Sean Tibor: All right. Hello and welcome to teaching Python. This is episode 103, all about chat, PT, and I guess the nuclear armageddon for teachers that it represents. My name is Sean Tyr. I'm a coder who teaches. Kelly Schuster-Paredes: My name's Kelly Schuster pride is, and I'm a teacher who codes. I don't know why drew Carey and your last line anyway just came out. It must be the holiday season where. Sean Tibor: The points don't matter, but the coding is real. Kelly Schuster-Paredes: That's right. Sean Tibor: Well, it's been good. Our first episode of 2023. Welcome to the new year, Kelly. Kelly Schuster-Paredes: Thank you. Happy New Year, everyone. It's been a long two days. Sean Tibor: It's true. Today is my official first day back at work, so it's a lot of dusting off things and getting ready and making plans. But I did spend a lot of time over the holiday break. Not really taking a break, but more just giving myself a mental change and working on the stuff I really wanted to work on. So it was kind of fun. How about you? Kelly Schuster-Paredes: Super cool. I traveled. I tried to zen out a little bit and not work, and I have an extra day. So now I'm like playing catch up and trying to get the house organized because it was a disaster. But it was all good. Good break. Nice. Mental vacation. Sean Tibor: Nice. Well, let's do our official wins of the week. And I think I went last week, so it's your turn, Kelly. What's your win of the week? Kelly Schuster-Paredes: Well, huge win is I finally figured out why I kept failing on my file path. It wasn't me. My file path was written. I was trying to write a file path and I was trying to do it all right in vs. Code. And I was coding it for the entire package properly because it was two main files within. So technically it was two separate projects, which I didn't understand. And they didn't want it that way. They wanted it for each folder separately, so it was not a coding error. So that was a huge win because for the life of me, I was not understanding how I was not figuring out something so simple and it was driving me nuts. And it wasn't me. Now, what it was, it wasn't wrong. Sean Tibor: It just wasn't what they wanted. Kelly Schuster-Paredes: Exactly. And I'm almost through my pandas challenge and I'm feeling very data frame happy right now. I love Jupiter notebook. It's my new favorite. Sean Tibor: It's the best. Any day when you get to use some Jupiter and Pandas, it's a good day. Kelly Schuster-Paredes: Yeah, I'm excited. So those are my wins. How about you? Sean Tibor: Well, my win has been really working on well, I guess I'm going to cheat and do two because one is all about coding. And the other thing is just something really cool. The first one is this project that I've been working on over the holiday break to give myself a mental change is event driven architectures. So it's pretty simple to grasp, but just, I guess a little more complicated to get into. But essentially the idea is that when things happen, run this code, right? So when this event happens, run this code. And what's really great about this is that your code can then trigger another event. So you can chain these events together and instead of writing really big long code that does a lot of things that's kind of brittle and fragile and prone to breakage, you write short bits of code. So when this event happens, do this one small thing and then send another event off. And then the next item or the next event triggers more code and you can have multiple bits of code triggered from the same event. So there's a lot more to it than that. But the really cool thing has been getting this all working and then also getting it working with a program called Event Catalog, which lets you write really simple markdown files. And it generates an entire site of documentation around all of the events in your code. So I just got all of that working in my job and I'm going to do some show and tell on it later this week. But it's really cool. It's letting us automate some integrations between systems and do it in a very robust sort of way. Kelly Schuster-Paredes: It's like a coded if this, then that kind of thing. If Tea thing written on a backbone or something. Sean Tibor: Exactly. So if this thing happens, then do this. If you've used any sort of system where you have some sort of trigger condition and then code that runs as a result of that trigger, you've used an event driven architecture. Kelly Schuster-Paredes: Wow. Very cool. You're so smart. Sean Tibor: I'm just trying to keep up. Kelly Schuster-Paredes: The world is so fast. You can't keep up. I give up. Sean Tibor: Well, the other Win is a local community Win. So they just opened a new train station here in my town for the high speed rail line that's been running for a few years now between Miami, Fort Lauderdale and West Palm Beach. They opened a station in my town along that route, and I think I've taken it three times now, up to West Palm Beach and down to Miami. And just the way that it changes the way you travel, I know that for many people around the world, they're going, yeah, of course, Sean, that's like how modern infrastructure works. But here in the States, we don't always have that. It's pretty amazing. The station is five or ten minutes away from my house by car. I can ride my bike there. You hop on the train and in an hour you're in Miami