Inside Tech Comm with Zohra Mutabanna

S4E4 AI and Tech Comm - Using ChatGPT Smartly to Enhance Your Career with Shaughn Kern

Zohra Mutabanna Season 4 Episode 4

Shaughn Kern, my nerdy friend, graces my latest episode. As a technical writer and a code geek, Shaughn has been experimenting with ChatGPT to simplify and automate mundane parts of his day-to-day. In this heart-to-heart conversation, Shaughn shares some great examples that any technical writer can use to move on from the mundane and get creative with ChatGPT's assistance. If you are a technical writer wondering how you can leverage ChatGPT, or for that matter any AI tool, this episode is a must-listen for you.

Highlights of the conversation you don't want to miss:

  • Examples of how Shaughn uses ChatGPT to simplify the grind and get creative with his research.
  • What kind of prompts did Shaughn give with his content to benefit from ChatGPT feedback?
  • How to use ChatGPT to track legacy content or convert documents from one format to another?

Guest Bio

Shaughn Kern is a technical writer with a background in technical communication, UX/UI design, and presentations. During his career, Shaughn has mostly worked for small research institutes and software companies where he was the only technical writer. This sparked his interest in writing scripts and utilities to automate mundane, complex, or error-prone tasks. He spends several hours a week researching or building tools that make technical communication more efficient.

In addition to his technical communication experience, Shaughn has worked as an adjunct lecturer, writing center coach, and Brazilian jiujitsu instructor. He holds an M.S. in Rhetoric and Technical Communication from Michigan Technological University, where he also earned a B.S. in Scientific and Technical Communication with a concentration in Computer Engineering. His non-work time is devoted to parenting, cooking, and martial arts.

Credits

  • Intro and outro music - Az
  • Audio engineer - RJ Basilio
Zohra Mubeena-Mutabanna:

Hello listeners. Welcome to Inside Tech Comm with your host Zohra Mutabanna. In season four, I hope to bring to you different perspectives and interests that intersect with our field. Let's get started, Hello friends, welcome to another episode of Inside Tech Comm With Zohra Mutabanna. Today, I'm having my best friend, Sean Cohen, on the show, and I pick his brains all the time. He and I work together. And he is a nerd. If I'm using it in the right context, he is super nerdy. And he and I were talking about ChatGPT. What else, right? And he started sharing all these cool things with me and I'm like, Dude, you have to get on my show. And I'm honored to have my dear friend Shaughn Kern with me here today. Hey, Shaughn, how you doing?

Shaughn Kern:

I'm doing good Zohra. And thank you so much for that introduction. I do really take the term nerd as a badge of endearment, something that I strive to earn and wear proudly. So thank you for that.

Zohra Mubeena-Mutabanna:

Thank you. I'm so honored to have you. So Shaughn, tell us a little about yourself.

Shaughn Kern:

I have a bachelor's and Master's of Science in Technical Communication with most of my focus being on software documentation, UX and slideshow design. I am a bit of a computer nerd, and the word nerd. So it's I find to be a fun intersection. And in doing so, a lot of times I'm looking for new tools and new ways to reduce my own workload and make my work more accurate, more targeted.

Zohra Mubeena-Mutabanna:

And we will do a deep dive into that. But tell us, you and I worked on some really interesting projects. And one of the reasons why I love collaborating with you is you're such a, like a coding geek. Or at least from where I'm sitting and watching you do stuff. So tell us how your background in tech comm has kind of led to your passion for coding. And what have you done to become the scripter that you are today?

Shaughn Kern:

Yeah, well, I don't know if I've earned the scripting or coding geek title quite yet. But I'm certainly an amateur coder, particularly with regards to finding things that make my life as a technical communicator, easier. And it's funny because my first real life tech comm gig was about 15 years ago, at a research institute, they had built an internal toolkit of something like 1000 scripts in various languages. And the only documentation they had was the command line hope statements. My job was move to all that content, from the command line help statements into a confluence wiki, and then take that same content and create a user guide in Microsoft Word. I think I had three months for this project, maybe four. Anyways, it was an absolutely insurmountable task to do manually, especially within that timeframe. Developers like how do I even begin here, and they're like, you're gonna have to use regular expressions, and you have to use Python, and you have to use some bash. And I'm like, I think you hired the wrong person, like, but I'm going to try and figure this out just the same. I did it. It was, you know, week after week of figuring out how to use Python, and bash and macros and Microsoft Word. And all these more technical elements that are available to tech writers, but maybe not taught in a traditional curriculum. They're not usually not listed on any sort of application to a tech writing position. But what stuck with me the most is once you know that this technology is out there that you have these tools to help, you can vastly simplify both removing bottlenecks from your job functions and automating more mundane parts of your work, you can reapply those. And to this day, I still use a lot of the same tricks. I learned from that, that three month four month gig. And I use that in the same projects as at Blackbaud, as I did with previous companies, and it's been working great.

