big topic. It still is a big topic of, you know, using it to produce academic papers, using it to produce, you know, I remember being in school, we had to write like five page papers on nonsense, right? That we all, well, game system, right? Had to be five pages, double-spaced, 12-point font. Well, you make it 12 and a half point font. You make it 2.25 line spacing, right? Like you tweak it just a little bit to kind of squeeze out a little bit more, at least I did. The games have always been there and you just did it enough to where like you were within, you could get away with it. AI is no different.

I don't know what you're talking about. ⁓

And I have seen examples, right? A few people on our team, or at least one, I think a couple, their wives are in academia. And I've seen a couple of papers from some kids that they forgot to take the pieces of chat GPT telling you, here's the paper, be sure to do whatever, the instructions that they... that the AI gave back were also in the paper that got turned in. So like, that's pretty blatant. ⁓ think anybody could probably sniff that out. My guess is that with the senior thesis, I would hope that the kids are not that dumb. ⁓ So at that point, like how are you detecting it? And the only way to detect it has to be through ⁓ using AI detectors. And that becomes really dangerous because AI detectors have been proven to be wrong. ⁓ I think OpenAI used to have their own and they killed it based on their reason for killing it was that it just wasn't accurate enough. So that's like the people who are building the shit. But I know that Turnitin, which is one of the really big ⁓ platforms that schools are using, they have an AI detection thing.

Mm-hmm.

It's been proven like multiple times to be wrong. There's some schools that have talked about why they've disabled it even ⁓ if you search around online about it. ⁓ Cause you can pass like existing works in. think one of the funnier ones was like, if you pass in the U S constitution, it says that it's, it was AI generated. If you pass in passages of the Bible, it says that those are AI generated. If you pass it pass in, ⁓ the Declaration of Independence, it says that portions of it are AI, like these are things that ⁓ well predate AI being even a thought, let alone, so it calls into question, right, like how the hell is it detecting that it's AI generated? ⁓ It goes even further that like, if you're a good writer, you have a higher probability of being flagged as AI. mean, hell, on social media now, if you use dashes, you're automatically. assumed to be AI, even though dashes have existed for a long time and are an actual ⁓ piece of like how you should write. ⁓

I use dashes all the time. Most of my stuff will probably get flagged.

Yeah. ⁓ Or an interesting thing, just kind of like we talked about this, you know, sort of loosely yesterday. So I was digging into it. ⁓ English as a second language. ⁓ People have a much higher probability of being flagged. So content that they create and write from scratch are more likely to be flagged as AI, probably because they write it properly. Right. Us as native English speakers.

Yep.

maybe have a little bit of a different, we didn't, you know, we learned the textbook way, but like, we all talk a little bit differently. So we, you know, don't necessarily write exactly properly. But if you're, I'm just spitballing and assuming that that would be the reason that, you know, somebody who, wasn't their primary language would be more likely to be flagged than not, because they may write something that is,

Sort of.

is proper English, but isn't the way that you and I would say it. And it isn't the way that you and I would say it because they didn't grow up speaking English ⁓ and kind of learn some of those nuances of language.

One of the, one of the things that I think about in thinking about that situation of student expulsion is like, what, what's the boundary? Like, what is the, what is the, the gray area between critical thinking ⁓ and plagiarism? Yeah. Between, you know, critical thinking, plagiarism is this, are these my thoughts that I get here because of my thoughts, but I'm using it as a tool to help me accelerate my thoughts. Like there's, there's weirdness here that has existed from.

Like how much use is too much use?

the days of the calculator and then translated to the days of the internet with, with, you know, search and then sort of translated to even social media becoming that source of source of truth for things that you could find quickly. And then now it's evolved into AI and all of those things, all that, like, you know, those paradigm moves over the years are similar, but this one is accelerated in a way that is ⁓ very difficult to ⁓ kind of put your finger on the pulse of and say like, what's right and what's wrong.

