What we are doing with GenAI now... (it's pretty cool)

Jason had pneumonia when he recorded this very long pod with Inger.

It's a testament to his brainy-ness (and Inger's sound engineering skills) that it turned out as well as it did, given he coughed and sweated his way through it.

Warning: in the chat section at the top there's a lot of talk about TEQSA. If you're confused, here's their webpage. Jason was too sick to do mailbag, but please keep sending in your letters and speak pipes!

At we switch to our work problems section (26:04)  Large Language Models (LLMs) - we recommend you have a look at our Discussion Guide as we talk, or after you get back from your walk/gardening/gym/cooking or what every you are doing. You can read the briefing note that Inger wrote for her boss here.

We skip the reading section and  coast out with some two minute tips (1:38:22)  for planning large complex tasks and reading all the instructions before you start doing a thing.

The good news is that Jason did not die of his pneumonia and is back on his feet again - phew!

Things we mentioned:

On the Reg discussion guide for this episode (free to share)
The briefing paper Inger wrote for her boss on how researchers are using ChattieG
Timothy Snyder On Tyranny Book
Manchester Phrase Bank
Randy Olsen 'and but therefore' writing tip is in 'Houston, We have a Narrative'
Team Human podcast

We've changed our email address!! You can email us on <pod@ontheregteam.com>
Leave us a message on www.speakpipe.com/thesiswhisperer.
Visit our (pretty terrible) website: www.ontheregteam.com

Jason is having a break from the Socials, but you can still find Inger as @thesiswhisperer pretty much everywhere (and she will friend anyone on LinkedIn). You can read Inger's stuff on www.thesiswhisperer.com.

It costs us about $1000 a year to produce this podcast. If you want to support our work, you can sign up to be a'Riding the Bus' member for just $2 a month, via our On The Reg Ko-Fi site

Chapters

• 0:00 Chat
• 26:04 Work problems
• 1:38:22 Two Minute Tips

Transcript

season 5 - episode two

Yeah. Pneumonia. It's like, that's great. It's awesome.

Oh yeah. It's like, whatever, pneumonia, come get some.

It'll be fine. I just, I imagine that's not at all toxic masculinity, my friend. No, not at all. I imagine there's a bunch of people who are listening right now. I just go like Jason. Bloody hell. Switch your shit out. Go to a doctor. I will. Thursday. Okay, fair enough. You're driving the bus. Do you feel like? Oh crap.

Am I? Yeah. I'm terrible at this. Okay. So I'm going to start with just an apology. Just straight out of the box. I, I can't just sound terrible. As I said, no, I think you sound fine. I think you sound like husky radio voice. Like I'm not, as, as our sound producer, I'm not worried about it. Okay. All right, cool.

Oh man. Welcome to On The Reg. I'm Dr. Jason Downs from La Trobe University. And I'm here with my good friend, Professor Inga Mewburn from the Australian National University. But she is better known as the thesis whisperer on the internet. We're here for another episode of On The Reg. Where we talk about work, but you know, not in a boring way.

We are all about practical implementable productivity hacks to help you live a more balanced life We'd also like to make it clear that even though we work at a couple of awesome universities, that this podcast is not connected to either of them, nor are we representing them.

On the reg is our own endeavour. Welcome Inga. Have you been since we last caught up? Well, I don't know. I think you should tell the listeners about your husky radio voice, which sounds good, by the way. Yes, but it comes at a cost. You are walking wounded. Take one for the team today. Yeah, we have to, we have to record.

We've got to get, we've got to. Got to keep up our production values. Um, uh, yeah, I've got what's, um, the doctor referred to as community acquired pneumonia. And that essentially is just pneumonia that you get when you don't get it in a, like a medical facility. I had to Google this, right? I go like, community acquired pneumonia.

Is that different to any other kind? Um, and that's the only difference between pneumonia and pneumonia and And the kind that I've got is that I just didn't pick it up in a hospital or somewhere like that. But, uh, oh man, it's, it's knocked me around a fair bit. I can't I, it's, I, I'm wheezy. You can hear that in my voice, I think.

And a little bit short of breath and I'm coughing so hard. I had a coughing fit the other night. Probably went for 10 minutes, um, while I tried to kind of shift that stuff from my lungs. And I coughed so hard that I think I bruised my ribs. Like, it's just like, it's all, it's all just painful through my ribs and stuff.

And if I cough now, the muscles up, up the run up the, up the back, up your back, they are like, I can just feel them just going like, stop that now. Have I ever told you the story about when Brendan got pneumonia? So he's about seven and he was at that age where, you know, they skive off from school. I'm sick, mom.

I'm sick. Anyway, he, he was not, he was skiving off from school. And so there was a bit of a pattern of that. And then, um, he, he of course was well enough to go to the Royal Melbourne show. All of a sudden, and so he did have a bit of a sniffle. I noticed when I took him to the show, we took him to the show with a couple of friends and then he was kind of just a bit not himself and I thought, Oh, that's unusual.

Maybe he really is sick. And then we got home and he's like, I'm really sick. And I'm like, you're not that sick. Just, you know, go to bed. Here's some lemon and this went on for three days and eventually I took him to the doctor and the doctor's like, your child has pneumonia. Put us in an ambulance. Oh man.

He took us to the hospital because he was so sick with pneumonia. So like I've over corrected parenting, you know, it's easy to do, it's all I'm saying when they lie to you, when they learn to lie, it's hard, yeah, anyway, yes, well, thank you for taking one for the team, they're struggling on through, like, yeah, yeah, it's fine, right.

My memory of pneumonia though is, um, I was a young, very young lad. And dad caught pneumonia and I can't, I can remember him coughing like, like I was really young. I can remember him sitting in the lounge room coughing, it was the middle of winter, um, we had the fire going in the lounge room and, and, and I can remember just the fear in the house.

Do you know what I mean? Him having it. Yeah. Like, it's like, this is really bad and yeah, all that sort of stuff. And I've been walking around, I've been walking around like it's fine, whatever. Like maybe I need to start paying a bit more attention to it. Like don't call it walking pneumonia for nothing.

Like. Yeah. Yeah. Yeah. There's four stages and if you read the Googles about those four stages, it's like, it's like Jesus. Yeah. I think the fourth stage is death. Isn't it? Something like that. Resolution is how they call it. Resolve one way or the other. That's it. Right. Healthy resolution. All right. We are on the clock.

We're on the clock because you've got pneumonia and I've got pneumonia. I've got door knocking for the greens to host. So anyway, it was Lunar New Year, Jason, the last couple of days. Yes. Yeah. How awesome is that? Yesterday was actually New Year's Day. And I was, I was lucky enough to be honored first New Year's guest at Nguyen and Tong's house, who are Vietnamese and celebrate the Vietnamese version of Lunar New Year.

So I learned all about the different sort of trees and colored blossoms and all the traditions. They, they gave me some traditional food. That was amazing. Then we went to the temple. We saw the dragons, we saw the firecrackers, uh, there was praying. I didn't pray because I'm not religious. And I always feel it's weird if you pretend to pray for someone else.

But luckily Nguyen played, prayed for me. Nguyen used to be my doctoral student, now Dr. Nguyen Bai. And, um, she, uh, she got, she prayed so that I could get a lucky orange. And at the temple, so I've, I've got, you had to close your eyes and fish an orange out of this bucket, and I got one that's labelled courage.

That's, I've got a courage orange. That's awesome. And that's meant to set me up for the year with courage. I thought that was great. Oh, man, this is, I'm going to show my cultural ignorance here. Are you allowed to eat the orange? Well, I said that and she's like, well, you can eat the orange. Some people do, but she keeps it all year.

She said, oh no, it keeps all right. And I'm like, does it though? Anyway, so we'll see. I've just got it on my desk. I don't want to throw out the curry, George. That isn't what you do, right? Like, yeah, it might just find its way into my garden. I was about to say you could plant the courage orange and then you could have like a whole tree of them.

Anyway, next week I'll need my courage because I'm going to the Taylor Swift concert next Saturday night. God help me. I actually. Why? I actually, from a, from a youth where I worked in a record store, love live music, I've come to actually hate live music. I love music, but it's so loud. I always forget.

Every time I go to a concert, I'm like, there's too many people. It's too loud. It's too noisy. I don't like it. And then it's like getting on one of those fun house rides, you know, where you're like, why did I sign up for this? Why am I on this roller coaster? I don't like roller coasters. I did it again.

Anyway, anyway, so I'm going to Taylor Swift concert at the MCG, so it's going to be massive. Um, but Anitra, my sister Anitra, got into the Taylor Swift poseum, Jason. This is a career highlight. So it's a True Dinks academic conference. On the thing with Taylor Swift, they put out calls for papers and she said to me, do you want to be on my call for papers for the Taylor Swift conference?

I didn't say heck no. I just was like, Oh, that sounds great. No, I don't have time to do that right now. Anyway, there were 400 papers, um, submitted to the Taylor Swift poseum and Anitra's paper got selected. But she'd written one of those abstracts, you know, that you write for the paper that you. Then when you actually go to write the paper, you're like, there's no possible way I can do this.

She wanted to demonstrate something like the dark Taylor halo effect online or something. And then she had to actually do it. So I ended up in these endless text messages with her about Gabrielle Tard, this like economist from the 19th century, 19th century. Anyway, and she said in the end, I had to get an acknowledgement because I did some of the intellectually rigorous thinking, but I will admit that I did have to ask chat, JPG, Afro and chatty.

To explain TARD to me because I couldn't remember it. And um, anyway, so there was that. So I'm going to Queensland on Monday. I'm going to spend some time in Queensland. I'm going to give a talk on Thursday in Queensland. That's what I've been invited to do, but the rest of the time I'm going to see my mates in Queensland that I haven't seen for a while, my nerdy friends, talk about papers and talk about, you know, have meetings and stuff.

It's going to be great. And then I'm going to fly to Melbourne. I'm doing a big triangle. Canberra, Brisbane, Melbourne. Back to camp, Melbourne for Taylor Swift, and then back to Canberra next week. So it's, it's busy and work's just been nuts. It's been nuts, but not as nuts as you've been at work. Have you been nutty?

You tell me what's been going on. Deadlines are, um. Deadlines are here, so, it just means this is the kind of work that, um, no one likes to do, right? Like you, there's no choice. You just have to grind it out. Yeah. And you know, Texa doesn't wait for anyone really. Like, you know, you submit, you submit your re registration, um, they give you a date and that's like, they don't move it.

So, you know, you just do what you need to do to make sure, make sure that you're ready. And, um, You know, this kind of work, right? Like this self assurance work and quality work that there's no red line, right? Yeah. It's just like, you know, you have to make, you've got to make your own judgment as to whether or not you satisfy the standards or not.

And of course we do, but like, then there's, it could always be better. It can always be better, right? Yeah. It doesn't matter. It doesn't matter whether you're on the jujitsu mats or whether or not you're doing quality work for university or whether or not you're baking a cake, right? Like it could always just be better.

Yeah. And also you're not the one who's really in the best position to judge at that point. Like, do you know what I mean? Like when you say something could be better, you're just so close to it by that point. You like, you lose all perspective. I think, you know, yeah, that's that's certainly a risk the but there's, we've gotten so many checks and balances around all of this sort of stuff to make sure that, that that doesn't happen.