Zohra Mubeena-Mutabanna:

I mean, definitely, you know, as I've shadowed you, I've seen how you kind of have applied some of the tricks that you've learned before, to new role. The reason I asked this question, Shaughn is one, I believe I personally believe that as we grow in our careers, we have to pivot, right constantly to keep our skills to stay current and to stay, I guess, to stay relevant. But the coding thing can sometimes be intimidating for many technical writers. And, right, the purpose and asking this question to you is, what advice would you give to a technical writer who has absolutely no experience in encoding? What could be the simplest way to start somewhere.

Shaughn Kern:

I would say just learning regular expressions and learning how to open a file in Python and find a chunk of text, you know, a string that's maybe like our most recent collaborative project, the alt text script, where you and I set the script that basically looks through all the HTML files and says, Okay, find all instances of an image tag, and give me a list of the ones that do not have, you know, an alt equals quotation marks. And so it involves a living doing regular expressions, while a little bit of knowing how to open a file and put over into a new file. From there, you start to realize, okay, is this just a matter of cracking open a file, finding a type of string, maybe changing that string, and then moving any content that you want into a, basically an output file

Zohra Mubeena-Mutabanna:

or files? That's a simple enough example that is not insurmountable. That is not intimidating? And of course, I guess Google is your best friend, right?

Shaughn Kern:

No, I think at this point, ChatGPT. Thank you. Yeah. And I'm not pandering to the subject of this show. Because, you know, I spent a ton of time on Stack Overflow, just trying to find snippets of code. And it's like, Okay, does this work? I think, as writers, we have Google for a very long time, I think software engineers and developers have had Stack Overflow that's been like one of their go to resources. And so something like that, just the basic of saying, Okay, we'll find every instance of passive voice and put that line of code in the file into a text document, you know, that can be helpful, that can be helpful in symbol for just exploring how to write things in Python, how to open files. But you can get far enough into that, where we can do things like say, Okay, I want you to convert all of our Freemaker, or sorry, all of our Madcap flare documentation into markdown, or all of our Adobe FrameMaker did a documentation into lightweight data. So once you're able to crack open files and manipulate content, and put them out in whatever file you want, your ability to migrate legacy content formats into a new style, or clean up legacy content becomes something where, again, if you had three months to do it manually, there's no way you're going to pull that off. Yeah. But if you have the right script for it, you can do that with hypothetically, any amount of content in a couple of minutes.

Zohra Mubeena-Mutabanna:

Excellent examples, I would agree that they're simple enough that they can really pay off, starting small and experimenting. So Shaughn, you know, I've been experimenting with Python. I haven't told you that. But since I started shadowing you, I've been writing small scripts, and I've been doing exactly the examples you mentioned, to find patterns. I just introduced dumb patterns into a file. And I tried to read to see if it picked it up. And it did. So good examples there. And I feel so motivated, working with you and just shadowing you. So thank you for that, buddy. We talked about ChatGPT. Right. And before I do a deep dive into that this was sort of the precursor to that, you know, you've been doing coding. What do you think about ChatGPT? From your perspective, is it going to replace us?