and you ride the whole way with WiFi and air conditioning and nice bathrooms and you don't have to think about driving and navigating. Kelly Schuster-Paredes: And you can bring your bike too, right? Sean Tibor: I can bring my bike, so I haven't done that yet. But that's next on my list. Kelly Schuster-Paredes: You're one of the few, because I know there's a lot of controversy over that going through up north ways after it leaves Boca or West Palm. But I love it too. I love public transportation. Sean Tibor: I think it's one of those things that we just have to start using more of for so many reasons, not the least of which is just driving sucks. I like riding. It's so nice. I sit and I have a cup of coffee. I could work. It's a great way to get from one place to another. Kelly Schuster-Paredes: Yeah, I agree. Let's chat about chat. Sean Tibor: Let's talk about Chat GPT. So if you haven't seen a story about Chat GPT, you've seen a story about Chat GPT that didn't call it that. So there's been a whole range of news stories that have been written, probably some of them by the tool in the last month, month and a half, about this new machine learning model called Chat GPT. It's produced by OpenAI, and I guess the best way to describe it is you get a chat window and you can type in various prompts and it will respond back to you. So you can use it to make a conversation, you can give it various instructions. And I think what everyone's kind of amazed by is how broad and how robust this model is compared to anything that's come before. So, Kelly, anything I'm missing in the description? Because I feel like that's a very dry, clinical sort of description. Have you tried it? What do you think of using it? What is it like? Kelly Schuster-Paredes: Well, I have to admit, when I first opened it up and I started online and you log in, you have to make an account. I haven't found the free option where you can just do it without logging in. So if there is out there, I'm not sure, but obviously you log in and it's a little bit scary because I'm always a person that says, don't log in. You don't know what you're doing, but get in there and log in. And you type a conversation or a question and that conversation threads keeps going if you want. And it's like opening up one of those horrible chats where they're like, how can we help you today? Pick a but instead of having to follow the menu and get the same automated response, you get a response tailored to your question. And the more you ask, the better it is. Sean Tibor: It's surprisingly good, right? Kelly Schuster-Paredes: Surprisingly. Sean Tibor: Give an example of this, a real example right now that we're going to do live, which is kind of scary. Okay, I've got the window here. I've got the window here. How would you describe yourself? Kelly Schuster-Paredes: Oh, you're talking about me or chat? Sean Tibor: I'm asking the chat bot to describe itself, see how it does. How would you describe yourself? So it's sitting here, it's thinking for a moment. It says, I am a language model. Trained by open AI. I have been designed to assist users by generating human like text based on the input provided to me. I do not have personal experiences or feelings and my knowledge and responses are based on the data that was used to train me. I am here to help users with a variety of tasks and questions. I'm always learning and improving my understanding of the language and the world. Is there anything else you would like to know about me? That's pretty cool, right? You can ask it these questions and it will give answers. I think what has puzzled people the most or fascinated them the most is that you can ask it nearly anything. So the folks over at Talk Python to me and Python Bytes were saying that they had something, whereas explain the worst case scenario of the bubble sort algorithm. But do it as a 40s gangster, right. So you can prompt it with all these things and it will explain it. What's amazing is that it will generate code. So you can have it generate Python code. That looks pretty solid, right? So I can ask it to generate the bubble sort algorithm while you're typing that. Kelly Schuster-Paredes: I actually use that to help with my files. And it was like an AHA moment when I read the response, which I already knew was a response, and I was like, it was really good. So go ahead. Sean Tibor: Yeah, so right now it's actually generating the bubble sort algorithm and it looks pretty solid, like it has a couple of nested loops and it moves things around and changes their position and everything. It even gives an explanation of how it works. Right. So it describes the algorithm in addition to writing the code. Kelly Schuster-Paredes: Yeah. And so that's where that game changer comes in for a lot of educators. And I don't think the computer science teachers have necessarily fathomed. Or maybe they have. I'm generalizing, but I didn't fathom this until, I think, yesterday when I read a tweet about someone writing an entire app and then showing how to put it online and going through the Flask library, et cetera, that I went, wow, so much for searching for the way to do something. It's kind of like the photo math app that's out there where you give it a math problem and it solves the whole problem for you right in front of your eyes. So as a computer science teacher, part of me is like, wow, that's