Um, you know, there's more than just, you know, my set of eyes on this thing, right? As you can imagine. That's why we have teams. Yes. That's why we have teams. And, uh, yeah . You, you, it's standards based regulation, right? Like you've got to hit the standards. So you make sure that you hit them.

And I'm the kind of person who likes to hit them and then nail them into the ground and then keep hitting them until they're well and truly buried. That's why you're a good match for what you do, Jason. It's like, yeah, it's like, we've got this. Absolutely. And we'll show you how I've got it. We're 79 different documents.

You're welcome. Oh, yeah. Thank you, Texa. We got our text to reread through it got announced in the all staff newsletter. I think I texted you. Um, I screen grabbed that bit and I said, I'm probably one of the like five people in the university that goes, wow, we did it. Yay. Cause I only know how much work it is being your friend behind the scenes.

But I know I had a little bit to do with the team. Like I know they're working on it for bloody ages. Yeah, there's, there's so much work that goes on to make this all happen. And, and of course, you know, the way you want to do this sort of stuff is not to panic anybody else. Right. So you try and have as low impact as you possibly can on the rest of the organization.

It's like, Hey, have you got some, you know, I'm looking for this kind of document. Have you got something like that? Thanks very much. Kind of carry on sort of thing. And then in the background, it's like, you're starting to assemble this, um. You're putting together something that is going to be meaningful and in the format that the regulator wants it because of course they change their format.

Um, uh, in, I think about October, uh, yeah, just, just like that, like through the work that you were doing on it, they just went, you know what, not a thousand word essay. Now you have to give a Kabuki performance, kind of in, in interpretive dance. Yeah. Yeah. So that, so that, that was fun. I can just imagine, um, like your team, um, at ANU because you would have had to have been much closer.

Right. Yeah. Yeah. For sure. Like if you've just got it announced that you got it, I don't know how long, six months or something for them to make that decision about whether or not, so, um, you know, you, you guys would have done all the work under the old regime and then they would have changed it. And then, you know, you're kind of waiting for, waiting to hear.

So really they let us hand it in. Like in the other format. Yeah. But as I said before, the folks at Texa are like, I've, you know, my dealings with them, very smart, very nice folks. Yeah. I I actually know more people at Texa than I thought I did. So I went to the Texa conference, uh, which was in, I think November.

Last year, um, bumped into some old colleagues from RMIT and they have moved across to working in Texa now. So it was like, you know, it was good to sit down and have a bit of a chat with them and, uh, yeah, so career move to you maybe Jason. One day? Oh, I don't know. I don't think so. Not cut out for government work?

Oh, I don't know. I saw that they were, um, they had an announcement, a job advertisement for the CEO of Texa. Oh, they had a CEO? Yeah. I'm like, wow, that'd be a job. I wonder if he gets to meet the Prime Minister. And when you go to the prime minister, like one day a year and the prime minister says, how are all the universities you go, fine, it's all good.

So yeah, like maybe that's the job. Yeah. Right. Yeah. We've got good quality here in Australia. You can come to Canberra. I'll take you out for a good dinner. You go back home. It's all done. I like as much as, as much as, um, uh, we make a little bit of fun of it though. I, you know, I have this real. Um, regard for the way in which the legislation is set up and the way in which, um, we go about ensuring that our universities are a world class standard.

Because they are. They are a world class standard. Yeah, a lot of bad things about the sector, and God knows I do, right? But I think I do that, and then I go other places and look at other systems and go, nah, it's fine. It's all good. Yeah. Yeah. Yeah. You know, and I think. You know, the legislation has really got its heart in the right place, you know, it's aimed, it's aimed at students, right?

Like it's aimed to ensure that our students who study here in Australia get a world class education. Um, and we have, we have. Specifics, um, legislation that supports international students who study here onshore as well to ensure that, that they are well supported through, that they're not, taken advantage of and, or any of that sort of stuff.

And that's important. I don't know if you've noticed in your travels around, but there's a lot of jobs and units and people doing work that I reckon the university just wouldn't. Like I, first of all, think to pay for, and second of all, when, when they're strapped for cash would cut if they didn't have to do it.

So it's like, it's quite good that we've got this check and balance in the system. In fact, um, you know, my job is, exists because of Texa. Oh, really? Yeah. Director of Research and Development. Um, in the Higher Education Standards Framework, which is the document that documents a lot of Texas, Tertiary Education Quality Standards.

Seven domains. You'd say, you're an internationalist and you're like, what the hell are they? They've said Texas about 20 times, but, um, but yeah, I, I was saying to a group of students the other day when I was doing a workshop, it's basically, you know, I am the risk management apparatus at the university to make sure that you, uh.

Don't take too long to complete and you don't have a terrible time. So yeah, you're welcome. Yeah. Yeah. So, you know, there's a lot of us who have jobs because we, we are in the quality space, quality of experience, quality, you know, equity of outcomes and stuff like that, you know, but then you, you know, you look at how that system's under attack in the U S and you just think, Yeah, there could be worse places to be.

That's all I'm saying when it comes to caring about people and their experience of things. Anyway, so but you haven't been able to go to the old BJJ and stuff because there's no Brazilian jiu jitsu for you while you've got the pneumonias. Well, it's contagious, right? So, yeah. You haven't given it to any of your family.

They didn't give you COVID. You haven't given them pneumonia. That's good. Sharing is caring, but not that much, right? Not that much. So, we're kind of, we're keeping. At home, we try and keep a bit of a distance between everyone, which, um, as much as you possibly can. So you have to sleep in the snoring room then, do you?

Yeah. I'm afraid so. It's also known as the coughing room. Right. The go away, you've got COVID room. Everyone needs one. I'm actually sitting in our room, I've moved office. So this is, this is now my office, which is also the spare room, but also the biggest room in the house, other than the lounge room. So you can see I've got my porcelain collection, I've got my plants, I've got my cushions, you know.

It looks all very nice. Thank you. We used to joke that, um, I got really cranky about this one particular seat on the couch because everyone else in the house had a room but me. So Brendan had his own room, Luke has got a room to work in because he telecommutes And then we have the shared family office, which is what I'm in now.

And I was like, I just don't have a room to myself because I share my bedroom and I share my office. And now I do. Everyone's happy. Everyone's got a room. Nice. Yes. Excellent. Somewhere for me to put my porcelain. So no one else has to look at it. Except for guests of course. Excellent. Yeah. I know. Cause you drag them from the front of the house.

It's like quick, quick, quick, come and have a look at my porcelain. Well, no, they have to stay and they have to sleep and they, I've got a little porcelain lamp there, you know, with dancing ladies on it, but it is porcelain paradise in here. Right. There's a story behind my porcelain collection, but maybe I'll save it for another time.

Yes. Yes. Hey, um, what else has been going on in the world of work for you? Anything? Uh, no. No. It's just too much of it. Just too much of it. Just too much of it. That's, that's the only thing I have. Yeah. Yeah. I think I clocked 55 hours last week. That's crazy. Um, I did, I did 46 and that was bad. Yeah. Yeah. And I, you know, I don't think I captured every hour that I, away from the keyboard where I was thinking about it either, or where I woke up at three o'clock in the morning and thought about it and things like that, because it's just like start of semesters.

Like we don't even work on semesters really, cause we're, you know, PhD, but something about the start of semester just is like collective crazy that kind of rubs off on you or something and everyone comes back and everything that you shoved onto their desk just before Christmas, they shoved back at you, you know?

Excellent. Yeah. But on the other hand, um, orientation, uh, a week coming up very soon for us. Yeah. Yeah. For, for the, most of the unis, they kind of line up a little bit. Um, It's always the new fashions, like, it's been a poor, I look, I've just got to say it. I I'm the mother of a gen Zed, but their, their fashions are boring, man.

Like they all look the same. They've all watched the same Tik Tok, Tik Tok videos. There's no self expression there anymore. I just, I like, am I getting old? Probably, but I'm hoping this year we're going to just not see baggy jeans with tiny midriff tops. And for the boys, that eche look with the like slung over, like designing, it's not, it's not a great fashion time, but like I can relate because it's like, you know what it is?

It's back to the early nineties when I was that age. Right. So, and I would say that our fashion time was not great. And now they're recycling the not great fashion time. Do you remember the bucket hats, Jason? You remember the bucket hats. I do. I never wore one. It was ill advised. Yes, I wore one in a nightclub once with a big fluffy floppy rose on it.

I'm just waiting for those to come back

In Thailand they had all these little bucket hats for tourists to wear now had marijuana Like motifs across them and so I'm like why would you yeah, I don't know No, that is a very 90s vibe Hey, I've been, um, I have been lucky enough to get an early look at the new co pilot feature in the Microsoft suite.

Yes. And you're going to tell us all about it. In the word problem segment though, Jason, yeah, the gen AI stuff. So, um, yeah, is that what we're moving on to now? We're not doing mailbag. I make, I made an executive decision. We actually had some really great letters and you and I, you know, replying to all those letters and having conversations with people and, you know.

Big shout out to the people who've written to us and, you know, and also taking advantage of the kind of like free advice that we offer in debugging your obsidian and all that kind of stuff. I really, I promise I'm on it. Yes. Coralie, Coralie English, friend of the pod continued our conversation around Alfred versus LaunchBar versus Spotlight.

She did, yeah. And she pointed us in the direction of a little. piece of software, which is free and I can't remember the name of it now. Well, you find copy clip, clip. I think it's called, I think it's called copy clip, which solved, solved my problem, which is the, it looks, it does one job and it does it bloody well.

I'm telling you. When you copy something from one document, it kind of holds it in its little memory. So you can, you can copy three or four different things. You can go to a spreadsheet and copy something. You can go to a document, copy something else. You can, you know, copy the details of a copy, the details of a contact, like if you're a phone number or something like that.

Um, and it just kind of holds them in this virtual list. And so if you need to go back to something, there it is, right? Yeah. And we changed our lives with that Coralie, by the way. Thank you. She said in a, in an email that she uses it every day and I can bloody see why it's awesome. I downloaded this thing.

It took about 10 minutes for me to realize its value. And then I was working in a spreadsheet where it was a lot of kind of data validation and, um, you had to, one of my jobs was to validate that the information that we had was categorized in the right category. And so, um, Which meant if it wasn't in the right category, I had to change it to the right category.

So of course, it was just a case of making sure that I had all of those categories in copy clip. And then when I, when I got to one that was in category A and it should have been in category B, I could just go up to copy clip. Copy paste, there was no messing around. There was no typing. There was Excel. I hate you.

There was no anything that I like because Excel on a Mac. Right. I know it's just, I mean, I love my Mac, but if there was anything that would ever made me go to a PC, it would be just. Just being behind in those, those office releases, you know, and just having a different feature set. And none of the things on the internet, when you search for it quite work the way you expect, cause you haven't got the same things as PCs.

Yeah. Yeah. Yeah. Colleague, um, uh, sent through a, a tip, an Excel working tip. Um, and of course it was, it was based for PCs. And I'm like, oh yes, so the, the key binding that you want for the, on the, on a Mac is this other key binding. It won't work the way you wish it would. It'll take you half an hour of Googling to figure that out.