Shaughn Kern:

No, absolutely not. The first thing I asked of it was, I don't remember if it was trying to find a movie for my daughter. She was like, I want to see a superhero movie. And I wanted to be scary. And superhero has to be a girl. I was like, Okay, I could Google that and shift through sort through all of the results and trying to figure it out. And ChatGPT gave me a really simple, just, here's a list of six movies that would be appropriate for your six year old. Here's the descriptions, here's why we think it'd be appropriate. I was like, Oh, that's pretty cool. I should probably say that I'm relatively late to the ChatGPT game. I am not someone who's interested in having any new piece of technology for the sake of having it. I want to know that it's going to give me practical value. I don't need Alexa, turn the lights on in my house. And I don't know what the implications of that event are. So when you hear people saying, oh, ChadGPT is a singularity and the robots have taken over Skynet has arrived. It's like, probably not. And when you have people say, this is the most powerful tool ever invented, it's going to put people out of business because it's so powerful. When you hear heated sides of both arguments, it's like I'm gonna wait for this to die down. And check this out from a practical standpoint. I love it. I think it's an incredible tool. I think your average person in the tech industry right now is going to have any gven time in the workday, you know, email client, messaging client, maybe Stack Overflow, maybe Google. Chat GPT is going to be the next tool on our computer that is just open at all times of the day. And it's because whereas search engines help you gather information ChatGPT, it gathers it and aggregates it and transforms that information in a novel way that is eerily similar to what we do as technical writers. But it cannot do it better than us. And I guess once we get in some some of the case studies I've looked at, I could start with those next because, yes, my first experience looking at this as like, Oh, this is pretty smart. And the second question, I typed in the ChatGPT was, as tech writers, many of our projects start with research, whether that is us having to research original content ourselves, or get content from subject matter experts. Now we get writer's block, or we're not able to access a subject matter expert. We can't start your project. Yep. So right after I typed in my first question, GPT, I realized, like, wait a minute, I've got this bottleneck going on right now. So I'm supposed to be documenting some internal DevOps processes. The colleague that we're supposed to interview for the information is absolutely the best in the company of this sort of thing. He knows his stuff up and down. But because he's so smart, and so in demand, his calendar is literally booked two weeks out, the normal workflow for tech writers here, the either A, I start looking through the existing documentation on this particular DevOps workflow, or I wait until we can get this DevOps person in the room, right? She's gonna take two weeks, I was like, Well, I got this little chat thing open right here. I type in it's like, how do you? How do you build this, and you know, this infrastructure, I can't go into specifics here. But as a third party, DevOps system, and it's spit out 300 words, I typed in two follow up questions, each one was done your 100 word response, took 20 minutes to parse that down to about 150 words, oh, and then sent that the subject matter expert and said, Hey, I know you're super slammed right now. So rather than you trying to come up with something, send to me yourself, maybe just when you get the chance, go through this review it for accuracy, we can use as a starting point when we do have the chance to meet. And it worked out pretty well. And then I realized, wait a minute, I just shaved two weeks off of what had been would have been the turnaround time for me to just get started on this.

Zohra Mubeena-Mutabanna:

That's pretty neat. Now, in this case, there was information out there already available about this third party tool. Yes, right. So and ChatGPT had access to that data, but but the fact that you were able to use it to your advantage is an example of how you can partner with this, quote, unquote, tool.

Shaughn Kern:

But I was also able to take the edited version of the documentation that I worked on, and combine that with some existing internal documentation specific to our company that I had found. Because you don't want to go around documenting someone else's software, you want to document how to use their software within your institution. Exactly. So I was able to find internal documentation and incorporate that as well.

Zohra Mubeena-Mutabanna:

So this is very human ingenuity. I mean, I think from my takeaway from this example is, you know, at the end of the day, you nothing can replace human ingenuity.

Shaughn Kern:

Yeah, I mean, there was a lot of text to clean up. Because, as you pointed out, this was an existing, this was an existing product from another company that it was getting the documentation from. So depending on how old that company documents their product, that's, I think, gonna inform how well ChatGPT documents for you.

Zohra Mubeena-Mutabanna:

I think that's a very important point. I want to kind of elaborate on that. The fact that the data that is available to ChatGPT is data that has been created by a human. And as it's learning it as learning from existing data, at least at this point in time.

Shaughn Kern:

Precisely and as tech writers, I, one of the reasons I say is we're not going anywhere, anytime soon, with ChatGPT is not going to put us out of a job is because one, it still leaves us to edit this stuff. And two as technical writer we're oftentimes documenting brand new technology and sadly, doesn't exist anywhere else. It doesn't exist, both the components of our jobs where we create new content and parts of our jobs where we refine old content, we are still quite relevant.