Yeah, I could go for hours on this one because I have lots of thoughts on academia and just, you know, a lot of it is pulling on my own experiences in school because I was one of the good student, good grades hated school, right? Like it was, it didn't feel that it was doing what it should have been for me. And I have lots of reasons for that. But the. know, as I think back to like. just staying on the topic of written papers. One of the things that was in evolution when I was in high school was Wikipedia. And we had the same problem with Wikipedia. ⁓ I remember ⁓ for years, and it's funny, it's even laughable now to think, we ⁓ were scolded and you would be failed for using Wikipedia as a source. You had to use an encyclopedia or some sort of like... real sort, like don't even fucking exist anymore. I'm sure they do. So technically, maybe. But like Wikipedia ⁓ is a lot more accepted as a source, I would say, now. And like if you don't think, like the argument back then was like, well anybody can put anything on Wikipedia. Go try it. Go ⁓ try to put anything, go try to put something factually accurate about yourself or your own company on Wikipedia. I did it. I tried to just like, I was like, we're an advertising agency in Cincinnati. We should be on ⁓ this list and we could be like years ago. And it was before I could refresh the page. This is pre-AI, pre-automatic. Before I could refresh the page, one of the moderators had flagged that shit and kicked it out and been like, no, needs to be by an independent person. was like, holy shit. These people are extreme. ⁓ So like that by itself, and like that was always a question back then, right? We used to have, there used to be like services, or I'm sure there still are, I haven't written a paper in a long time, obviously, but you could run it through and it would tell you like, you know, if you were getting flagged for plagiarism. And like that was always even a challenge then of there's only so many ways to say stuff. Right, if you're describing like, you know, I don't know.

the life story of Steve Jobs or like the history of Apple, right? There are only so many ways to say how Apple started ⁓ or like these very formative things. So you wind up with like these plagiarization flags saying that you, know, a sentence is plagiarized when it's like, it's just ⁓ factual. Like Steve Jobs was born and blank, right?

And I think maybe that's the, that's the line here, right? I was talking to my son about this, who obviously is impacted by this because he's going to school there and has, you know, he's, he's at risk. And so he's like, I don't, ⁓ how do we navigate this? I was like, look, ⁓ the bottom line is this from my standpoint, the ethics involved in plagiarism is no different today than what it was pre AI. It's exactly the same. If you blatantly copy and paste something.

He's at risk.

Or in our day, go to an encyclopedia and literally type out exactly what the encyclopedia says. That is plagiarism. That is the definition of plagiarism. If it comes from your own brain and you're using the research that you found and you're using the new tool that is AI to accelerate your research and accelerate your thinking, but it's still your own words and you're not blatantly copying and pasting whatever AI spit out into your paper. That's not plagiarism. Full stop.

I can't of words for word.

If you are copying and pasting those things from AI, even though they came from who like come from came from you in terms of like your research or whatever, ⁓ that could still be considered plagiarism because you're copying and pasting words that you did not physically write. You prompt engineered your way towards something that's an output.

and this is where the legal world would disagree with you to an extent. ⁓ We've seen this in music a lot. ⁓ There are only so many chords that exist. It is possible to create a new melody only to find that that melody is actually already owned by somebody. ⁓

And that's fine, right? ⁓ if ⁓ you accidentally came across, like this is the ethical conversation, right? So the ethical conversation is self-policing. How can I self-police my, and make sure that I'm not ⁓ doing something incorrect, right? So if the school policy is that you should not use AI, and what it really means by that is don't plagiarize, like ⁓ in general.

But how do know?

Yeah, but ⁓ if I say, hey, I didn't plagiarize, I wrote this myself and you run it through some bullshit tool or something or however you detect it, you go find online where somebody wrote the exact same thing that I wrote. It just so happens we wrote the same thing in the same way or close enough. How do you prove that ⁓ I didn't write that? And it just happens to be the same.

I that comes down to the ethical dilemma, right? Are you a trustworthy source? ⁓ Can I believe in what you're saying? Is it plausible enough that that's a real thing? Just in your example of music, right? So yeah, ⁓ my bad. Like I created a melody that was almost identical to this other melody and so therefore it is legally protected and I have to move on and use some other melody, right? If it's copyrighted and there's a bunch of other legal stuff that goes into all that.

Yeah, usually you don't find that out until you've made like a hit song with it and it's making a bunch of money. And then now you have to give all of ⁓ the rights to ⁓ whoever you accidentally sampled ⁓ a melody. Like the patent trolls of the music industry are like the armpit of the fucking world. ⁓

But in ⁓ this.