Yeah. Yeah. That's right. Yeah. I mean, we do love our Macs though, you know, pry them from our cold dead hands, but. Oh my gosh. I don't think I've ever done the deep dive into the app store looking to solve very specific problems, right?

Like this one. Um, like I, I kind of tend to go for. Something like launch bar that, that has that feature built in that I really like, but it also does 50, 57 other things. Right. Right. And I'm like, I wonder whether or not there's. And so I think there's a little bit of noodling around in the app store for me to do in my near future, just to kind of see what other little pieces of software, little gems like that, single use, they don't do anything else.

They're lightweight, they do one job, they do it properly, they do it well. And if CopyClip's any example, they're for free, right? Yeah, absolutely. Think you're right. Like, you need to do that deep dive. Maybe if you get some sick leave. Are you taking any sick leave? Uh, no, no. I have, um, No? We have, we have deadlines, right? Yeah, but isn't, isn't that what sick leave is for? Uh, yeah. Like, I'm not, I'm not dying.

Right. You could, like, that's the thing about pneumonia. It's like not, not death.

Is that right? Should I get that double negative right? I'm not quite sure. I'm just saying it's, its, it does fall under life threatening conditions. That's okay. It's okay. Sometimes one doesn't recover unless one just rests Anyway. Yeah. It's okay though. Alright. I just, you're gonna be fine. It's fine.

Everything's fine. Everything's fine. All right, are we doing work problems? I'm sitting here sweating. That's okay.

Yes. Let's do our work problems. This part of the show, we focus on one aspect of our work and well, we tend to nerd out on that. So we sometimes tackle problems we've had at work. Or we discuss a theme that has been suggested by a listener. So please feel free. To ring us on speak pipe. Do they, do you ring people on speak pipe?

I'll pay it. Yeah, fine. Yeah. Yeah. Ring us on speak pipe. Um, if you've got a, if you've got a work problem that you'd like us to turn our minds to, um, please tell us. We love that stuff. Um, we always try to be practical, sharing our own tips and hacks and our feel opinions. And this week, our topic is. What we're doing with generative AI, gen AI, as Inga has put in the little thing here.

Um, and I noticed, Inga, you've written, uh, Inga has a grumpy rant about how it's not really AI. It's not intelligent. I'm sorry. It's just not, this is marketing hype. It's technically a large language model or an, or a model, a machine model trained on images. It's fancy statistics still all the way down, very, very fancy statistics, but if you call it intelligence, that's really good for getting seed money in Silicon Valley.

Oh, yeah, yeah. Yeah. Um, and we've got, we've got a discussion guide. You've whipped up. We did. A discussion guide for this. People have been telling me that they're missing our discussion guides. We haven't done one for so long. I realize it's been like nine, ten months since we did a discussion guide. Oh, really?

Yeah, yeah, yeah. So this is like when we do a PowerPoint slide deck, like we're teaching. Yes. And then we put the link in the show notes. And then you can just follow along. And so when I say look at the images on slide number eight, for instance, which I will do shortly, then you can look at the images on slide number eight.

So it's sort of like a storybook. It's like we're there with you. Right? Yeah, yeah, yeah, yeah. So, um, on Google Slides and as part of Inga's philosophy around the share economy, free for use. Please, you know, be kind and cite appropriately if you're going to use it, right. You can use it, it's fine. You can use it.

So, um, link, link will be in the show notes. Um, if you need to press pause now, go off to the show notes, click the link and you'll find a slide deck there that it's coming in at a healthy 16 slides. Well, I think that's healthy because, you know, sometimes when I do slide decks, I look at the total number at the end.

I'm like, Ooh, 72. Yes. And I've got 15 minutes. Exactly. I've called it what we're using Gen AI for now. February, 2024, because we did, if you go to slide two, we did do a, um, another pod about a year ago. We were ahead of this trend, Jason, we were so ahead of it. We were like busting it out. And we did an episode called generative AI pearl clutching panic fest or are the robots coming for our jobs.

So the robots have not yet come for our jobs. Uh, but we have certainly witnessed a lot of pearl clutching over the last year. Um, untold number of articles, like the students are cheating. Everything is terrible. Everyone's going to lose their jobs. I don't know. I've lived through this so many times in my life now, and no doubt you have too, Jason.

And it seems to me like every time a technology is invented, it's more work, not less. Yeah. But look, you know, maybe this time it'll be different. I don't know. But I'm not seeing any less work. There are some, I, I'm sure we'll talk about this, but there are some things that I think, uh, there is. It'll change work, it'll change work, definitely.

And people will lose and are losing their jobs. I shouldn't be flipping about that. People that will, it will make a difference, but I think more like the work itself moves and changes and what we have to do as humans, as we work with machines. And so it is no, like when we did our first episode, I was just playing around.

In fact, I got chat GPT three, and I now call chat GPT because my sister gave it this great name. We couldn't make it into the word of the year, but we tried, um, chatty G cause we got chatty. I knew we succeeded when someone called it chatty G and didn't realize that I'd been one of the ones who tried to make it happen.

And so I was like, this, this is great. Um, so it got called chatty G at the Texas conference. That's awesome. I know, right. You said that to me. Yeah, yeah, yeah. So the text of folks circling back to the top of the episode, nice little call back there, Jason. Um, yeah, so we, we, and I looked back at our slide deck from then, and I'll include the link in this slide deck to the other slide deck, and I got chatGPT3 to write the slide deck for me last time.

And then we just reflected on what it did. And, you know, I looked back at some of its answers then and bloody how it's improved since then. Like CHEP GPT 4 is something else. So like last year, and I'm going to slide three now. It's like my, I had this like list of reflections on using it. And I said, it was fun.

It was quick. It was scary and reassuring. Um, in the, um, it couldn't do everything. There were lots of limitations. And then I said, it would take a lot of effort to make something genuinely useful. I was, yeah, it's genuinely useful now. And you'll talk about Copilot towards the end and maybe through the episode, but it is being now built into our platforms and built into the programs and that's happened with lightning speed.

And I said at the end of that list that if they charge for it, I'd say, take my money. I did. In fact. Say, take my money, I've been paying for it most of the year. It is not cheap, Jase, like it's 30 a month, which is not, I don't think that's cheap considering how many subscriptions I have coming out of my backside.

Like seriously, I've got like, when I start to add them up. It gets into serious hundreds of dollars a month in streaming services and Canva and all sorts of other things. So thank you to those people who support me on Patreon, which sort of doesn't quite cover everything that I have for the pod, like Riverside and, and everything.

So, Um, but I, I pay for Chatty G happily. It's one of those things that would be last on my list to be struck off because it's an indispensable tool of work for me. Now I have it open all day, every day. Um, I use it all the time and every time I have a task, I think, well, I wonder if Chatty can do this. And I try that first within the limits.

Of which I'm going to discuss and and I've been talking to people around ANU and a lot of the efforts have not concentrated on researchers and people like me, how we're using it, what's ethical for us. And so I kept complaining about that to my boss. And every time someone sent me a, well, can I do this with AI?

Is this allowed in a PhD? I just would send it up with a shrug emoji. Yes, and eventually she's like, okay, okay. Will you write me a paper about it? Cause I kept bitching at it. So I've sort of taken bits of the paper out, um, for this chat. So yeah, so slide four, um, and I'm sure you've got a lot to say about this, Jason, cause this is.

More your area than mine. I'm sure it's been talked about a lot of your words. I've called it the debate in inverted commas, um, I think has, as we predicted, centered on the moral panic of undergraduates cheating and not much discussion about people like us and how we might use it for work. And the panic is focused, I think.

On legitimate area of concern, which is hallucination and deep fakes and misinformation, and these are problems, but that's sort of overlooked or you get less discussion of the problems with copyright, with privacy, with exploited labor of these platforms of people doing work in African countries to moderate content and greenhouse gases, it's as much of a user of.

Carbon intensive type of activity is crypto and I hate crypto and I think crypto is a blight on our society. There's never any ethical consumption under capitalism. Jason, your thoughts. This is more your thing than mine. I mean, it's like the way in which we think about, um, academic integrity here. So, you know, the debate, the big one that everybody's been having like that in January.

When Copilot became real, like it's available for Microsoft's Copilot, which is their own kind of generative AI, um, that they bake into all of their platform. So it's in PowerPoint, it's in Excel, it's in Word. When that became a subscription that's open to the public. That game changed, right? So, um, built on top of the chat GPT or the open AI chat GPT stuff.

And it's just brought it to the masses. So, and you know, it queries the web in real time. And so it's now a case of you go to chat GPT inside word and you say, Hey, tell me about this particular topic. And it goes away, queries the web, comes back, tells you about it. Yeah. Provide you with, provide you with the links so that you can go and validate that information.

Yeah. So the fact checking part is much easier. Much easier, right? And it's everywhere now. So the expectation I think of people will be that, oh, this is just another tool that comes ship, ships with, this is the modern way of doing work now, right?

Right. So. Universities need to really start to think very, very carefully about the way in which they go about talking about the use of gen AI, uh, when it comes to things like academic integrity, it's like, it's incredibly important that we do have academic integrity. It's incredibly important.

The, we do cite our sources like it, it shows us the way in which we, or it shows us and others, the way in which we've been able to develop our knowledge and the way, and that kind of our thinking, it's a thinking map, if you like, for what, for people to be able to put arguments together. And that, that's a real quality issue.

And so it's incredibly important that we, that we still maintain that and in fact, strengthen that because. You know, others elsewhere might not, um, you know, quite so quite such rigorous approach to this sort of stuff. And, and as a result, you won't be able to trust it as much. I was thinking about this just the other day, because.

And I had you in my mind, Inga, when I was thinking about this, I was, I was driving along and I was thinking about this particular topic and I was like, what is the future of academics in the, in the space where generative AI is so ubiquitous and it's so everywhere? And I was comforted by the thought that this actually makes academics more valuable to society.

Like this is, this is kind of the moment, right? And you, and I think you see it when you look to the general press. I see more professor so and so from Monash university is quoted as saying, and, you know, there's a particular issue that's in the general press and what we're seeing, I think. Now, more than ever before, is journalists turning to academics to get their commentary on this particular.

Oh, I agree with that. And actually you can also see it in the converse. You can see the attacks on academics in, especially in the U S I think it's, but it's in every country that has authoritarian leanings like the U. S. does, unfortunately, I can't believe that sentence came out of my mouth, but it's real, we're here in 2024, that they will attack universities because we're trusted.

And the reason we're trusted is that due diligence, that citations, that careful fact checking, that weighing up of evidence, that doesn't change. And I said to a group of PhD students today, it should make you even more confident about getting a PhD. It Show you that you're even more valuable. And it might be the only degree in the end that people really have, um, you know, super trust for, because you've honed your, um, you've honed yourself to that capacity.

So therefore that some of the things we're going to talk about, like the erosion and atrophying of those skillsets need to be resisted at all costs. Like I totally agree. Like, I know we're getting very serious on the reg here, but you know, that's like, this shit is important because we live and die on our reputations.

And there's enough challenges to it with the kind of commodification of academic practice, right? With publication metrics and that kind of push to, um, to publish, publish or perish and essay meals and all that kind of stuff. All those sort of fake peer reviews, the, the things that are happening in the, in our publication systems are really terrible.