Zohra Mubeena-Mutabanna:

Thank you. I really like that take and I do see it that way, as well. I had somebody comment on my post on LinkedIn. It was like a doomsday sort of scenario for them. And I think a lot of there is a lot of excitement. But there's also a lot of panic from what I can sense. And I don't know if this person who responded to my post on LinkedIn was a technical writer or not, but they're like, Oh, my God, this is just going to replace us. And it's going to be the bane of our existence. And I'm thinking, are we creating unnecessary anxiety? With that approach? And that mindset?

Shaughn Kern:

I think it's absolutely okay to be cautious about this sort of thing. One of the earlier things that I read or listened to, I will have to get you her name, I forget the name. She did a really cool little YouTube video, you know, things that taxpayers can use chat GBT, for, she compared it to what happens with translation devices, where all of a sudden you had the feel of translation, we hadn't cracked the code. And yet, you still needed a person to do good translation from one language to another. But once the technology hit, like, good enough, a lot of that was automated. It was just, you know, it's like, well, we can either hire a full time translator, or we could just get something that's fairly close with this automated system. So I do think a company could say, well, we don't feel like hiring a tech writer. So we're just going to have our POs put their stuff into ChatGPT. And they'd get something that's better than if you didn't have a tech writer at all. But it's not going to be good. As an actual tech writer, I would say, my advice to anyone that thinks that their job is going to be taken by the system. Go in and play around with it, but just play around with it in terms of the area of your work. That is the where you're the most valuable. So for example, the scripting and programming that I do to build utilities make my day a little bit easier. That's important to my job. That's not what companies hire me to do. ChatGPT can definitely Oh, code me. But I know that it can't outwrite me. I know that it can't out and me. Now, I don't think Chat,GPT can out code, a professional software engineer or developer. But I think probably I'll edit them. Does that make sense? Yes. So it's gonna prop up the areas where we're not experts. And it means that whatever domain of expertise that we're in, we're gonna really have to double down on it. And we better darn well be passionate about it. Right? Because otherwise, like you said, we're not going to be able to keep our competitive edge. And we're not going to be ahead of this technology. come ChatGPT, five or six.

Zohra Mubeena-Mutabanna:

Great advice, great insight, and a great, I think, great perspective there, Shaughn. Absolutely. I think I think I agree with you. So that kind of naturally segues into my next question. You have done some cool stuff at work, with ChatGPT. If you can share those examples. Can you share some interesting examples? Because I thought as we will, as you and I were talking about it the other day, I thought, hey, you know what technical writers out there who are not thinking of how to leverage ChatGPT as a tool, and start thinking about those. And some of the examples that you shared with me, kind of just led me down that path. So, you know, why don't you just jump in and share those cool examples with

Shaughn Kern:

I mean, my favorite example so far, and I me? think it's the one where it's my favorite because I had the highest aspirations to it, but it was the most lessons learned. So it was maybe the fourth or fifth thing I typed in ChatGPT is like, can you read this URL? To me, it was the release notes for one of the products that I work on. And I said, Sure, and gave the content and it's like, this is not the right content. And I was like, What's going on here? And it turns out, like ChatGPT cannot read live websites. It was reading a website from a couple years ago, which I think a lot of people that are roughly familiar with the engine, they understand that now. It's not reading live websites, it just somehow knows everything that was on the internet three years ago, or whatever it is.

Zohra Mubeena-Mutabanna:

Yeah, I think 2021 That's interesting that so you just randomly went and typed a URL and asked iChatGPT to read that,

Shaughn Kern:

I just went into it with a project. And it was like, let's see how much you can help. It's like, okay, so you can't read live websites. So that's, it's like, okay, well, what happens if I take this raw pseudo HTML file for my Madcap flare project, and just paste that right in there? Sometimes I had success with that. Sometimes I didn't, but it was like, okay, read the content in this HTML and then edit it. And it could do that but I also need you know, before and after of the plain text so that I can defeat them and figure out what exactly ChatGPT changed because I don't trust it enough yet to just give it content and then Ctrll C Ctrl V that into a document. That's not safe at this point. So then it became okay, well, you can't upload a file directly. So now I'm pasting code. And it doesn't like HTML, once you get past 70 to 90 lines, there's a chance that it will choke. And then it's okay. Well, if I paste in the HTML, can you convert that to Markdown and then I take the Markdown code, and use that for the original, then ask it to edit the content in the markdown for clarity, and grammar and the style of technical writing new Markdown and compare that. And so essentially, I'm trying to throw my content at it in as many ways as possible to figure out what type of content that works the best with, and to be able to track the various edits that it's making. I mean, my conclusion at the end of the day is drop it in their habit, convert markdown, and convert it back, the process itself was not worth it, the edits that it provided, and I cycled my entire flair project through it. So about 80 topics, small project, sometimes it picked up really good examples. A lot of times introduced passive voice, a lot of times it overrode our, we have an internal style guide at our company. So it would change bolding or italics or, you know, switching around things as a as it soft feel fit. And then one time, it just took the Markdown and tried to make it into a Python script, just randomly random. And there was no discussion of Python in any of our discussions. Oh, okay. So it's just a complete glitch in the Matrix. In very few occasions, did it introduce anything that was completely wrong, I would say that it offered very diminishing returns on editing content that I worked on recently, however, working on a product that we had acquired from another company, there was some very old content that it saw and just marked right up. And this was good, because that was like on the landing page, or some of the most frequently visited pages of our product, where everyone has been staring at the product description and the content on that page for so long. That no one's thought to edit it. And many years, people just kind of glaze over it. So I think it's great if you're going to use it to edit legacy content, content that was written by someone that maybe wasn't a technical writer, or was from a different company, in terms of having an edit my own writing, not because you know, like, oh, I you know, I can out John Henry, this this machine here. I would take any tech writer at Blackbaud to edit my content before ChatGPT hands down. Will it find you know, random things here? And there? Yes. Is it recycling an entire flare project through there? Probably not until there's a more efficient way of doing it. I don't know if I have found what I would consider an efficient way yet. Getting

Zohra Mubeena-Mutabanna:

Okay, so what kind of prompts did you closer. provide Shaughn, you know, you went through a pretty big process to get to discover whatever one you know, and all that you've shared so far, but just as an example, can you based on what you've shared with us? What kind of prompts did you give with your content to benefit from ChatGPT's feedback, or edits?

Shaughn Kern:

Your question is perfect, because it addresses the areas that I got burned the most in in this Flare project. And that is if you know how to use chat, GBT, you can get content in a code editor that has syntax highlighting, and you can get it in tabular format. You have to specify it. But I would say for example, okay, convert the following HTML to markdown ignoring any madcap elements because Madcap Flare adds additional proprietary elements to help their software process HTML better when I didn't tell it to ignore the madcap elements that says this is not HTML, this is XML, and I cannot read it, which is true, right? Which is true, but I was able to get it to read Madcap Flare content. And it did somewhat fine with that. And it was like convert this to markdown. Then I'll just give the content and it's like, this isn't markdown. That's literally where I type I go, this isn't markdown then it says, Oh, yes, here. Let me display this for you in a code editor. With the raw markdown. It's like perfect. And then the next time I'm dropping an HTML to it, I go, okay. Convert the following HTML to markdown, ignoring any madcap elements, display the full Markdown syntax in your code, which it wouldn't do it every time I had to have a follow up. Question says display this in a code widget. So this is not something where you can put in a single line and have it do everything that you want it to. It has to be iterative, you have to do follow up questions. Now I tried the same thing and ChatGPT-4 for it did everything right away perfectly? Oh, it did? Yes. It takes a lot longer. And both ChatGPT 3.5, which isn't the currently the most recent version and chat GPT four, you're dropping in an HTML file, that's depending on the complexity 100 lines or so it's gonna choke on it and might choke, I mean, it might get the line 70 The one time it might get to line at the next time, it might get the line 60 the next time. So there is a lot of treating this as an iterative process that has to happen.

Zohra Mubeena-Mutabanna:

Okay. So if somebody is, you know, a lone writer, and is trying to figure out, hey, how do I put ChatGPT to my advantage? Is there anything that you can recommend them from this exercise that hey, maybe you can? There are XYZ things that ChatGPT could help you with?