⁓ That's a whole different podcast. ⁓

But I mean, that's the challenge, right? Is like, have legal precedents for how some of that gets proved. And obviously we're not talking about that here. Here we're talking about something that honestly is a little more dangerous, right? Like at least that has to go through a real process with real judgment, with real rules, with real precedent, with, you know, all of these layers of things. I'm guessing that like in the classroom or in any sort of academic setting, that could just be a teacher. Right? A teacher just decides if you pass or fail based on if they think you did something wrong or not, or whether some tool, like they ran it through some online, like I guarantee there's teachers out there who run papers and things through free online AI detectors that do something. And it probably made decisions with those and ⁓ that have negatively impacted students not realizing.

Yeah, and.

Right? Because it's so good at providing such a confident answer and telling you, it'll not only tell you it is fake, but it'll tell you why. It'll probably tell you, like, I could probably feed it one of my papers from college. If I could dig it up. think I still have some somewhere and tell it, I think this student cheated and it would, used AI to generate this. And it would probably tell me that it did. I wish I knew where those papers were right now. I know I have a backup of like all of my stuff back in college.

think I have mine. I might do that just for fun. That would be fun to do. ⁓ But like the real conversation ⁓ here, ⁓ just to put it on a record, ⁓ yeah, well, I don't.

These are all the philosophical, like, ⁓ of the scenario and like, could it, would it, should it.

I don't think the headmaster at this school, because the process at this school is teacher probably flagged it. It ⁓ escalated to the headmaster headmaster probably dug in and did more thorough investigation ⁓ and ultimately made a decision and probably even got a committee of some sort to evaluate whether, you know, whether this is the right decision. Like this, this is, I don't think this is like the school out to get these kids. really don't. Yeah, I really don't. I know they're going to, you know, they did a thorough process and whatnot.

re-jerk. Yeah.

But ⁓ it prompted this ethical question of like, ⁓ AI is a wonderful tool and ⁓ how and when should we use this wonderful tool and not demonize it in the way that Wikipedia was demonized when we were ⁓ in school. ⁓ Yeah.

Or one further is calculators. You've heard me complain about this forever, right? I remember there were ⁓ two chief complaints I always had in school. Well, three, they made us write things by hand and bitched at me about having terrible handwriting. ⁓ calculators, ⁓ we were not allowed to use calculators in the classroom for years. ⁓ And it was just absurd to me, right? I was like, we have a... We have invented as humans, we've invented a tool to solve a problem. And you're saying we can't use the tool because we need to know how to do it the old fashioned way, right? And like maybe to an extent there's some value in understanding some of that, but like I've never been in a position still to this day where I need to calculate the cosine of something by hand.

Right. ⁓ And I get it. This is ⁓ the conversation we've had about academia too, right? There's a balance between understanding the critical thinking side of how to solve problems. And the best way to train that in your brain is to understand how formulas work. And so you sort of, you understand now how formulas work. And so you apply that methodology, not that exact formula, but that methodology to other things in life, right? Like,

This just not happened. ⁓

that level of critical thinking is important as a foundational element that you build up, but where's the boundary? And like, you know, now, okay, I understand how formulas work. And now so let me use the freaking calculator. Let me use the tools. It's like, it's like a construction guy that isn't allowed to use a hammer. ⁓

⁓ We ⁓ had plenty of when I was in ⁓ AP Calc, we had a lot of those conversations. ⁓ I was ⁓ friendly with my teachers back then, so I had no problem debating anybody. ⁓ But my argument was always like, should just give us the formulas. Because if you're giving me a test, and the test is going to require that I use those formulas. Are you testing whether I've memorized the formulas? In which case this is not a math test, this is a memorization test. Or are you testing that I know which formula to apply and therefore I understand the topic at hand? Because if this isn't a memorization test, you should give me the formulas. Because if in a real setting, right now, right here, if we were like, we need to calculate the surface area of a triangle and we have two of the sides or two angles or something, be like, hmm, okay. There's a, I actually do remember that one, but like there's a formula for that one. We can look that formula up and we would just Google that formula real quick. And then we would apply the formula. The important part is we know A, that there's a formula to do it, B, what formula it is, and C, how to apply that formula to achieve whatever goal we're trying to achieve.