Like we really need to keep an eye on that because that is undermining our integrity, big time, more so than AI is. Yeah. Even though it's not AI. It's large language models. I'm still like, I'll die on this hill. Hey, I've, um, not to, not to swerve off into business too far, but I'm reading very, very slowly.

Yeah. Very, very slowly. But carefully. Like, but, but for that particular reason, like I'm reading, um, an essay and then I'm thinking about it for two or three weeks. Rather than just diving in. Anyway, I'm reading. Oh. Timothy Snyder's called On Tyranny. I've heard it. 20 lessons. Yes. Oh, have you? Yeah. 20 lessons from the 20th century.

Yes. Yeah. A couple of times actually. It's good. I'm really enjoying it. I'm really enjoying it. It's one of those ones where the, the lessons are really, really short, you know, four or five pages long each sort of thing. I like that. Um, yeah. Snack size lessons. I mean, the whole book is a hundred and there's a hundred pages, a hundred and what?

126 pages. Including front matter. So, um, you know, 20 lessons. They're not very long. Each of, each of these. And the second one, the second lesson is called defend institutions. Yeah. Yeah. And, um, he talks, he talks about the need to be able to, um, defend institutions because we expect that our institutions will be able to defend themselves, but of course they can't.

Right. And they don't. Yeah. And so it requires an engaged citizenry to defend the institutions, otherwise they will get undermined by, you know, the lessons here are, you know, of authority, authoritarian regimes come in and they attack these institutions and de legitimate. Them. Yeah. And so unless you know you've got an engaged citizenry who are prepared to stand up and fight for those institutions, you can and will lose them.

Right. Speaking engaged citizenry, you reminded me that, uh, my engaged citizenry today is I'm hosting the first door knock party for the. Uh, coming election, hopefully where we get our seats, preserve our seats and our part in government because we are Greens in government. It happens in Australia.

I know, I know Labor don't want to say that the Greens are in government, but we are here in the ACP and unless I get out on the streets and doorknock, then, you know, That might not be the case. So I'm going to like carry on if that's all right, because we, you know, we could talk for hours about that because it's really worth talking about.

Can we talk about On Tyranny? Can we make that an episode? Because that looks really good and I, I should read it. It's a great book, um, but it's one of those ones, as I sort of said at the top there, um, I read one lesson, and then I go and I think about it for like a month. We could read it all year, because also it's scary, and like reading a lot of them all at once may be not so good for one's mental health.

Yeah. Um, but it, you know, it's, it's a challenging book in that it really makes me think about the way in which I engage with all of this sort of stuff as well. So yeah, you know, am I being lazy and complicit in some of this sort of stuff? Maybe. Maybe. Maybe I'm not defending it. But on the other hand, Capitalism Keeps Us Very Busy, Jason.

Yeah. Yeah. One of, one of, one of its tactics is to erode the amount of time that one has to do door knocking. So that one feels tired and one goes, I won't, but you know, that whenever I feel that feeling, I'm like, this is when you have to dig in harder. Right. That's right. That's right. Yeah. And like kudos to those of you outdoor knocking for the liberals, for the labor, labor party, like the problem with any kind of democracy is when it becomes too easy for one party because they gerrymander it or whatever, and then you get real problems.

So I actually come to treasure those times on the hustings when I'm talking to my political opponents in a civilized way without guns pointing at each other, because that's democracy, right? Yes. It should be a contest of ideas. If you don't agree with the Greens and you agree with Labor, get out there, knock some doors.

We've got 12, 000, we have to knock 12, 000 doors, Jason, before October. Hence why we're squeezing in podcasts whenever we can. Okay. I'm going to get on like, so slide five, I've just got a little bit of guff at the start of the paper, just so that I like to say what the paper isn't, what it is. And what I, when you're asked to write a paper like this, it's really like, where do you start and where do you stop?

Like, like write me a paper about large language models and researchers. Like it could have been. So what I've done is just limited it to only things that every researcher in every discipline could do with desktop tools, not what they're making for themselves or specialized software that they might use or any of that sort of stuff.

I gloss over that. So a lot of it does really concentrate on image and text for that reason. So I've just got my caveats there on slide five. And so on slide six, I did. Go to eight use cases in the paper, but we only have time to talk about two of them, I think. And so we're going to talk about, and also there might be the ones that listeners are most interested in because these are ones you can use yourself and play around with.

And I'd be really interested to hear what people do with these kind of prompting techniques, basically what they are asking the machine to do things and see how they, how they work out. So I've got two use cases, which is, I'm just calling language transformations because it's not exactly copy editing.

Um, it's not exactly, it's more than that. It's, it's actually getting the language and treating it like a, a lump of clay that you can do all sorts of things to, that would take a long time if you were doing them yourself, but when you do them with the machine, those. Um, and then also using chat, chatty as a critique partner or a sounding board to help you develop your ideas as a sort of thought partner.

So on slide seven, I've just got straight out of the paper, what I say about this use case. So, you know, the, Of course, you can just put your text into chatty and say, or co pilot or whatever, and just say spell check this or correct my grammar or make it longer, make it shorter. Um, and it's quite good at just doing that without much instruction.

So unlike something like Grammarly. Where you're having to be there line by line and actually, um, participate in the process with it. And the good thing about Grammarly is it also teaches you why it's making a change. It gives you the general rule and gives you examples. So I think if you're going to do that kind of language transformation.

If you're still learning your voice and learning to be an academic, I think Grammarly is a better tool because it actually teaches you the principles behind it. Um, but I, but you can if you're a fully fledged writer like me and look, I've been teaching writing for nearly 20 years. So I'm, I think I, I do need to practice my craft.

Yeah, I've written a couple of books on writing. I do need to practice my craft. And by the way, I was not a good writer. I left school and became an architect because I hated, I got terrible marks in English and I actually hated writing and I wasn't very good at it. It didn't seem to matter how much people tried to instruct me.

I didn't get any better. It wasn't until I started teaching myself out of books that I got good at writing. And they do say that the worst students are the best teachers. And that's because they know what it's like to struggle. So I think. I, I've got a, I'm a, a good writing teacher, I think, because I understand how people, uh, the dysfunctional nature of like, sort of not just not being able to fix it and not having a person tell you that one thing that you needed to know.

Um, so, uh, you can also just. Just adjust reading level, you know, you can say, write this text for, um, a 12th grader, write this text for a 7th grader, write this text and it will adjust the, the complexity of the words up and down. And I, I do actually really enjoy that. Or it will adjust it for a particular audience.

I'll talk about that more in a moment. Um, it's good for doing things like taking a complicated academic paper that you've written and turning it into talking points because our written language is different from our spoken language. Spoken language, it doesn't have things like left branching syntax.

That's hanging clauses before you get to the main subject. Um, verb, object, if you're technically inclined, or it's all that throat clearing stuff that happens before you get to the point of the sentence. That's very common in academic writing, um, that it takes away all of that. You don't have to instruct it.

You don't have to know that. To to get it to do that. See, I know how to do that. So I can go into a piece of text and just get rid of all those hanging clauses at the front. 'cause I know that they're problematic when I talk to 'em. A lot of people don't have that level of technical knowledge. Yeah. So you can just say, make these into notes, spoken notes, and it will do that for you, which I think is really handy thing actually.

And then the more technical knowledge you have about those kind maneuvers left and right, branching syntax, you know, subject verb objects. Theme, rain, transition, all that kind of stuff, um, the better it is. And, and it does actually understand that kind of technical language. So you can talk to it if you do have that capacity.

A lot of people who've learned English as a second, third, fourth, or fifth language have better skills in that technicality than the native speakers doing, because they've had to learn it from scratch. And so those people who are listening now, you probably can. Tell it stuff that a native speaker won't think to say.

So you should use those skills, try it out. Um, there's also a growing market of custom GPTs now in, in chatGPT. Um, so you can upload a body of knowledge. So I've made a thesis, Whisperer GPT. I gave it, um, How to Tame Your PhD, which is a book of my blog posts. Um, which is about 42, 000 words of text. Um, and I gave it some of the stuff that we wrote about text expander, not everything, but I, like a blog post or two was in there, Jason, so it does know about text expander and things like that.

Um, and then you can also design prompts. Yeah. And then you had a play with, um, um, thesis, whisperer, chat, GPT. You can find it if you're in open AI by searching. The GPT store, it's thesis whisperer, two separate words, if you want to find it and talk to it. What did you think, Jason? I did basically, that took five minutes.

Do you know what? That's all I spent on it. I think that this is. Well, certainly for me, right? This is an area that is a slow burn. Like it was, there was unexpected value in these kind of specific GPTs that you put there. So, so with yours, I went and had a bit of a play and it was I, there was, you helpfully had given me some prompts.

At the start, right? Like, you know, these four things that you can kind of kick off with and one of them was, you know, how do I manage the stress in, doing a PhD, um, and it was really, it was bloody good, right? Like I just, I just clicked that thing. I was like, Oh yeah, let's see, let's, let's see, right?

Like, you know, how good is it? Some bloody helpful advice there about how to manage things, including, um, Text expander, and I think I'd send you a text, uh, text message to say, I'm like, I'm so impressed that you managed to get text expander into that thing. Um, uh, but you know, I then went and I had a play with a whole bunch of other ones as well.

So like you, I, um, subscribe, I, I pay the monthly fee for Chatty G4, I think. Yes, it's four. Whatever. Yeah. Four. Um. And so that gives you access to all of these kind of bespoke GPTs, um, as well. And so I, I, um, looked at, well, there's one thing that you can get logos designed and web pages, web pages. Yeah. Um, there's one there for data analytics.

So I had some data and I, um, you know, I made sure it was clean data. It wasn't anything super secret or anything like that. And there was, there was certainly. Um, I don't, I punched that data in there and it came and I said, I want you to do this. Can you do it? And it came back and it came back with answers.

And I'm like, love that for us. Yeah. Yeah. Yeah. Statistical analysis. Yeah. Like my son's doing a, doing a course on, on survey design and statistical analysis. And his, his lecturer said, uh, you know, they're teaching RStudio. His lecturer's like, this is, this is the money class. And Brendan, of course, is like, I'm listening because he's very motivated.

And, um. And I said to him, right, yeah, it is the money, but like, this is the money. What this class is the money in is knowing what to ask for what problem to solve and what multiple kind of, not for actually the doing of it, because that is going to be machined, but it's going to be, can you assess the output?

Can you ask the right questions? Can you troubleshoot problems? And he said, and he was like, yeah, mum, because he did the eye roll. Like, we love, we love how our kids know more about this stuff than we do. They know so much more than us, Jason. Like, we should just bow down. Yeah. I mean, my, my problems are smaller.

My child is smaller than yours, child. Yeah. They get more opinions. Younger. They get bigger. Yeah. Uh huh, yeah. But, like, I, I was saying, like, blah, blah, blah, blah, blah, do this thing, do it like this, blah, blah, blah, blah. No, Dad, I'm going to do it this other way. And like, just, he wouldn't budge on it, right? He wanted to do it his way.