really cool. And the other part of me is like, that's going to change the way students learn and what are we going to do about it? So it's something food for thought. Sean Tibor: Yeah. So I think what was interesting about this is that it can generate just about any text that you can imagine. So I recall seeing a previous iteration of this. I think it was GPT-3 based and maybe even GPT-2, but it was a language model at scipi 2019. So almost four years ago now, three and a half, four years ago. And they were showing examples of the text that it would generate. And the example that they gave was a news story written about the discovery of unicorns in the Andes Mountains. And it even generated, like, quotes from a researcher that it just completely made up. And they showed everyone this, and we're all just blown away at how amazing this was. But they were like, look, the state of this right now is that it is at the time, in 2019, was that that was the best one that they came up with, that was the most fascinating. But there were a bunch of others that were just low quality, very kind of badly written news articles. I think the advancement here is that everything that you see that it comes back with is plausible, right? Like, it looks, at least on the surface, like it's the right response to that. So, for example, the Python code that it came up with looks plausible to me. But until I actually test it and run it, I'm not sure that it will actually work. Right, but now nearly every response is plausible. Not just one out of 100 or one out of ten. It's like nine out of ten. So the implications for this are kind of staggering. And I think what we should probably do is talk about a few of those implications. And one of the things I want to try to avoid or sidestep is putting value judgments on it, right? Because I think it'd be really easy to decry. The sky is falling, and everything we know about education is going to implode upon itself because we now have an AI chatgpt. But there's a lot of good things that can come out of this, there are a lot of bad things that can come out of this, and we don't really know which ones are going to be good and bad yet. So maybe we'll just talk about how this could change or affect the way that we teach and we learn. More importantly, how do we learn? Rather than trying to say this is good or this is bad? Kelly Schuster-Paredes: Yeah. And I kind of wrote them in a couple of ideas in forms of questions, because for me, for Sean, sean knows a little bit more from the Sci-fi Conference, but for me, it's just what everybody else is reading out there. So we're just taking in a lot of information and trying to process it every single day. Right. So where do you want to start? I kind of want to back up. And before going into learning, I kind of think we need to touch on this question. Is Chat GPT for helping students cheat better? Kelly so this is from an article that I actually read, and it was a college student, and it was a couple of different articles that I read about where the college essay is dead. Oh, no. English Teachers Turn It in, launched an article that was like, don't worry, we've known about Chat GPT, or it's similar products in the past, and it's about the education and it's about the topics that the educators provide. But the thing that stuck with me is this idea of helping students to cheat better. So if a student puts in some code, as you said, and it generates some code that solves a problem, it could be easily submitted, just like that, and it could be right. It could be wrong in some form. But it's that student that can look at the code and look for the errors that might be happening, that allows them to be better cheaters, should we say? And I put my air quotes up there, and if you're listening to this post, live Better cheaters. And I say that with the air quotes because a lot of people would claim that we allow our students to cheat when we allow them to use Google during class challenges. So is it anything and I'm posing this as a question, not as an answer for anyone, but is this any different? Because eventually the answer is all out there, because I'm assuming Chat GPT is just scraping the web and the answer is out there. Sean Tibor: They trade it on a massive data set of information out there. I mean, you can see that the style that it takes on when it responds to coding questions comes very much from, like, blog posts, probably even some of the blog posts that people in the Python community have written, right? Or books or things that are freely available. So there's this huge language set of content that it's been trained on, and you can see that in the responses. I think this is a good point also to just pause with a disclaimer because Kelly is a teacher at a school and I work in technology. It's way too early to really make any sweeping statements. We're kind of, like, reacting here and thinking about this. So if in, like, two years you come back to us and say, like, why did you think it was okay to cheat? Just know we're talking about this in a very open minded, very thoughtful way, but not necessarily wanting to make predictive statements that reflect upon either of our employers or anything like that. We're just trying to figure this out. So let's just put that out there first. Kelly Schuster-Paredes: Everything's posed as a question, and again, these