Shaughn Kern:

Absolutely. I mean, just as a research assistant for getting the crappy first draft out of the way. Okay. I think most tech writers, well, I might be wrong here. I think tech writers that have a healthy ego, understand that the first piece of content they put together, it's not going to be great. And that's fine, because we can figure out and edit it as many times as we want. But getting that first chunk out there and going through whether it's Stack Overflow or a bunch of Google hits, I mean, like all this is an ad, I have to drop my email to receive a white paper to access this information. It's a great place to start for that original research and just getting a chunk of content to parse down and hopefully figure out and a complex concept right off the bat for coding, something that I had not considered not only does it write, from my perspective, pretty decent code, it comments all of its code. And our colleague, Bob Costanzo, has been throwing his scripts in there and saying, Yeah, write comments from my code. And it's like, How brilliant is that. And then meanwhile, you've been using it for translation purposes, which, of course, I had never thought of. So I think aside from generating content, getting over writer's block, or using it to comment some code or that, just be creative and play around with it, if you're willing to say, I want to see how this tool works for my project, I think there's a lot to be learned no matter what you're working on. And I really look forward to hearing from other tech writers in terms of what they're using this for.

Zohra Mubeena-Mutabanna:

Yeah, I think I think this is a great period for us to really get experimental with it right? And, and have fun, really more than anything at this point, and not feel intimidated that this is going to replace us. That's for another day. I'm going to leave the worry for another day and have fun with it while it's accessible to us right now. Right? Because I think very soon it's going to become a paid subscription thing. And we may not have access to it for us to experiment. So I think it's great that writers such as you are experimenting with it and having and discovering some new ways model ways. That brings up my next question for you. And I think you mentioned this is you were going to throw our style guide at it and do something cool with it.

Shaughn Kern:

Well, I don't know if that's within my scope of abilities. I would really like to, I looked, one of the things I have enjoyed the most is, you know, if you can't find what you're looking for on Google, you can't just say Google help me figure you out. With ChatGPT I typed in. Okay, so what do I do if I want to give you content? And have you format it and for APA, MLA, Chicago, anything like that, and it's like, I can handle more style guides? And it's like, what about an internal style guide for a company, not a widely known one? And it's like, okay, like, you can do that through the chat interface. But it's gonna be really loose paraphrasing, by the way. ChatGPT is trained to give much better responses than I am much more articulate ChatGPT basically said, You can do it through the usual chat interface, but it's going to be very difficult and potentially inaccurate. Your best bet is to feed it through our our API. Okay. And that's it. We're like, Okay, I'm gonna get myself an API key. I don't know I'm going to get to the point of building my own style guide through it.

Zohra Mubeena-Mutabanna:

But you've tried something at the end of the day. Right?

Shaughn Kern:

And I think the biggest thing I learned there is, ChatGPT understands its own constraints. And if you ask it if it can do something, or if it's giving you a hard time and say, How do I get you to do this? It's really helpful in that regard. And it takes a weird type of creative thinking that we've not really developed with search engine. Search engines, you try and type in a string, and you click enter. If it doesn't work. Yeah, try again. You switch a couple of words around, whereas ChatGPT is iteratively building on what you're doing.

Zohra Mubeena-Mutabanna:

Thank you. I almost felt like I was having a conversation with it. It's like a two way street with a search engine it's a one way street. And frustration can build up very quickly.

Shaughn Kern:

So Zohra to your point about, you know, this has to be a conversation. It's not Google. I think this is a new type of you brought up the term prompt engineering, which thank you for that on your last episode, because I didn't know what the heck it meant. But product engineering is basically, do you know how to use this in a way that's not Google. And one of the things I'm trying to use it as is more like Wikipedia. You know how you go on those Wikipedia? Like, you just click, click, click, click, click, click, and eventually you wind up somewhere completely different. Yep, wasn't necessarily what you set out to do. But you sure learned a whole lot getting there. That's how I'm starting to look at it. And that's happened with some of the, the tech comm stuff that I've discussed, where it's like, oh, how do you work with Markdown? Can you do this with your API? How much does your API costs? What is the token? So having those conversations where you're drilling down more and more, I think is a better way to interact and you know, the Google one off, can I get what I need at the top of the search engine? And I think by treating it that way, you're less likely to take whatever ChatGPT spits out, and just rolls. And so last week was my wife's birthday. We're in the middle of someone's house. We're super busy. And House showings. We're both still working. I was like, What do you want for your birthday? They're like, anything involving smoked salmon? It's like, okay, well, that's one of my favorite ingredients. But she's like, not the pate. Like, alright, well, there goes my ace in the hole there. And it's I was like, okay, ChatGPT. What are some cultures that use smoked salmon as a primary component of dishes and it spits out some stuff. I was like, Okay, let's, let's figure out a salad. And I work with it to build a salad. I was like, what are some dressings that would go off this salad. And they didn't want and it was sour cream and horseradish based, and one that was Dijon mustard, lemon and sour cream based. And at that point, it's like, you know, this is someone's birthday dinner. Specifically, my wife's birthday dinner. It feels like I'm cheating at this point. So it's like, you know what, we're gonna end the chat here. I'm gonna look at these dressings. I'm going to combine them, I'm going to change them. Because, you know, I do have the ability to cook everything on my own. So I very much think that ChatGPT. Whether it's for the professional workplace, or just stuff around your house, you can use it for a lot of really cool things. Great place to start, bad place to stop, especially if you're doing something where you have personal talents that supersede this chat engine, and something where you want to be more meaningful,