Yep. And in the end, I said to him. Look, how about we start with world's best practice, and then you can adjust from there. Like you can work, you can work like, but let's just start with world's best practice. Right. And then away you go. Like if you, you can make it worse, you know, as much as you want after that.

Right. It was just like, ah, Jesus, the ego of it is quite stunning. I, yeah, mine's starting to grow out of it a bit. Like he actually sat down to me and talked about mortgage. mortgages with me. I was shocked to find myself in this conversation. So, you know, maybe there's hope for us. And he actually seemed to be listening to what I said about mortgages being as a person who's paid them for 20 years.

Yeah. That's it. Yeah. And you know. You know what the, yeah, yeah. Okay. So on slide eight, I've got two pictures there. One's from Dali too, where I, the prompt was, give me the photograph of a nerdy research intern. Do you want to describe that picture to the listeners? Uh, it's black and white, um, and it is a young man.

Uh, I think quite an attractive young man, quite an attractive young man. And he's got kind of 1960s glasses on and almost like, I would almost argue an Elvis haircut. Yeah. It's almost a quiff. Yeah. Um, but he's, uh, there's certainly no lines or any wear or tear or concerns on this young person's face. No, he's handsome and he's white and he's white and he's a man.

I didn't ask for a man. Yeah. Oh, didn't you? You just asked for a research interview. I literally asked. Give me a photograph of a nerdy research intern. That's what it gave me. So. Did you notice that he appears to have two different coloured eyes? He does. I'm not quite sure if that's shadows, but in black and white, it's hard to tell.

Like, you know, Chaddy's getting out of it by Dali is one of the models that you can access through Chaddy. So, um, that's the image transformer. So, you know, so I didn't give it a gender. I didn't give it a color of skin. I didn't give it any of that. And it defaults to young, handsome white guy. Cause that's its, its idea because it's learned from us, Jason.

It's learned sexism and racism and it's learned it really well. And so what, um, the term that's sort of come out in the communities that do this kind of image transformation is called force styling, where you try to work it against its own biases. Okay. Yep. So you say, not to you, you tell it, I want a photograph of a nerdy research intern who's a woman of color.

Yeah. So the other picture there is a woman of color. Um, you'll notice something about her. She's really beautiful. Right? Yeah. Yeah. Yeah. And she looks really fresh faced and kind of cartoony, doesn't she? Yes, yes, I tried really hard, the number of things that I threw at that prompt afterwards, I'm like, can you make her look tired?

Can you mess her hair up a bit? And that's as un groomed and tired as I could make her. So even my forced styling, my really, really, really, really forced styling it, she's still just beautiful and fresh face. So. Algorithmic bias is real. It just sits there and sometimes there's nothing you can do about it.

It doesn't, it will do what it's going to do. It's the limitation of the tool, but often those limitations are invisible or unstated. Like it doesn't say, Hey, look, my idea of a nerdy research intern is a man. Is that your idea of it? Doesn't ask you those questions. It just does it. So you really don't, you know, this is where again, that expertise is so important, again, that critical faculty to try and see what the biases are doing to you, um, one thing that I've noticed that Chatty G does lately is when it gives you something back, it describes what it's given because what's happening is, um, it's taking your prompt and behind the scenes, it's recrafting it.

So it changes your prompts before it feeds it to the machine. That's the big difference between chat GPT 3 and chat GPT 4, is that it's sort of cleaning you up. Yep. And so then you're like, no, no, I want it to be dirty. It won't. Yeah. It'll still clean it up. All right, so like those guardrails, okay, they stop porn and they're designed to sort of stop sort of copyright violations such as, I'm not even going to go into that topic because it's so huge, but but suffice to say it doesn't do a very good job because it's still got heaps of assumptions built into it.

So it's always good to be aware of that. Yeah. And it's really important to understand what those rules are and to recognize that not all models have the same rules. Yes, exactly. So there was an article in the newspaper just the other day it was talking about Gen AI and a banker in Hong Kong was convinced to.

Make a transaction of some 23, I think 23 and a half million dollars. So what had happened was he this banker had got this email from the chief, purportedly from the chief financial officer of this bank saying, Hey, I need you to do this secret transaction. And he was smart enough or they were smart enough to go, this looks dodgy.

This email doesn't look legit, right? Like why would the CFO want to make this dodgy secret, super secret? So, um, cleverly, I jumped online for a video chat with the CFO and some other executives. Yeah. Yeah. The whole thing was deepfake gen AI. Amazing. What in real time? In real time, convinced this guy or this person.

I don't know. I don't know if it was a guy or a woman. I'm, I'm making assumptions now. Convinced this banker that, um, that it was legitimate and the banker then went ahead and authorized the transaction of 23 and a half million dollars. I hope they didn't lose their jobs. Cause what are you meant to do with that?

Yeah. Right. Like tried hard, tried their hardest to verify the. The, what it was that I had to do, but yeah, like, so that model of gen AI, where they're, they're taking the voice of, and the images of a particular person and then manipulating them to be used in nefarious ways, that's got a different set of rules.

Boundaries than what the, the public facing Chatti G and Copilot and Dali E and those have as well. So you have to be, you have to be aware of what you're dealing with and those, um, you know, it's, it's amazing that we're going to have a, um, an ecosystem of. Bespoke GPTs to be able to choose from to be able to solve particular problems.

But each of those is going to have a whole bunch of design parameters around them that you probably are not going to be all that particularly aware of. And so that's why like telling you what the design parameters are is so important. Um, I did notice on threads the other day, the head of threads actually published a whole article.

How does our algorithm work? And it was in plain language and detailed and I thought this is a good start, right? Like if I'm going to get an algorithmic feed of what I'm supposedly interested in, how does it determine whether I'm interested or not? Like, um, yeah, so absolutely like we're entering a new era where trust is going to have to be done in all different sorts of ways.

Yeah, the, um, I was on, I was on Instagram the other day and there was a deepfake video of an attractive girl who was very much in the style of the one that you've got here on the slide who was saying, you know, before you do anything else, you should watch these five TED talks. And so, and then, you know, went through and just listed the, the five TED talks and what they were about.

But, you know, it got me thinking, right, like there's no humans involved in any of that anymore. Yeah, they, so somewhere, someone spun up, a deepfake video, a gen AI video of an attractive person and then they've cobbled together these five videos these five TED talks, put it together as a post, stuck it on Instagram presumably to have a it monetized in some way, shape or form, right?

Yeah. No content creation anymore? Well, this is the thing. What is content creation anymore? Is it a person who's been teaching researchers that for over 10 years? I don't know, I retired all my classes because I was like, I don't know what I had to do when Twitter, you know, um, it, I mean, Twitter, Twitter really just exposed the power state of social media for me.

Like it was just sort of last straw, but I haven't spun that class up again. Cause I'm like, do we spend our time doing that? I don't think so. I think we just become experts and have the journalists come to us. Like they are increasingly coming to us. As we were saying earlier in this conversation. So on slide nine talking about this sort of forced styling a lot of people said to me, Oh, like chat GPT, it's not very good, is it?

I'm like, well, it's not good. Cause you're holding it wrong. I'll get that. I mean, I know it sort of sounds not safe for me and me and my husband do like to say, you know, blah, blah, blah. And they said I was holding it wrong. Um, but it was when in the, I don't know if you remember the iPhone four when it was.

It had these like, um, receivers on the side of it or something. And people held it in a certain way, it would block wifi or something. And Steve Jobs said, they're just holding it wrong. Like, it's like, no, it's up to you to design it. Yeah. Yeah. But anyway, I do think that people who think that chat GPT is no good.

Air quotes doesn't work for me. Air quotes will never replace me. Air quotes. They're just not using it properly because it can and it does. Um, would you like to read this quote, which is from my favorite writer on the subject of Gen I is Ethan, Ethan Mullick, um, from Wharton and he's one useful thing, substack, and this is my favorite quote about, um, Thank you for throwing slide nine under the bus, like under the bus, like that nine because, um, as a, as a person living with pneumonia, being able to speak and read is great.

Right? , you want me to do it? I'll read it out. I'll read it out. Please. Thank you. I forgot you are, you're doing such a good job, by the way. Oh my God. Like I'm, I just forgot you had pneumonia. Okay. I'm dying here. Can you, can you do it please? Ethan says, there's one major trick that will make your conversations work better.

Provide context. You can, brackets, inaccurately, but usefully, close brackets. Imagine AI's knowledge as a huge cloud. In one corner of that cloud, the AI answers only in Shakespearean sonnets. In another, it answers at the mortgage broker. In a third, it draws mostly on mathematical formulas from high school textbooks.

By default, the AI gives you answers from the center of the cloud, the most likely answers to the question for the average person. You can, by providing context, push the AI into a more interesting corner of its knowledge, resulting in you getting more unique answers that might better fit your questions.

So this is called in the trade, we've talked about forced styling. So this is called priming. So on slide 10, um, you can prime it overall in the settings. You'll see there on, um, a screen grab from my priming of Chatty G for me. So I've asked it, I can't, it's too small for me in these classes, but I've asked it to call me professor.

And sometimes Inga, because it's nice when something calls me professor. I want it to be analytical and try to be neutral. I want it to be my research assistant and I want it to help me with academic papers and talks and so on. So it knows who it's meant to be. By providing a couple of words in there, I've pushed it into the area of its word cloud that is to do more with academic analytical language rather than business casual, which is it's.

Default kind of language. And so, but you can also override this prompt, um, uh, strategically. Um, so sometimes I say, you're no longer my research assistant. Now you're a tabloid journalist and then tell it to do something. So you can strategically override it. And the, and the idea is when you go in to do any prompt, treat it like an enthusiastic, but naive research assistant.

It's prone to making mistakes if you're not very exact in what you tell it to do. So on slide 11, I've got some of my favorite. Language transformation prompts. And I use these and as I did on Friday, I spent all morning working with them on fast writing technique, free writing, writing to prompts, the end, but therefore, um, storytelling technique from Randy Olson's Houston, we have a narrative.

I do move step analysis. I use the Manchester phrase bank. I get them to write the conclusion before they start all these kinds of things. And I say to them all morning, I want as many. crappy words on a page as you can muster. So they get like, hopefully about a thousand to 1500 words of original thoughts on a page before we start using ChatGPT in this way.

 Sort of as a. Uh, to like a potter's wheel, you throw the, the, the lump of clay, the shitty words that you wrote, but they were interesting thoughts and you put them down on paper in intentional ways and you throw them down on the, um, on the thing and you start to shape them in various ways.

So you can, um, the first one is, and wherever I've got a square bracket there, it is where you insert your own. So rewrite this text about topic, um, to make the argument about. Whatever your argument is, clearer, and then insert the text after a couple of colons. So, for instance, rewrite this text about neurodivergence to make the argument about the need for progressive assessment clearer, insert the text where you're trying to make the argument.

And By gum, it does a great job. So a lot of people were like, that's fucking amazing. Then the other one is rewrite this text about topic to make it more, insert the academic discipline in which you're in, in tone and then insert the text. Um, and that's an example of sort of priming more, more finely on the fly.

And then you can adjust the reading level about the level of this writing and you tell the machine what the writing's about. Um, you know, you give it a general topic kind of description and then insert the age group or stage. Or you can say something like based on these notes about, put the topic or subject in there, write an outline of a possible paper for the journal, put the title.