are all questions that I contact you with, and I'm just like, wow, I need to really talk this over. I need a lunchtime walk with Sean. Sean Tibor: So the policy, like, actual, how do we handle this as educators and as institutions of learning? We're not there yet. Nobody's there yet. We have to figure this out, and this is part of that process. But whatever we're saying now is definitely not policy. This is us trying to understand, didn't. Kelly Schuster-Paredes: Charles Keturine say a problem well stated as a problem half solved? So I'm throwing the questions out there. Sean Tibor: Yeah, we have to get the questions out. Yes. Here's the question, and I think this is where we get into the real semantics of what do we mean by cheating? Right. So in reality, I think most educators would agree that cheating occurs when learning does not happen, when it's a shortcut to actual learning and gaining of knowledge. Right. So if you'd use that definition, the Internet and test banks and all sorts of things have been around for a long time that have made it easy for students to circumvent their own learning. This is a much more effective way in a lot of areas of circumventing that learning. So as educators, we do not want to circumvent the learning. That being said, I think there's a role to be played for tools like this and in fact, this tool to help enhance the learning. Right. So that's really our question. How do we help students enhance their learning with tools like this and discourage or find ways to prevent them from circumventing it, from taking those shortcuts that don't actually provide learning and knowledge? Kelly Schuster-Paredes: Yeah. One of the conversations I was having on a chat with an ISTI group was I think we posted on LinkedIn. What would it be if we had Chat GPT write an essay for a particular prompt and let's do an English one? Because that's the one that's trending right now, because it is the essay. The essay is like a big thing in college, and then you have a lot of online courses. So it is right now the essay. But what if that essay was written and the class the students analyzed the essay graded on a rubric, et cetera, et cetera. So could that same process be applied to the computer science classroom? Let's deconstruct the code from Chat GPT and see if we can refactor it, see if we can rewrite it better, see if we can use one liner, even though we don't want a one liner, or maybe set it up differently or add some comments in there. So there's a lot of learning that can still happen. And I guess it kind of, for me, is this thinking of how do we overcome these challenges of this AI? That is the first of many to come, right? Sean Tibor: Right. And here's a benefit, for example, just a small one. If you were to do that same thing with a student's code, right? There's all this emotional social stigma that gets attached to it when you're deconstructing a person's code up on the screen. That's one of the reasons why you and I would always take the time to write or share examples of bad code that we had written in the past and use it to deconstruct ourselves and use that as the example of what not to do. But if you used a student's code for that. There's a stigma with that. There's a social effect, there's an emotional effect. The student may or may not care that it's their code up there, but an AI certainly doesn't, right? So you can remove all that and make it the AI code and try optimizing that. I bet you could probably even ask the AI to write badly optimized code for you. That would be fun, right? Kelly Schuster-Paredes: First of all, we have to learn to have to teach what optimization is for the kids. But that would be cool. The other thing I think one of my questions is, and I'm thinking for myself, we took this and I think you were in the course with us. We did. This ISD AI course two, three years ago that was just presented from ISTI. We're talking? ISTI international Society of Technological Educators. Sean Tibor: Technology Educators. Kelly Schuster-Paredes: Technology educators who's now combined with, I think, Ed surge or some Edutopia, one of those Eds, and I'm sorry, I don't have exactly the right one. So these are the people that are like the people that I turn to when I want to find answers. And that was three years ago. So it's this lack of knowledge about AI. We've learned a lot, we read a lot about AI, but it's not something that we all talk about all the time. We just kind of use it. So this alone, this AI, and not understanding what is happening can also be kind of fear or not. I don't want to say fear, but a question that should be asked, what is it? What do we need to know about this AI and other AIS that are forthcoming? Sean Tibor: Yeah, that's an area where if you are a technology educator in your school, one of the things that you can do is learn as much as you can about the implications of a tool or technology like this. And then summarize that, simplify it, summarize it and share it back with your teaching community for people who aren't able to necessarily synthesize all of this and bring it together. So being that kind of maybe not an expert, but at least a point of contact or someone who's thought about this and has some ability to share and maybe dispel some of the misperceptions or myths about what AI is and what it isn't. Kelly Schuster-Paredes: Do you have