When these are like in our little example here, these are high school kids, right? So presumably they're they've already gone through this process of laying a foundation for critical thinking and how to problem solve and how to do all these things. And now they have these wonderful tools at their disposal to be able to get to a place where they can write this senior thesis. So the question becomes like, why demonize ⁓ something that is literally a tool. So long as it's not straight up plagiarization, right? Like if you're using it AI in this case, if you're using AI for research or you're using it for, um, you know, idea generation, or you're using it for checking your own work or any other application that would be a tool, quote unquote, not the primary creator of, right? Like that's, that's, think the balance between this, right? Like.

Right, here's the real question, right? Like flipping it on its head, like we're still operating on the assumption that like you shouldn't produce, you shouldn't use AI and obviously they, you know. they created some set of rules of engagement going into this process. And if you break the rules, that's a different story, right? Regardless of if I agree with the rules or not. But kind of setting that aside for a minute.

Should it matter? I ⁓ tell our team all the time, I tell everybody. ⁓ You're still responsible for whatever your AI produces. Using the excuse, I just used AI, that's not a valid excuse that's acceptable to me for anything. It's not an excuse for anything to be shitty that you put out. Ultimately, if you ship it, it's yours. I don't care if you hand wrote it from scratch. I don't care if you used AI to generate it and didn't look at any of the results. I don't care if you use AI to look at it and then you review the results. You're still responsible because in my head you produced that. It doesn't really matter how you achieve the goal. You still got to the goal. And in some cases that's using judgment, right? I have plenty of things that are very small, very low risk, very like square peg, square hole that I'll throw AI at. I mean like just make this and ship it. I don't even look at, I'm not even gonna look at it. Maybe I'll go look at the code later or look at the results later. just to double check, but I know that like, this is a very simple thing. And there are much more complex things that I know that it would completely fuck up and I won't ship. ⁓ And there are some things that you just don't turn it really so at all. So the argument is like, if, ⁓ for your senior thesis, right? Like, or anything ⁓ for that matter, but what's the goal of it? Is the goal to... write a paper? Is that really what we're testing here? Is that really what like the whole of my senior year hinges on is the ability to write a four-page paper? That seems kind of stupid. The whole point of it at least, yeah whatever, but like when I was in school ⁓ my school did kind of a

It's like a 30 page paper.

weird thing because we had the small schools ⁓ experimental thing that ⁓ they did for period they don't do anymore. ⁓ We didn't have a senior thesis. had a exit project or something. I don't know what hell it was, but it was a year long thing that we had to do that of course everybody crammed in at the end. But you know, the point of it was to like choose a project and like get some, some level of experience with something. And I honestly don't remember. I remember what I did, but I don't remember what the whole perp, honestly don't, it outlines to me that like the purpose was kind of lost in that like it was a, it was a set of boxes I had to check and it was responsible for 30 % of my grade as a senior. And ⁓ all I needed to know is how to check those boxes and not necessarily what it was trying to accomplish for me. ⁓

Yeah, in this particular case, these senior theses are ⁓ designed to prove that you ⁓ can be very persuasive in whatever your topic is that you've chosen. And usually they're kind of controversial, purposely. ⁓ so that's exactly 100 % what it's designed to do is show that you can be persuasive in arguing whatever your point of view is on whatever subject it is. And it goes deep into...

Yeah, I'm gonna need 30 pages for that. can be persuasive in about 10 sentences. ⁓ And persuasive and argumentative. ⁓

⁓ Well, ⁓ so this paper is. It has to be extremely well researched and you actually have to stand up in front of the whole school and give a speech associated with it. So it's it's it's multifaceted and multi layered, right? It's not just like writing this simple four page paper in this case. ⁓ So like.

Sure. But that becomes a question, right? If I'm able to use AI intelligently to be, you know, in that case, I envision that sort of like a litigator, right? My job as to litigator or to debate my point. And if I'm able to use AI at my instructions to effectively be my paralegal and go do research and gather information and produce that information and help me formulate my arguments and ⁓ mold that into something that whether it writes it or I write it, it's written with my input and guidance, I would argue that should be allowed. I would argue there, and like, if it sucks, right? If it's a shitty argument or it's a wrong argument, it's still my fucking argument. So I should be judged ⁓ on that, but it was still my work and my thought and my inputs that led to the output. I'm just using a vehicle to do it.

True.