Or a journal for research educators, or I could just say studies, research and higher education. And then the text does a pretty good job, if it does a good job of that, say write me the abstract for my paper and see what it does. One of my favorite things to do is to take, I like to write messily and in chunks, sort of like I'm piecing together a quilt.

And it takes me a very long time to piece together and join the text and smooth it over and make it good, as we would say in the building trade. Um, but it's very easy and chatty. And the second last prompt I've got there, join this piece of text about topic, paste the text, with the following piece of text about topic, paste the text in a way that preserves logical flow.

If it does a good job at that, and often it does, um, I just type continue. To see what it says and sometimes it says something quite interesting. Yeah. Yeah. Sometimes it says something. I would say, sometimes it says something I wouldn't say. Um, and then, then I will take that sometimes and say based on our conversation so far, after you've done a series of this transformation, it remembers in memory everything you've talked to about it.

Yeah. Its own text and your text, which is really useful. I say write a lesson.

And sometimes in doing that, what it does is it summarizes and simplifies it enough that I come back to the text with, um, what we call in the trade, a reverse storyline. It's a very common technique in, in teaching writing, which is when you've got writing, that's a mess. You take it and you try and write a very simple version of it.

By telling it to do a lesson plan or a lecture, it will often write a very simple version of the text, but can serve as a platform for rewriting. And so the students said to me, they did this and they were kind of blown away. Some people said this is shit, um, especially with statistical writing. They found that it sort of mixed up correlation and causation, which you don't want it to.

So I think writing about statistics, um, so your mileage is going to vary significantly. Like some people were like, this absolutely blew their minds. And they're like, I never have to clean up my own text again. What the what? I could just write dot points. What the what? And other people were like, Oh, this won't work for me for whatever reason.

But now they know, right. And then they were saying to me, well, what do you do? How do you acknowledge it? Blah, blah, blah. And I was like, well, you know what I do quite simply. If I'm going to use this text, I take it into a document. I run a highlighter over it, like in green, and then I start hacking into it.

Like I do with all my texts. And if it ends up that there's only fragments of green there, it's my text again. Um, and if you've started with your own text and you're using this in the way that you're throwing clay on a wheel, I actually don't have any, any conceptual problems with that. It's just a tool that you're using to enhance and extends your creativity.

Um, on slide 12, I've, I've said, I then had to sort of say to my boss, when I wrote about this in the paper, you know, what, what should you be worried about? What shouldn't you be worried about? Cause that's what she asked. So I said, you know, it kind of does in some ways replicate Grammarly because Grammarly will do things like adjust reading level and has for years.

Language tools and MS Word will correct your grammar. It sort of looks uncontroversial on the surface if you don't think about it too much, but then there's sort of these deeply worrying kind of bang on consequences. So writing is thinking. Writing and thinking are together, so by externalizing your thoughts on a page, you can look at it, you can critique it, it, it, it takes concrete form.

And it, and I write and think, and sometimes I can't think unless I write. And I like, I think more logically and persuasively if I do it through writing. And if, if I can outsource this to machine, am I thinking in quite the same way? And also, am I, are my basic writing skills just going to atrophy, like use it or lose it?

Um, I just think back to, there was a widespread moral panic about calculators when we were growing up. So we've lived through these decades of technological change. And I remember the year where they decided just to give up and let us have the calculators and they just changed math. Right? Yes. Like no longer did you need to do mental arithmetic or anything, you just, you had to do harder math because you had a calculator.

And so there's that element of it. I think you pointed out to me, which I put back into the paper, cause I sent this to my brains trust and we had a really good conversation, the 12 people that sent back and wrote back to each other. You would have seen some of that, Jason. Some of their reflections on it were quite interesting.

But you pointed out that we often use the PhD proposal, that piece of writing, to assess whether a person can write and think. Um, and will we be able to do that in the same way if we know machine assistance? Like, can we really assess the abilities? They're bringing to the to the degree and again, back to that algorithmic bias, you know, there's no algorithmic transparency.

Those people who are not as accomplished writers who haven't developed their own voice, uh, are they just going to be nudged by that algorithm in ways that they don't even see? Um, but on the other hand, the people with. Who have less language skills, but really well developed, creative and insightful.

And I think plenty of these people all the time. People who, who maybe have never been diagnosed with dyslexia or are working in a language. They haven't had as much English instruction because they haven't been able to afford it. Jason, um, is it also a tool of equity and are we actually. Like, does it allow access for people that previously couldn't have access to this level of thinking?

Because it enables you to do that, um, that thinking through writing that you probably for physical reasons couldn't do before. I don't know. See, no, like no tool is like evil or good. There's everything shades of gray. Um, but I mean, I think the ultimate worry for me is that we're just going to become an illiterate society.

Like, at the young kids, your kids age, like where's the incentive to spend hours and hours and hours and hours like we have to become good writers? I don't know. Don't know the answer. Um, and you don't have to look very far, right? Go to a retail store and then, you know, your. Shoes or whatever it is that you're buying, uh, a certain amount of money and then hand over cash and watch the, watch, watch people struggle with change.

Right. Yeah. Yeah. Yeah. So, you know, it's 75 for a thing and you hand over 125. Yeah. Right. And, and they go, but there's too much here. And you go, yeah, yeah, I know, but I want the, I want this change. Right. And then they have to do that mental arithmetic and it's painful to watch while they try and figure out, Oh man, can I, can I give them that change in the way that they want it or not?

Right. That's so funny because when I started working at Coles at 15 as a checkout chick, right, I haven't stopped working since then, basically. And, um, there were no bar scanners. They came in the year I left. This job was like three or four years later. So it was like all key entry with your fingers and you had to make change.

Um, and when someone said, I want a 20, you know, dollar note with that, you had to like work the change and I could not do it when I started this job because I grew up with calculators and I remember the person teaching me was like muttering the whole time about young people and the reliant on calculators and gave me a few kind of heuristics or algorithms to be able to make change.

And now I can make change really easily. Like I have it, but it's a dead art. You're right. My son works in retail and he has a calculator on his phone, just does it on the calculator, you know? Yeah. But he still manages to do it. But yeah, you're right. Like, does it matter that we're losing a skill? Well, your point, you know, will that, will that skill atrophy?

Yeah, I think it will. Yeah. Like, I think, um, I think it will. Atrophy. And so we have to wrap our heads around that and sort of say, that's what society is going to look like. I think. But if we're thinking through writing, how are we going to think? Yeah. I mean, am I having a moral panic myself or will find other ways to think?

I mean, I'm kind of team human in lots of ways, but I'm a bit worried about this one. I am. Yeah. Um, there are lots of other ways to think I, um, um, I think and draw. Yes, I can draw too. I think through diagrams. Yep. I do drawing a lot. Yep. Um, so, you know, mapping stuff out that way helps me to be able to help me to be able to think.

So it's actually better in lots of ways, you know, it's clear up. We have other. Other mechanisms to help us to be able to think, but you know, the writing, yeah, like, Hey, I want you to, Hey Chetty G, I want you to do this particular thing. Can you write me a thousand words on this particular well known topic that I can't be bothered really having to go away and, and think through.

Project management, for example. Yeah, I did that the other day. Oh, there you go, right? I wrote a chapter on project management. I'm like, Ooh, I could research this or I could ask Chatty. So just ask Chatty. I didn't use what Chatty gave me. I wrote it in my own words, but I verified what Chatty told me. And then I, then I wrote it in my own words and made it contextually specific to the audience that I was talking to.

But boy, did I use it instead of doing a bunch of Google searching? Cause I'm like, what's Scrum? What's Agile? What's Waterfall really? I mean, I've heard all these words. I kind of know what they are. Can you tell me? Then I was like, can you tell me more about this aspect of it? And geez, it saved me heaps of time.

Did I do a chapter in four hours that would have probably taken me a week? Yes, I did. Yeah. Yeah. And that's where I, that's where I think that these, um, these tools will. Yeah. But they do rely on your ability to be able to effectively critique the output. Sure. And, and, and I think the way in which we're engaging with that here, this conversation is deeper than what most people will, you know, the Microsoft word version, the copilot version of this sort of stuff.

Um, that's available to the general public at large, um, might not understand the need to really stop and think about what's come back from that, uh, in a way to ensure that it's correct. Um, and I don't mean correct in terms of objectively right or wrong, but that you, what you got back actually does the job that you wanted it to do, or it was, you know, it seems, it seems right.

So funny that you should say critique your output, Jason, because it can critique you as well. And actually I think it's bloody good at it. So on slide 13, and this is a use that I use increasingly in my work. This is a sounding board. So one of the things about writing academically is you, it's a genre piece, right?

Readers have expectations and they want to see certain moves that you make in the writing. And you've got to, one of the moves that you have to make in academic writing is anticipating someone disagreeing with you and in the text planting both that disagreement, like you might think XXX. But let me tell you why, why, why, why, okay, that's a classic.

And in fact, one of the real differences between an experienced academic writer and a novice is that we do a lot of this. We do a lot of this imagining the person reading it, imagining them critiquing us and then talking back to them. So you have to make an effort to think of these. It's actually part of the creative, the most deepest, difficult creative work is really thinking, uh, what would.

In the case of a thesis examiner, what would my thesis examiner And hopefully you even have an idea who that might be, or at least, you know, what sort of person would be a good standing for that. So I wanted I wanted a guy called Juergen Strick to to examine my thesis. And so I had a little profile of him up on the wall.

And I would and I knew what he liked and he didn't like, and I would always say, what would Jürgen think of this as I'm writing it? Would Jürgen believe me? What does he need a little bit more? Would he believe that? You know, so I, that is a really important move to make in writing. And a lot of people have to have that explained to them.

I did. But once you do make that. Change. It makes your writing heaps seats better. Now you can ask chat tvb to be that reader for you, that intended reader. 

It's quite good at. You know, uh, take this piece of text and tell me if there's any logical fallacies. If you use that term, it will name all the logical fallacies in your writing, which is confronting. Um, but it will do a peer review and it will also have opinions. And if you prime it correctly, it will have actually quite useful opinions.

So recently I wrote a piece again about neurodivergence in the PhD and assessment and the idea. Staged assessment or progressive assessment. And I, I put some of that text in and said critique this text from the point of view of someone who believes PhD examinations should never change. And so it gave me six arguments that people, and I, if you're going to a conference and you're saying something controversial, it's very handy to have it give you in advance.

You know, possible objections. You can build that in. You can think about what your answers might be, especially if you're doing a PhD by now, like a high stakes examination by voice. Um, so on page 14, I've got some critiquing prompts that we go use after those last language transformation. Sometimes I say, you are a tabloid journalist.

Or you can put in, you're an economist or you're a, and sometimes you can put in the opposing discipline, like you might be in sociology and say, what would an economist say about this? Which is quite, when you're going to an interdisciplinary audience can be quite useful. Ask me six difficult questions.

And I'll see the questions. I'm like. Damn, Chatty. Yeah. I don't actually know the answer to these questions. I'm going to have to go and figure that shit out. And you can say things like, I'm going to give you some text that argues, and then you insert the gist of your argument, you know, arguing that the PhD should have progressive assessment.