another one? I have more. Sean Tibor: I think my question is, is this Plagiarism? And I ask that honestly as an open ended question, because does the introduction of this tool redefine what Plagiarism is? Right? Maybe it needs some evolution or some nuance added to it. But the question of Plagiarism is not necessarily that you're copying from another person or another site. Now you're using a tool, a technology tool, to generate answers for you. Is that plagiarism? And that's where I think the policies will have to catch up with the technology. Kelly Schuster-Paredes: That's an interesting point because your response from Chat GPT, from what I've played with, depends on how one you pose the question, how much you dig into the question, how much you manipulate the question in order to get the answer that you want. That's a huge scale. It's like asking the right question and isn't that what you want students to do? So that is a great question. What constitutes plagiarism and where it is? Again, going back to the learning, how you can spend this, to what type of questions do you want to ask Chat GBT? Which is kind of scary as a teacher, for me at least, because designing lesson plans and designing new challenges that aren't Google is hard enough, but now we have design challenges that are not only not Googleable, but not easily answered by Chat GPT. So, wow, it's going to be fun. I get chills because that's going to be some fun learning for teachers, I think. And I mean fun, honestly, like, for me, that sounds like fun. A challenge accepted. Sean Tibor: As an arms race, right? This is a better tool and in a lot of ways, it's worse in other ways, but this is a better tool. As educators, we have to respond and react and come up with better policies, better procedures, better learning, better lesson plans, better environments. It's an arms race. So that if it's easy or really easy for someone to get the grade or get the reward by using chat GPT, that means we probably need to change the way that we assess, the way we teach, the way. We create those learning environments so that something like Chat GPT is useful, but doesn't replace or doesn't allow for easy trying to think of the right words here. It doesn't make it easy for students to shortcut that learning. Right? They still have to do the learning and Chat GBT maybe helps them with that, but if the environment is right, it doesn't provide that much of a benefit to them. Kelly Schuster-Paredes: Yeah, speaking of lesson plans, I saw on Twitter, I don't know if you saw that, but a teacher wrote an entire five lesson lesson plans for how to teach plate tectonics and it was fully printed out. I'm thinking of the benefit for all the teachers out there who need to present their lesson plans every two, you know, they have to have them present them, like Austin and in England, you have to have specific written lesson plans. And on the Twitter feed, it was brilliant and you just have to do some tweaking. So, again, is it plagiarism? Is it yours? Because you can't just submit the lesson plan that Chat GPT, because it might not fit your form. Maybe in time it would after you fine tune it. But how cool is that to save time for a lot of those teachers out there? I think that's, no offense, a win. I know we're not supposed to give opinions, but that's a win. Sean Tibor: That's a good example of where this is a tool, right? Like it's a tool that we can use it for. So I'll give you an example. I had to sell some old technology equipment I had at my house, so I needed to put them up on ebay. And the worst part of putting it up on ebay for me is writing all the descriptions. I can take the photos, I can come up with the titles, but like, writing the product description of what is this thing and why you should buy it and everything I can't do. So I use chat GPT. So I said, write an ebay product description for this particular piece of technology hardware. And it went and wrote a reasonable, maybe not the best, but like 80% better than I could do description of that specific product, including technical specifications for the product that it came up with from somewhere, right? And maybe it worked because it was old hardware and so the content was part of its data set, but it found all that, it put it all together and it wrote it. So as a productivity enhancer, it was like having a professional technical writer right there that could write this for me. That was a big win. Kelly Schuster-Paredes: That's a huge win. I'm thinking all the Twitter bots that are going to come out from Chat GPT, right? Sean Tibor: Yeah. Kelly Schuster-Paredes: But here's the question I have, and obviously I don't have the answer. And again, just throwing out this question is chat GPT launched. And I don't know if it's Chat GPT or if it's another person that took the open AI open source code and manipulate whatever. I don't know a lot about the background, but they opened up the Chrome extension. And this is a Chrome extension that doesn't necessarily have to work only on Chrome. It can work on DuckDuck Go, the place where you're supposed to be safe and not tracked everywhere. It can work everywhere. It can even be downloaded as a desktop extension, which will be interesting. And this is something to think about when you go to and I didn't install the extension because I was like when you open up and just says, just