Zohra Mubeena-Mutabanna:

Very interesting examples. Shaughn, thank you for sharing something, something so personal with us, you know, we really appreciate it. But it's like you said, you know, use it smartly. But don't stop with it. I don't know where I read this. In some book, the word that came to my mind was Wayfinding, where you're trying to find something. And the path you follow to get to it, you may not end up where you wanted to be. But you eventually land somewhere where it's better than not having landed anywhere at all. Right?

Shaughn Kern:

That's a wonderful way of putting it. If you ever do a Google search, and you see like, the list of like the 10 questions that have been asked, and how banal and nonsensical those questions are. It's the opposite I I'm starting to look at prompt engineering as the opposite of the Ask Yahoo or ask Google I don't know what those categories are called. But where it's like, it's like, no, no, you can't get pregnant from that. Where did you go to school? Who asked this question? Who asked this question so many times that it became a topic in Google, it's being the opposite of the people ask the bad questions and in Google.

Zohra Mubeena-Mutabanna:

Oh, interesting perspective again there, Shaughn, I agree. Another thing that I'm going to be trying out actually, right after this interview, something that I wanted to share with you is a very dear family member of mine is looking to pivot their career. And based on their skills, they're like, I can either go to a career coach, which they will, but just as a starting point, just as a discovery, what kind of careers can I look at or start talking about? And when I go to the career coach, I'm like, This is what I want, and how can I make it there. But it's kind of being more prepared and informed in in their career choices that they're going to make. They're thinking of throwing their resume at ChatGPT and asking, based on my skills, what career options can I have? So that's something that they want my help with. And that's something that I'm going to try. I haven't tried that yet. But

Shaughn Kern:

Doing a lateral movement, a career wise, because we don't always know where the fine overlaps are with our current roles. And that's brilliant. And it's great that they are also going to a professional career course afterwards. Because again, this robot cannot outperform sorry. So language model, language model cannot outperform a lot of professionals yet. But at the same time, if they look at all the stuff that ChatGPT spits out and they know immediately, like, I don't want this. I don't want this one. What even is this clear? No, that could help them have a more focused conversation and get the most out of the professional career guide guidance.

Zohra Mubeena-Mutabanna:

Yeah, that's what we are expecting. And I think, to a large extent, that's how the story will pan out. But this kind of brings us back to the point where use it for your research, use it to edit, but don't make it the the end all of your work, right, use it as a means and as an aid, really, but nothing more than that. So this has been a fantastic discussion, Shaughn, I have enjoyed picking your brains as always. And thank you for your time on a Sunday afternoon while you're in the midst of a storm trying to manage everything else.

Shaughn Kern:

It's absolutely my pleasure and Zohra I am sincerely looking forward to whatever we collaborate on next with the ChatGPT stuff at work. This is a fun time to be in the discipline. And I think it will continue to be.

Zohra Mubeena-Mutabanna:

Yes, and I'm just going to add to that. And it's a fun time for me to be picking your brains. Um, so thank you for being my buddy and for helping me with everything that I want to learn. I'm always curious, and it's people like you that make it fun for me. So thank you.

Shaughn Kern:

I'm very flattered, and you're very kind.

Zohra Mubeena-Mutabanna:

Thank you. Have a great weekend. Shaughn, whatever is left of it. And I'll see you on the other side at work tomorrow. Bye Zohra. Bye, Shaughn. Subscribe to the podcast on your favorite app, such as Google, Apple or Spotify. For the latest on my show, follow me on LinkedIn or visit me at www.insidetechcomm. show. Catch you on another episode.