Please critique this text from the point of view of someone who will disagree with this argument. So from the point of view of someone who thinks that PhD assessment shouldn't change. And then I could say simple things like, what do you think an audience of? And then instead become specialists, like research educators would think of about this text, which is about topic X.

Remember, you've got to be specific. Have I missed anything important? Like, well, that's how I did that. It, I, most of the things that it listed, I'm like, yeah, they're not really important. But one of the things really was, and I hadn't thought of it. I'm tired and I'm busy, Jason. Yeah. I hadn't thought of it.

And then a peer review, um, before you send your paper for peer review, get chatty to do it first because bloody hell did it save me a lot of time last time I wrote a paper. Um, you know, some of the things I didn't want to address, you know, but some of the things I choose to. And I when I talked with my boss about this particular use of chatty, she's like, yeah, I don't, I don't really have an argument.

Against that, on slide 15, I sort of then said, what should she be worried about? I think it's incredibly powerful and will become more so as we have more expert, like my thesis whisperer, GPT would think about a text differently than the standard GPT because it's being fed and goes to specific information first.

It can be hard to fact check this kind of output because it's opinionista, it's being an opinionista. It's not an objective fact, it's, it's a, it's an idea. And in my testing so far, the machine's at least as good as me about this. Okay. Like granted, I'm writing about something that there's a lot of writing on, on the internet.

If you're in a much more niche area, maybe not. Okay. So your mileage is going to vary. And I think that's kind of exciting to see where the limitations are. And that critique partner role, is this one that we reserve for human supervisors? You know, this is deeply human work. But if you're going there with your faculties really attuned to it, and you're going there in the spirit of, you know, bringing all your critical faculties to bear on what it gives you back, is it any different from having a coffee with someone?

And I think this is. This is actually where machines really helpful because some people really have trouble engaging their supervisors or anyone really with their ideas because everyone's busy or not everyone's interested. And and it's sort of this variation on notes to self, if you like, you know.

And then certainly there's a risk again of being too reliant on a machine if the machine misses something and, or insights, insights come by that you would have missed if you hadn't gone to talk to a human who might be more creative, maybe not your thoughts. Yeah, I I sat and sat and had to think about that one a little bit when I saw it in the paper, you know, is it any different from having a coffee with a colleague?

And I came down that I think it is different, and I think it's still, you know, go right back to the start. It's a statistical model, right? Yeah. Like it's not an intelligence. And so, you know, some of the most insightful. Conversations I've ever had have been with other humans over coffee where you start a conversation and it kind of veers off to one side and circles around very much like on the red pod sometimes, um, to, you know, and then you eventually come to something that is this kind of.

Dual meshing of ideas and interactions in such a way that you, you end up with something kind of much more creative as a result. And I, I don't know, maybe I'm holding it wrong, right? Maybe, you know, I'm not doing it. Maybe I'm not doing it properly in Chedi G. But I still think that there's sometimes those creative leaps that humans can do.

I'm not sure that the machines are quite there yet. Yeah. So, um, starting that process, yes. Right. Like let's, you know, does it, does it give you a leg up? Yeah, sure. It certainly does. Right. Like you can ask all these kinds of questions and it will find, um, it will find things that you have not thought about that you can address, like, as you were saying before, you know, tired and grumpy and, and like didn't have much time and it gave you something to consider, which is good, right?

Yeah. But ultimately, but ultimately before you send it off, you're in a unique position, we are, we're in a unique position to be able to judge our own work, um, which is what the PhD does for you. It's a lasting value add is that you become your own best critic, at least you try. At least you know when you're not being.

Yeah. But so when you get it to the point that you think like it's ready to go, you know, you do pass it by colleagues, you know, can you read this paper? I'm about to send this paper off to this particular journal. Can you read this through just to make sure that, um, you know, I'm not saying anything dumb.

Yeah, I don't think I am saying anything dumb here. But can you do that? I, I remember I was standing next to the photocopier and I was talking about my PhD with a professor and he said, blah, blah, blah, blah, blah. Have you thought about this? Yeah. . And I was like, holy shit. No. Right. Yeah. At the p Yeah.

And that was it. Yeah. And, and he, and he walked off at that point, like, and I was like, and I'm like, holy shit. And Right. That ended up, that conversation was so important to the, to my PhD in the end, right? Yeah. Like it sent me down a completely different, um, and, and it was like, so, I dunno, I maybe. But maybe it will get to that point where it's able to do that sort of stuff.

But you still, I think you still need those creative insights that come from another human sometimes to really make that big difference. So maybe this is more in the category of enhanced, like, you know, Ripley at the end of Alien 2 where she fights the alien and she's in an exomechanical suit. Yeah. Yeah.

Ripley. Yeah. Come at me bitch. Um. Yeah. Yeah. Yeah. Anyway, I'd like to encourage people to try those out right back to us. Send us a speech. Let's have human to human conversations. Team Human. Team Human. As Douglas Rushkoff's excellent podcast, which I also recommend. Team Human. Team Human. You wrote a bunch of notes.

I stuck them in this slide. Oh, um, just in the interest of democracy, I'll whip through them really, really quickly. We've talked about them a little bit through the podcast already, but I got early access to Copilot. So are you on a PC when you do this? Do you use a PC? No, I'm on my Mac. Okay. So it works on Mac?

Hmm. Yeah. Okay. I don't know. Yeah. I don't know whether or not it's, um, I'm not, yeah, I don't know cause I don't have a PC. Whisperer has got a PC and he said, all I hear is him going fucking co pilot and he said it's as bad as clippy, the paper clipper. He seems to think it's as bad as clippy, the paper clipper.

He's always suggesting that he does things in searches and he's like, just go away from me. I don't need your help right now. He said it's a bit pushy on Windows. I think it's still got a long way to go, right? They've released this thing really, really early, um, in order to be able to get market share.

And I still think it's got a very, very long way to go. So for example, in Excel, it will only look at data if it's formatted as a table. So yeah, so it has to have a specific format, right? Like, so you have to take your data and then you've got to format it as a table. But doesn't it already a table in Excel?

I don't understand. No, no, no. That's very specific formatting. It's like, it has to have alternate, uh, the rows have to be colored in alternate ways and yeah, like, so yeah, yeah. So, so really one of the things about it is. It's not like that. It just assumes the user's doing nothing but sitting back. Whereas Microsoft is so like, you will do it Microsoft's way.

That's like a good thing. And it's, the promise of this stuff is huge, but it's, it's not very helpful in helping you to understand how to get it to where it's promises. So, um, I took some. I have designing a workshop and that's coming up, we're giving, uh, next week. And so I, into ChattiG, I, uh, had this running conversation about what I wanted and, you know, what the topic was and all this sort of stuff.

And got it to the point where I was like, yep, that's great. Like, you know, we've got something useful here. And then I said, build me a slide deck that can go with this. And so ChattiG output. This outline of a slide deck, cause I can't actually, can't actually make the slides. So it just gave me an outline, you know, and it was title of this and sub subtitle of this and bullet points around that and all that sort of stuff.

I took that and I tried to put that straight into straight into PowerPoint. So like copy paste and. PowerPoint choked on it would just, just wouldn't generate it. It was just terrible. Like the copilot version of PowerPoint, the copilot. So you gave it an outline and it was just like, I don't under, I don't get it.

Yeah. I can't do this sort of stuff. That's not computer. Computer says no. Yeah. And so I. And it didn't tell me how to help it. Right. So like, so I went down the path of, okay, I need to build a slide by slide by slide. Right. So create a new slide. Copilot would create a new slide. Insert this title. It would then insert the title.

Insert these bullet points. And then we would insert bullet points right under the slide. So you had to kind of build it like Lego. Isn't that just like building a slide deck anyway? Like what's the value at? So it's basically, So, so yeah. So I was like, oh, this is shit. Right. So it didn't take me long. Like the, by the third time I'd written, create a new slide, I'm like text expander.

Let's go. It's like, I'm going to make that. If I, if I'm going to have to do it this way, then. Then I'm going to, you know, I'm just going to get, I'm just going to get Taxi Spanner to help me to do all that. Um, and then I think I got up and went and had a coffee or something. Um, and I came back to it and I was like, huh?

So I took the text from Chatty G, put it into a Word document, save that Word document to a OneDrive file. Right. Then pointed PowerPoint at that word document, like the, by copy and pasting the link from my OneDrive, the unique document link. Yeah. And I said, go and have a look at this file and create me a PowerPoint deck 40 seconds later.

No. Formatted properly. Yep. With images, all that sort of stuff. It like, it just. Presented back to me, um, a PowerPoint deck that would, yeah, would I use it? No, because my audience is discerning. Um, but, but I tell you what, hours and hours and hours of manual labor, because I don't spend a lot of time in PowerPoint saved.

Yeah. Like it just, it gave me the outline, it built the thing, it had transitions, including like, you know, fading, the titles would fade in and out and all that sort of stuff. Like it was bloody good. Yeah. It was good for that. This is what I think, it's like 70 percent of the way for most things. Which is a big improvement in terms of time and thinking.

The other, the other use that I've found it's in, um, it's in Teams as well. And it's also in Outlook, but I don't use Outlook. So I'll come back to Outlook, um, once I've had a bit more of a play with it. But in Teams, I had this really long conversation with a colleague, um, in like chat, Teams chat, and then I said, summarize this.

Or no, what decisions were made and it gave, and it said, Jason decided to do this and Colleague decided to do that. And Jason decided to do this other thing. And you know, like it's, how many times have you been in Teams chat where you've, you've made all of these kinds of decisions and then you kind of disappear into the ether a little bit.

Sure. Yeah. Heaps of times. And also. Yeah. Does Teams will do a transcript of a meeting that you have in Teams now? And so if you have a meeting, you could say, you know, cause I spend laborious hours in my bullet journal writing like lists of things to send summaries of meetings to people. I'm kind of pretty attracted to the idea that I could just tell the machine.

Summarize this. Especially what's, that's pretty like next level. It's weirdly next level on some, some stuff from the sounds of it. And just not all there. I'm interested that it's okay if you put it through its own program. So this is another algorithmic rise, right? Like, Oh, well I'll help you, but I'll only help you if you use Word.

What if I want to use. Scrivener or Obsidian or, you know, what, I've got to buy your crappy word, just paste it in there just to make you do it. Like this is pretty cool. In fact, you do that's no, no, in fact, you do, if you want Copilot Pro, you have to have a family or an individual subscription to have that.

Yeah. I do have a family subscription. But you, you can't just use it like institutional, my institutional subscription. Okay. I didn't know that you could use it on your family because I refuse to have an ANU computer. Because they won't give me the one I want and then they make it suck. So I have, I buy my own, I buy my own Microsoft family subscription and I have it for Brendan's for his schoolwork and stuff as well.

So like I was going to have to buy it anyway, figure I might as well have it on my own computer. So I could enable Copilot. Yep, you can. It's open to the public. It was like open to the public. Copilot Pro was open to the public in mid to late January. I can't remember the exact date. I might have a big play with it.