so you know, this can read and write everything from your browser. And I was like, no, cancel. So there's a thing that you need to think about. How does this affect cyber security or what are the risks, what are the implications? And I'm not talking about in your house, I'm talking about in the educational sector. What do we need to prepare, what do we need to plan for in order to think about this? Privacy and all that fun stuff. Copa so that was a question that I would love to hear opinions on that about from some really, I shouldn't say knowledgeable people on Chat GPT. Sean Tibor: Yeah, there's a lot to think about there. One of the other things that actually could be a positive is that somewhere. Every single request in response, every chat session is logged. So somewhere in the open AI hosted space, there's a log of every single prompt that was given and every single response that was generated by the model. So if you were to pair this with something like a blockchain or a ledger, where you could see all of those requests and responses, or it could be looked up or searched or something like that, you could potentially handle things like plagiarism by having an integration with one of those tools, like turn it in or something. The problem is that this model is portable. It doesn't belong really to open AI. In fact, I think they're publishing it and sharing it and everybody has access to it. So unless you make the model automatically write these responses somewhere, then there really is no way to control it. So on one hand, there's this logged, tracked, validated source of information that you could search through to look for items of plagiarism or for what was generated. On the other hand, you also have a log of everything that everyone has typed into the tool in one place where it entirely searchable, as well as who did it. That's a data science person's dream and a privacy nightmare. Kelly Schuster-Paredes: Exactly. Sean Tibor: So you have these double sided, like, two different ways to look at this. Yes, there's all of this generated content that's logged and searchable and could be found. And all of it is a treasure trove of information for people who are maybe looking to find out more about students or more about what our kids are looking for, things like that. So there's that double edged sword of this. Kelly Schuster-Paredes: Yeah. And it's not new. I think that's the thing that people need to keep in mind, the people that are learning about and reading about Chat GPT, this is not new. In fact, I think today or yesterday, I read the Google. I don't want to say Google one of the home, because I don't know if it's google one of the home kind of Siri Alexa kind of things was able to leak or it was recording information. And that information, I don't know if it got leaked, but it was found out that was recording. So it's not just chat. GPT. That is an AI that's out there to change the world. There were a lot of other things. In fact, I sent a post on LinkedIn. There were like ten other AIS that are already hitting the market. You've also got the one that reconstructs photos and generates all these photos out there. So Chat GPT is just taking this stage because it's pretty cool and you can have conversations with it. But all these AIS, these generative AIS are out there still. So there's all kinds of different venues. And so I keep going back to that. What do we need to do as an educator, maybe to talk about the ethics, not just plagiarism, not just cheating, but the ethics of these AIS and how not using technology for good can really cause harm. Sean Tibor: Yeah, I think it goes back into that whole digital citizenship conversation that we continue to have about as the world continues to evolve and change and technology becomes more advanced, how do we equip our students with the right social, emotional, ethical, and moral skills to be able to navigate that world in an upright manner? Right. I know this has been shared before, but I'm going to share it here because I think the analogy works really well. Steel that was manufactured before 1945, specifically before July of 1945, is very precious. And so people will gather steel from shipwrecks, they will gather it from old cars, places where the steel has been left in its original state, even if it's rusted over and they have to get to it. The reason for that, and the reason why it's July of 1945 is that July of 1945 was the Trinity Test of the Manhattan Project, where the first atomic bomb was set off. The radiation that was released into the atmosphere from that one test was enough to contaminate any new steel that has been made from that date forward. So steel before that date does not contain the radioactive markers that steel after July does. Right. So the steel beforehand is precious because it can be used for things like Geiger counters, sensitive medical equipment, things where the background radiation is really important. And someone who's an actual nuclear physicist can probably explain this way better than I can. But it became an epic, right? It was an epic date where steel before that time was new and pristine and steel after that time was contaminated. We passed a similar epic in AI years ago. I'm not exactly sure when, but where AI became good enough to create content like text, like photos, like videos, that is indistinguishable from that