And let alone some of the more technical sides of like how you can instruct AI differently. ⁓ How smart were you in your process? Did you make skills so that it would follow a certain workflow and process as part of its research? Did you give it ⁓ a tonal voice that you specifically wanted to ⁓ steer clear of or to lean into because it's more your voice? ⁓ There's lots of technical ways that you can enhance how you're using that AI as well. ⁓ I think to your point is ⁓ debatable as to whether that's actually my creation ⁓ or it's AI's creation. Is ⁓ it created by me by way of tools? In the same way that if I'm building anything by hand, if because I use a hammer or a screwdriver or a drill, is it the product of the hammer screwdriver or drill or is it my product?

And ultimately that's a big enough thing that like, can like, AI is not gonna one shot that. Like, I think there's probably a vision and a fear that somebody is just gonna be like, produce a well-cited argument about XYZ that's 30 pages long and, you know, it just goes and does it. And like, there's no thought and there's no work necessarily that goes into that. Because of what we've talked about before, that's a task that like AI is specifically trained against. ⁓ unless you're using like more agentic approaches and you've built ⁓ architectures to solve for that. like, ⁓ generally speaking, it doesn't want to write a 30 page paper. It also doesn't want to go do a bunch of research. It just wants to give you a quick answer that it can eject on the screen in less than 30 seconds because you lose the attention otherwise. So. that by itself, like the work and the output by itself becomes a big part. And especially if it's something that has to be presented, I'm assuming there's some amount of the presentation and the argument itself that becomes kind of judged in the process. So ⁓ I don't think it's a fair question, right? At what point does AI go from ⁓ being classed as sort of like ⁓ a cheater tool, right? In the same way. that Wikipedia was in the same way, because Wikipedia was just a shortcut for not having to go to encyclopedias and do all the stupid bullshit that you had to do to look up stuff in encyclopedias, or the same way that calculators were, or, you know, there are other examples I'm failing to, like I've seen since then, but that I'm not thinking of, like.

Well, even handwritten papers versus typing, Like typing a paper, you can do spell check and grammar check, whereas handwriting a paper, you cannot. And so is even that tool something that we shouldn't use was a debated subject for our generation.

Yeah, that's good example.

Yeah, so at what point does it go from being something like that, right? Spell check is a good example. I remember that. It was like ⁓ you're not allowed to use spell. If you type it, you can't use spelling and grammar check. And I was like, well, how are you gonna know? But also, there again, you're penalizing somebody for having good grammar. So you would intentionally just mess something up. So you're like, whoops, I gave you something to grade.

I remember, I remember being in school and, ⁓ you know, my ⁓ seventh grade history teacher, I was the very first person in the history of his, ⁓ his teaching experience to turn in a paper on a floppy disk. I asked him ahead of time and like, Hey, can I type this? And he's like, sure. Nobody's ever done that before. ⁓ So I turned it in and I actually met with him. He then became the principal of the school many years later and he's like, you know, I always remember you because of that moment. ⁓ Like, what am I supposed to do with this floppy disk? How do I grade this? ⁓

you

Oh man, I got stories from those days, because that was fun. I started bringing a laptop to school when I was in like...

maybe ninth grade, like into middle school or beginning of high school. It was early. It was before the school, the whole school had computers. So I was doing the same thing. I would type papers, I would put them on a disc, I would turn them in on a disc, you'd bring them in. The disc would just not work sometimes.

So, you know, there were, ⁓ it was fun because I mean, that was, that was when, again, it's always a game, right? We were playing the game of like, how I can outpace the pace that the school is going and like push that forward. And ⁓ there were certain advantages that were able to be afforded in that process. I think that's the fear right now, right? ⁓ And I think this is the question, like even for like outside of academia, it's for work, it's for other things. Like at what point does, and I think if you were to take this to organizations and ask the same question, you get a polar opposite result, right? An organization, any business is gonna say, well, hell yeah, I want my people to use AI to produce that thing and ⁓ produce a good result in a faster way, amount of time. ⁓ I mean, these things are impressive. They're able to do some really cool stuff. ⁓ and it doesn't necessarily take away, in some cases it does, in some cases it takes away some of the thought, but in my experience, there's still a lot that AI is not good at. There's a lot that it is, right? But when it comes to the critical thinking aspects, when it comes to the creativity, when it comes to all the pieces that makes humans human, it makes good ideas good ideas, it's just not there. And I don't know that it ever will be. because I think that the current generation of LLMs are ⁓ really sophisticated autocorrect, and at that point, ⁓ you're always trying to create something that's already been created. And the things that are truly foundational and truly, ⁓ the human aspect that kind of moves the needle on anything is always, ⁓ typically, something that hasn't really been done before. So, There's no pattern to be matched to, which means that by very definition, AI would never think to try it.