Yeah. So if you've got that, you can just, you can just fire it up. It's a monthly subscription. So subscriptions on top of subscriptions. Yeah. See, this is what I was talking about at the start of the episode. Thank you for your support. People that support me like a dollar a month and have people give me a dollar a month is actually, it's very substantial and it pays for our Buzzsprout hosting.

Actually it doesn't Jason. We need to do that bullet journal book. I taught Mr. Theis Blitzsprout how to bullet journal, by the way. That's so cute. And he said to me yesterday, Yeah. Bye. Bye. It's really satisfying crossing them off. I'm like, yes, it is. Yes, it is. That's right. Yes, it is. 

Um, I think this technology is amazing. I do think that it will help us once we get better at, you know, talking to it and holding it properly, as you say, uh, so my last little point before we kind of send you out into the world to do democracy, Um, I think that this will help work.

In that we will be able to do some jobs really fast. Some jobs that would take a long time for us to be able to do. I think this, these models will help us to be able to do some of that work, um, a whole lot quicker and that's, you know, valuable, right? Like you were saying before, did it take four hours where it would normally take a week?

Yes, that's, that's valuable, right? However, I worry about the way, if we're not thinking about this stuff properly, what that will mean for the average worker. Oh yeah. Um, you know, ideally what you would like to see is our, um, workplaces recognize this amazing, um, productivity enhancement or the ability of it, and then give us the time back.

Yeah. Four day work week. Yeah. Right. Or something. I bet that won't happen. But this is. That's right. This is an argument for the four day work week or the three day work week because like we know that knowledge work is killing us. You and I. Have really noticed its effect on our. On our health. And you and I push back against it.

You and I do things about it. I've got a lot of colleagues that just look really unwell and are really unhappily mentally, physically, spiritually in their relationships because the workplace intensity is already at a ridiculous level. So you add this, it's like petrol on a fire. That's right. So like, why haven't you done it for me in four hours, Inga?

Whereas my boss gave me two weeks, which was a short deadline anyway. Right? Yeah. But she knows I'm quick. It would be nice if this was, uh, led to some sort of de intensification of work. It would be nice. It's not going to happen. I don't think it's going to, right? Well, here's an opportunity for It won't happen unless we make it happen, you know?

Yeah. Um, speaking of democracy Here's an opportunity for all those workplaces that say they care about their employees, right? Shit yeah. Like to show, show that. Yeah, show us. Show us. I'm not holding my breath though. I know. For all of the reasons that That usually sent us into a rant. Anyway, that was a good discussion.

I'm glad we decided to do that. We didn't say this at the top, but we just got so into texting each other about it. We're like, let's just do this one. And I think, cause we were excited about it. And, um, I think that was really good conversation. Hopefully people got. value out of that for their own work and will tell us what they're doing, what they're thinking about.

Yeah, I'm going to skip, I'm going to make an executive decision because you've got to go make the world a better place. Um, I'm going to skip the what we're reading bit. I think that's a good idea because also you'll hold that over. You have pneumonia? Um, and we'll, we'll wrap up with our two minute tip.

So this is the section of our podcast. It's in honor of David Allen and his classic getting things done book. Yeah. I guess that if a task will take less than two minutes to complete assisted by Jen. Sure. You should do it then and there because it'll take longer than two minutes to capture it in your task system, schedule time to do it or yada, yada, yada.

Just get it done. We kind of take that idea, we riff on it a little bit and we look for hacks that can just generally help your, and your life, um, in one way or another. Um, so my two minute tip is a, is one that I actually, this legit, this worked for me. Um, it was, uh, the end of a Friday, I had a lot of work to do.

Um, and much of it was fiddly and that made that task seem overwhelmingly large, right? Because each little, there's lots of little steps that had to occur for this thing. I knew what the outcome was that I wanted, but I also knew that it was like a thousand steps to get there. Yeah. Yeah. Yeah. Um, I had no bandwidth, I just absolutely had no bandwidth at the end of the week, um, to do any of it.

So I did the least possible thing I could to keep it moving. I created a table in Word, like, uh, two columns, um, and it just had lots and lots of rows, right? Like just build a big table. Um, the row headings were what and how, and then I just bloody filled out the table in the smallest possible way. So, um, and I just broke it, broke all of the steps down into, this is what you have to do.

I had to, future Jason had to, had to like, just look at this table and go, okay, that's the next step. Just execute, execute, execute. So here's an example. In what? I, I wrote, take all the files from Wendy and then move them to Jay's folder. And then in the how column, it was like, drag, drop. Okay. So you got down to that kind of granular level, like.

Because that's all I could do, right? Like I, I, I could. There's thousands of ways that I could take that stuff out, extract, uh, build new folders, do all this sort of stuff. But you know, the outcome that I wanted was all of Wendy's files in Jay's folder. And then I just went through and I just like, that's all I can do.

Like, I just, I can't. I can't analyze, I can't conceptualize, I can't do anything. All I can do is just go do this step next, do that step next, do that step next. And that got me to the, um, to the point where I could close my computer down at the end of the day. When I came back to it the next time, when I did have bandwidth, I had all of these preformatted steps ready.

Oh yeah, that's where I was at with that thing. Yeah, yeah, yeah. And I was able just to smash my way through it. Yeah, yeah, yeah, yeah. Like. Dead simple, but bloody hell, it helped with the overwhelm. Yeah. Like gave me control back and, um, allow me to just keep, keep moving forward. Also made it probably quicker that next day.

So you've done all the thinking. Oh yeah. Yeah. And separating sometimes, sometimes with our work, it's hard to separate the thinking and the doing. Yep. Right. But, and so that's what makes it slow sometimes. But if you can separate it out of the thinking, then the doing can sometimes be just quite quick.

Yeah. And often, um, often I get slowed down because I overthink the doing bit. Yeah. And sometimes, you know, like we don't need to do that, but sometimes it's just like, just do. Yeah. So did you find on that table you could cross some things out? You know, just like. I just worked my way through, I just worked my way through the top, like I just top down and just went on my way through and I just trusted the pace past Jason had, you know, future Jason's back, thought it through, you were there for yourself.

That's a really nice thing. Yeah. Um, I was bloody surprised at how effective that was as a technique. So, um, if you find yourself overwhelmed, you've got a big task, you know what the outcome is, but you just don't have any bandwidth to be able to do it and you know, it's going to take a while. That, that's a good way just to index your thoughts, um, almost, um, and put those steps down.

You can come back to it next time. And as you say, just smash through. And also it gets it out of your mind. And every time you think about that open loop, you're like, it's all right, it's, it's all there. I don't need to keep this in working memory anymore. I can let it go. I can, especially if it's the weekend, you can actually turn off, which we now have the right to apparently that way.

Did that go, legislation go through the right to switch off? Bloody hell yes. It went through, I think, did you see the right wing press? Oh, I was about to say. Yeah. They were like, oh, that's, you might not have a job. Impact productivity. Yeah. , um, . I only, that's an excellent two minute tip. I would only say that last week we, we, where we did we record last week or the week before?

We did last week. So we are like, we haven't had actually like. Um, to actually do, and I've just been busy most of it, but last week, which won't be last week because it'll be like a month ago by the time I release this pod, um, I said that I did all my library cataloging. I got lots of letters. I got images sent to me on, on, I witnessed you, all the people out there who decided to do their libraries.

So I feel quite pleased that I kind of kicked a stone off the mountaintop. People like, yeah, yeah, I'm going to do that. And then sent me pictures of their library books. It's amazing. Um, cause I think I've maybe tapped a zeitgeist there at the time of year or something. Anyway, I did say library thing, which is an amazing, also has a phone app, which I knew.

Just to put this in context, but did I go ahead and buy the little old fashioned barcode scanner that looks like a cat slash sex toy? Yes, I did. Um, did I not actually look at the app properly until someone on BlueSky, thank you Air Minded, on BlueSky pointed out to me that there's a barcode scanner in the app?

No, I did not. Uh, did I open up that barcode scanner and did it work instantly on all the books that I aimed at? Yes, it did. Sorry. Sorry to Abby, my niece. For making her type in titles and ISBN numbers painstakingly when she could have said they're on her phone, aimed her phone at barcode and it would have imported it like that.

So if you're going to do your library thing thing, download the app. Um, and it does add more fuel to the fire that Abby said, why don't academics ever take the price stickers off the books? And I said, that's because of our tax later on, you know, like I just made up some shit. That's not really why. It's just because we're lazy.

It is the reason to take the sticker off though, because it gets in the way of the barcode usually. They put it over the barcode unhelpfully. So anyway, that's my two minute tip. Take the stickers off your books. Use the phone app, library thing's awesome, I'm loving my library, loving it, so just do it. Good job.

Well done. There's organized libraries all around the world now. I know, I feel good. Thanks to you. I feel good. Excellent. Good job. Impact. Exactly. Lisa Swisbra, that is impact. Yes, it is. Um. Thank you listeners. Thanks for listening. We love reviews. We're going to wrap this one up now. Um, we'd love it if you could leave a review on Apple podcasts.

Uh, we read everyone and we use them to actively shape our show. If you want to join us with a question, a great way to do that is to record it via our speak pipe page. And you can find that at www. speakpipe. com forward slash. Thesis Whisperer. We'd love to hear from you. If you've got a problem or work problem you'd like us to tackle, we're all up for that.

The best place to see us talking, um, to each other, Inga and I is, um, on BlueSky where Inga is at Thesis Whisperer and I am at Dr. J. D. Um, you can both find us on, I know, right? Yeah, it felt good. Felt like old times. I know. I know. And then she goes, you know, we're at the same institution and we've only ever caught up once.

I know. Sorry, Chin. Yeah, I know. Whose fault is that, both of you? Yes. I'm looking at us. You're looking at you both. Chin. Yeah. Yeah. Um, you can find us both on Mastodon. You can search for Inga on Oz. social as at thesiswhisperer, and you can find me as at Jason Downs on ravenation. club. And Inga is sadly addicted to threads.

And of course, as per all of her things, nice way to, um, to brand yourself, like, you have just like, at thesis whisperer everywhere, you know, Jason, one of the first conversations you and I ever had. Was about branding and you said you should make this thesis whisperer thing a brand and so I did and you did it really well I can thank Jason for that listeners.

Cause Jason was my little business whisperer back then like you said affiliate marketing And I've got many free books from Amazon that way over the years Yeah Because you said affiliate marketing to me once in a coffee and you said you need to make thesis whisperer a brand so

And we're at the point, I think, um, now where I had to revisit all those conversations again, cause the, everything's moved on. Right. Ah, sure. So, yeah. But Pieces Whisperer as a brand is still strong. It's still strong. It's still working. Yeah, it's still working. Thank you, Inga. I loved today's conversation about Gen AI.

Yeah. Um, I'm sure we'll come back in 12 months time and look at this again. Yeah, have it again, just make it an annual, like second pod in the year. Not that we ever do anything regularly on the reg. That's the joke about it really, isn't it? That's the joke. Please look after your pneumonia. Yeah, yeah. By the time we release that, Jason will be recovered from pneumonia because you won't be dying from it.

Hopefully. No. Um, it's, yes, I'm going to go and take some antibiotics now and I'm going to also go and have a bit of a snooze, I think. I think you should do that. All right. Take care. Have a great weekend. All right. See ya. Bye. Bye.