that was created before. Right. So we've now reached a point where you could see two different essays. One of them is written by a human, and one of them is written by a person, and you don't know by an AI, and you don't know which one was written by the AI. So now every single submission that you get is potentially contaminated. We don't know if it's generated by a human or not. So where that brings us to is this point of do we need some sort of way of fingerprinting or determining whether something was created by AI? Or do we need a new way of assessing and learning that makes the actual content and the assessment of content in that way less relevant or less sensitive to whether it was made by a human or by an AI. Kelly Schuster-Paredes: It goes back to, like, the Turing Test, right? Yeah, the Imitation Game. I read that the Alan turn in 1950 has to evaluate intelligence. Right. Can a human conversing with a human and with a computer tell which one is which, right? And that's the whole bit. And it is an interesting it is an interesting time. I mean, just interesting time in general, thinking back sorry, interesting. I'm trying to think and talk at the same time. Apparently, I can't do that today. It is an interesting time for all of us because we go back to this all the time that you and I are from the age where we didn't have cell phones, I didn't have the computers during high school. But our kids, my third grader is with Chat GPT. How cool is that? There's a world where they don't know that. Hey. An AI cannot answer. There weren't AIS to answer my questions. There wasn't a photo math to answer my math problem. There wasn't Google. What did you do before when you caught a fish and you didn't know what it was? And it's just this amazing moment in our life, and I kind of want to remember this. It's cool. Sean Tibor: Yeah. And I think this is the same way that when we talked about the use of technology and when we speak with students about how they use their phones and how they use social media and how they use all of these tools that have very positive effects but also could have negative consequences. It's important to remind students that adults are trying to figure this out, too. And in many ways, kids are better equipped to understand and apply these tools in new and interesting ways because they don't have any of the preconceived biases and notions that adults do. Right? So you and I are trying to figure out how Chat GPT works based on all of the other contexts that we've developed over the course of our lives. These students are saying, I don't have any of this context that's shaping the way I view this. How would I use it in a completely open, green field sort of new way that nobody's ever thought of before. So I think the lesson to me is we need to include students, young people, our learners, into the process of discovering and understanding how these tools can be used for good and for bad purposes. Kelly Schuster-Paredes: That poses another question that's not on my list, is when, how, where, to what extent do we introduce the Chat GPT in the classroom as, hey, we know you know it, we know you know about it, maybe some of you don't. That's cool, but here it is. How will that professional development look like with a teacher on how you can bring this into the classroom? Or how do we ask the right questions in order to make a plan? And I'm trying to think of other questions that are out there, but I just got a tweet, which is funny and it's making me laugh. You know how I had that trending tweet about what was the difference, why columns were properties and they didn't need the parentheses, not methods. And I just wanted to know the why behind things. Trying to get a metaphor, going to explain it. So Peter sent a reply from Chat GPT, and I was laughing about it. DF columns is an attribute of Pandas data frame, which is like a table of data. Attributes are values, blah blah blah. They don't have parentheses, blah blah blah. In this case, DF columns will give you a list in the column names in the data frame. Again, my question, though, which was funny on this trend, was that I wanted to know why. How can I put it into my head to explain to maybe students as why this is a dot something and not dot something, parentheses and using attributes and properties and all those words that are clearly written in documentation doesn't help that student learn about the differences. So again, going back to the party, funny as heck from Peter about that. But that answer, that response from Chat GPT still didn't answer my question, why? And maybe that's an answer for some people, like why in your opinion or for your mind, or for you to explain it to me like a twelve year old. Why does that happen? There you go. Sean Tibor: Well, I think you're touching on something important here for this conversation, which is what are the things that are uniquely human? Right? Our ability to regurgitate information doesn't make us human, but our curiosity, I would argue, makes us human. Right? The ability to formulate and ask questions that are interesting to us makes us human, I think the ability to be fascinated makes us human. Chat GPT can't be fascinated by any of the questions that it's asked, but a student can be fascinated by how something works or why it works that way, or when they get something working the way they want it