And I think that's the, ⁓ unless encouraged by a human, right? I could give it the input and be like, Hey, I use the Budweiser frogs as a good example, right? That was a piece of marketing that existed years ago, ⁓ that nobody would have fucking thought of like, Hey, let's sell beer using talking frogs and it's going to go viral. This is viral before virality was a thing. Like, ⁓ now I could probably, I could do that.

Yeah.

I could go to an AI and say, we're going to do this. How do we execute it? And it could help me riff on that, but the idea, the core of it, the piece that really matters, ⁓ the go-to-market strategy and all the other stuff that comes after that, some of that stuff is just prescriptive. It's just things you do. But the core of the idea, the critical thinking element, the piece that actually matters, and if you're grading somebody on their work, that was mine.

What? ⁓

Yep. And so that begs the question of like, what, and I don't think we have a good answer for it, because especially in something like academia, which is where we started the conversation.

I think the answer is gonna be different for every person, right? There's some people who gonna be like, absolutely not, ⁓ you shouldn't use it. You're have people like me who are like, absolutely you should, fuck you. Like use all tools at your disposal at all times. Not using tools does not make you better somehow. Like it really doesn't seem capable.

For ⁓ an adult, I agree with you. For, I think the debate becomes at what point for ⁓ students, ⁓ kind of where we started this conversation, ⁓ does that become, it's like movies, right? ⁓ When are you mature enough to see ⁓ or consume a certain type of content? And when are you mature enough to use a certain tool set

Hmm.

to its maximum capacity is kind of the way I think about it, right? It should be gated in some way, but I don't know that we have good ways to really figure that out right now.

I agree, it's different than how you describe it there, because I don't think it's like, you need to be mature enough to use this, although, you know, maybe, especially with like lobotomized models and things. ⁓ But ⁓ I do think you strike an interesting point in that like, so one of the things that we've found ⁓ with using AI, or I've found with using AI and other people have talked about too, ⁓

You get good at accomplishing things, but you don't learn shit. So, ⁓ you know, I could write, I already know Ruby, right? I can write Ruby by hand, I can write it from memory, and I can use AI to produce a ton of Ruby code. And that really doesn't kinda matter, because I still ⁓ know how to do it, I can review the code, I can write it, I can fix code by hand ⁓ if it makes a mistake, whatever.

Flipping over, you know, I've created some CLIs for a couple of tools that we're working on right now, and they're all written in Go, because it's just a good language for CLIs, there's some good libraries for it, it makes sense. I don't know how to write Go. I don't care to learn how to write Go. So I've had, have just let it produce it, and then sort of reviewed from as much as I know about just good general development and coding practices. But when it comes to the language itself, I couldn't write the shit. And that's okay because what I'm trying to accomplish there and what I'm testing is like, I architect the CLI tool that works? And it's just a thin wrapper around my API anyways, not can I write go? And I think that becomes the question is like, what are we testing? Right? Back to the same thing about the formulas and math. What are you testing? Cause I think that changes the equation. If you're testing whether I know how to build a house, give me power tools.

Mm-hmm.

If you're ⁓ testing whether I know intricate joinery, sure, yeah, give me a couple of hand tools. ⁓ What we're testing in those situations are fundamentally different. so the restrictions that should be placed on it are fundamentally different. So in the example of the ⁓ one at Mars Hill, ⁓ if we, there's Docs for kids, ⁓ if we, ⁓

I'll take that out. ⁓

If what we're testing is the ability for a student to make a comprehensive argument both in written and verbal presentation, who cares?

I think it's a fair, it's a fair point. And ⁓ if the, if the rubric ⁓ is around the idea that is, ⁓ are you, are you able to construct grammatically correct sentences and are you able to do all those different things? Maybe that changes the conversation. So it all. ⁓

Yeah, but if that's the case, in like today, in 2026, if we're grading like spelling and grammar, like, come on, the ship has sailed on that one.

⁓ My argument is different. think most schools produce people, produce students that know how to write in social media. And that's a fundamental flaw of future society.