to. How interesting that can be, right? So that social emotional education, I think, and that kind of what does it mean to really be human? Is at the forefront of this and differentiating between what the learning experience is like for a student versus a machine. So I think that's a really good point, because you want to know why, right? Chat GPT doesn't know why. So how do we engage that, right? How do we get learners to experience more of that curiosity, those burning questions that they really want to know about? And I would argue that the hard part as educators is getting them from the point where they don't know anything about coding to getting to the point where they're asking questions like, well, why do I need parentheses here and I don't need it over here. Kelly Schuster-Paredes: Yeah, that is pretty cool. And I want to add sorry, into that is that constant question everyone always asks through at least my 20 quarter of a century. Quarter of a century of teaching. Not quarter of a century yet. No. Yeah. What role does the teacher need or in the classroom, you know, what's going to happen to the role, the teacher. And this is not a new question, this is a constant question I've heard for the past 25 years. What is the role of the teacher and how is this affecting the role of the teacher in the classroom? And that's going to be continued conversation. Sean Tibor: Yeah. And I think that at least part of that role is going to be knowing about tools like this and knowing how to best wield them in order to enhance the learning. Right. Kelly Schuster-Paredes: I can't wait to hear more questions from our listeners. And again, just to get those questions going and start having those conversations will be really cool. Sean Tibor: Well, I've already used Chat GBT to write episode summaries for the podcast. I've used it, I don't think I actually published the chapter summaries that I came up with, but I gave it a list of chapters that were generated by another AI and then it rewrote those longer summaries into short chapters. It was pretty amazing to see what you can throw at this and see the responses. I think as you get into this and as you are looking for this, it's worth playing with. It doesn't cost anything to try it out, but you can start to recognize the styles and the patterns that it uses to be able to formulate its responses. And it's really a pretty fascinating tool. Kelly Schuster-Paredes: Yeah, I'm excited. It is pretty interesting. It's just a forefront of learning. I wonder if we can get a blog post written out of that. Maybe you can write a blog. Sean Tibor: I would actually like to write a human written blog post about it. It is a game changer in a lot of ways because the quality has gone up so high. So this is not new. But what is new is that it's very accessible and it's pretty good for what it is. Kelly Schuster-Paredes: Yeah. Well, I have a whole bunch of links that we can share in the show notes of just stuff that I've been reading that posed a lot of these questions. And I'm sure Sean has a list, a slew of stuff too, some really funny Twitter, and a mastodon post out there about Chat GPT. And it just makes me giggle and think why all that? Or what could happen with that? And then also, if you don't know, there are a lot of tutorials about how to use Chat GPT already streaming. I've got one that's about twelve minutes long. It shows you what to do and how you can formulate stuff. And there's a lot going on with this application and a lot still to be learned. Sean Tibor: Yeah, absolutely. So more to come. Definitely a longer conversation and I'm sure that the next iteration of this will be even more interesting. And I think we'll see more of this applied in other areas. There's already in Canva the tool that I use to generate all of our cover art for the podcast. There's an AI there where you can describe what you want it to draw for you, and it will make the. Kelly Schuster-Paredes: Image there goes five or jobs. Sean Tibor: Now the question is why and how do we use our human talents to create things instead of using the AI? Kelly Schuster-Paredes: Absolutely. Very cool. Sean Tibor: Very cool. All right, so I think more coming soon on the podcast. Kelly, you've been working on lining up some guests for us over the next few weeks now that we're into the new year and we're past the holidays, so keep an eye out for upcoming episodes from us. We've got some really great guests coming on the show. If you have an interesting idea that you'd like to discuss on the show with us, please send it our way. You can reach us through the website at Teaching Python FM. We're also on Twitter at Teaching Python. I think I have a set up on mastodon at teaching Python. Also, if anybody wants to reach out to us, there. Kelly, any other upcoming announcements for the show or for our community? Kelly Schuster-Paredes: Not much. Just check us out on LinkedIn if you haven't done so already, and there's a lot of posts going on through there, too. Sean Tibor: Yeah, that'd be great. I think we're only a few followers away from being able to start doing LinkedIn live broadcast as well, so let us know what you think about that. And I think that should do it for this week. Kelly Schuster-Paredes: Yep, that's it. Sean Tibor: Okay, so for Teaching Python, this is. Kelly Schuster-Paredes: Sean, and this is Kelly signing off.