No, what I'm saying is like, everybody uses spell check. Everybody uses grammar check. Everybody has Grammarly. Like that ship has kind of failed. ⁓

⁓ that's a different thing, right? But you still have to have to know like literary principles to write well. You can produce a whole bunch of stuff and it'd be grammatically sort of correct, but it still sound like poo poo.

Yeah, ⁓ I'm in there again, right? Like if I write a paragraph that kind of describes stuff, but I'm sort of an idiot and I write it in a stupid way, and then I ask Grammarly or AI in general, because like Grammarly will retool ⁓ whole sentence structures or whole paragraphs if I use those tools to just retool them. But again, I write it from scratch the whole way, start to finish all 30 pages. And I utilize Grammarly or even AI to say, me reword this to make this sound better. Is that violating the rules? Or going a step further, Microsoft has crammed Copilot down our fucking throats. So spell check and grammar check inside of Word probably have some AI component baked into it. The Google Docs is doing the same thing. So ⁓ if I'm writing my paper, right, if I'm writing a paper in Google Docs today, it will suggest autocompletes. And if I accept those, Did I break the AI rules and not realize it?

Right. ⁓ And taking that a step further, I do that on accident by hitting tab and it automatically accepting it? I'm like, I... ⁓

See ya.

Or just doing like spell check, grammar check, it's like reword this, do this, change these. I'm like, yeah, these all make sense. ⁓

Yep. Yeah. mean, those are all fair questions. Where's the line? ⁓ How do you, ⁓ especially in, like you brought up another interesting point, this isn't just going out to like a standalone LLM, like a CLOD or a GPT or whatever, and asking it questions. Now these things are baked into fundamental tools that we have to use to ⁓ write or create anything. And so ⁓ if you're using the tools that are in those, ⁓

Right.

like Word or Google Docs, and Gemini and GPT are just part of what it is, where's the line between ⁓ what I'm making and what the tool is helping me do?

It's a real challenge. ⁓ I don't have a good answer for it. And I don't think, especially academia where things are typically move a little bit slower intentionally, I don't think they're gonna like any of answers because I do think that like the commercial impact and the commercial benefit of this is ⁓ so high that companies are baking it in to just the root level of everything. But I don't think it's gonna be avoidable, right? I do think it is potentially viable for me to go sit down and write something just completely from scratch from my head and accidentally use AI just because it's baked into the tools that I'm already using and I don't realize it. So when you have sort of these like overly restrictive AI policies, I just don't think like in 2026 saying like zero AI use at all whatsoever. It was like saying don't accidentally breathe fucking air.

No.

is ⁓ I definitely don't envy academia with what they're, because I get what they're trying to do, right? They're trying to make sure that students learn.

Yeah, I get the purpose. get the general sense and the fear. The fear is you're just gonna use this tool to ⁓ subvert learning the topic, and that's a real challenge. ⁓ I just described how you can do it. I just described how I on an everyday basis, I've produced three or four really comprehensive CLI libraries in addition to converting a Python screensaver for a marshy to Go. and it's almost completely done. And I never wrote one line of it, and I couldn't tell you some of the most basic pieces of Like writing, I mean some of it's pretty basic, but like the syntax and those sorts things, just don't fucking know. Because I don't need to, and I didn't learn it. Because when you're using AI to produce things, you're not learning. And that is a whole other fallacy that we could go. go down the path of, and I think that is the fear, right? If you just use AI to generate the results, you're not learning. But if you have learned the ideas and the principles and you use AI to generate the results, that's a different story ⁓ and potentially yields fundamentally different results than I think is. ⁓ is a real challenge.

Yeah, I don't envy, I don't envy people in administration or in teachers in academia right now, because it is, it is very tricky. Um, and it's not going to get any easier. It's going to get, it's going to get harder and harder as time goes on and as, as, as these tools become more ubiquitous and they already are, and as they get better and better and, adoption continues to, to escalate. Like it's just like a, monthly basis, we're probably seeing, you know, an exponential amount more adoption across not just students and academia, across everyone on planet Earth. So, all right, let's bring this to a close. So we talked about critical thinking, we talked about academia, we talked about our own experiences in these things, and how they're kind of related to things like calculators, Wikipedia, things that have been demonized over time, and how AI is kind of the new the ⁓ demon, ⁓ if you will. ⁓ till next time.

next time. Stay in school kids.

⁓ Stay in school.

Those two are because they're going to fail. ⁓