Writing the Future: AI, Education, and the Craft of Critical Thinking with Leon Furze

0:03

Forget trying to keep up with AI. It's moving too fast. It's time to think differently about it. Welcome to the AI

0:10

readiness project hosted by Kyle Sham and Anne Murphy. They're here to help you build the mindset to thrive in an

0:17

AIdriven world and prepare for what's next.

0:26

It's a bop. It's a bop, everybody.

0:32

I I don't know if we can change our theme song. I thought we were going to, but now I'm like attached to it.

0:38

Yeah, it's kind of nice. We can change it, you know. But, you know, it's a thing. It's a thing. Good to see you.

0:44

I'm used to her. I know. Exactly. Good to see you. How you doing? I'm doing well.

0:49

Sounds like I met I missed a banger at the AI salon last night. It was a It was a really good one. We

0:56

had we had uh Mike Lavine from Movie Flow, which is it's like a it's like a

1:02

much more it's like a better thoughtout um

1:08

LTX studio, but instead of being built by technologists, it's built built by movie guys. And so we had him. And so we

1:15

were talking a lot about that. And then um you know, Liz Miller Gersfeld co-hosted Leh Leah is moving to

1:21

California. So she's going to be out for a few weeks. And uh so so Liz hosted and

1:27

and Liz is doing all sorts of really amazing work and she was talking about having to pull in all these different

1:33

tools and things like that. And so that's kind of what I I want to tee up today. But how about you? What's what's your week been like?

1:41

Well, a highlight I'll share is that today is the final

1:47

class of our first ever a historical first um of our f our first cohort of

1:54

the She Leads AI certified AI educator program. Oh wow. So as soon as everybody Yeah. So

2:03

all all that needs to happen now is people finish their internships. So, they finish the coursework, uh, they

2:09

finish their internships, they submit their media kit to us, so we know they

2:15

can go out and get hired. Um, and then they all graduate and they're officially

2:20

certified, I think, for a year. We don't haven't quite figured that part out yet, but yeah. So, we're open for business.

2:28

Nice. I love it. I love it. So, given that, how how readyit for AI are you this week

2:36

now? Now, now here's a question I have. If you're providing the certification, are you certified?

2:44

Okay. I figured that I just get to just wave a magic wand and I'm certified. So,

2:50

I think you get a waiver. I think you get a first year waiver. Next year, you're gonna have to reertify.

2:58

Oh, yeah. I'm gonna charge myself some money and make myself take some tests.

3:03

Yeah. Yeah. Yeah. How How are you doing? Where's your head? Where's my head at? Well, I mean, I'll

3:10

mention a little thing I noticed today, but I don't I won't want I won't dwell on it, but I want to share it in case

3:17

anybody else resonates with this. I've noticed recently I have a little bit

3:23

less tolerance for my just regular everyday foibless be because I can do so many things with

3:30

AI and I feel so smart and I'm like living at my super, you know, my

3:36

superpower era self today in the mail,

3:42

right? I can't do like we finally got um uh I've bought 86,000 microphones to use

3:49

when I'm making content. I finally got one. I've been doing content for like four or five years. Why

3:55

does it take me 86 microphones to find one that works? That's not normal. I can

4:00

do so many other things. So, I feel more impatient with myself now. I don't know

4:05

about other people but no I listen I think I think dependence

4:11

dependence on AI tools is a very real thing. Um and you know for me it's not that

4:17

it's funny that there's there's that trope out there and you know there was that MIT study that was this kind of flawed study that that said that you

4:24

know a AI makes you less smart. It lowers your you know critical thinking. Um, basically all it all it proved was

4:31

that, you know, if you don't use your critical thinking, you don't use your critical thinking. Um, it it didn't say

4:36

AI made you stupider in the end. But but um but yeah, I like what I what I find

4:42

is that I think the thing that I'm addicted to with AI is it allows me to

4:50

live in that higher plane of creative thinking where I don't get pulled down

4:56

into the into the weeds, right? And yes, I can just sort of live as a like as a

5:03

producer, like as a puppet master, right? Yeah. And I and I love being in that place.

5:08

And so when I encounter things in the world where I now have to do the little tactical details, it just feels like a

5:14

waste of time and it feels like a waste of my brain power and it feels like a waste of energy.

5:21

And so that's the thing for me like I don't feel like it makes me dumber. I just like I you just said it.

5:27

you have less tolerance for the for that kind of activity in the world,

5:35

you know. Yeah. And I'm I'm very happy to be living in our little AI bubble because I

5:42

can't imagine being in the work world right now working with people who don't use AI and that having that mean that I

5:49

have to waste my time because they don't use AI. I would be such an insufferable

5:56

colleague if I was in a normal business right now.

6:01

Absolutely insufferable. There's a I think there's a delay on my end, Kyle. Is there a delay on your end?

6:07

No. Well, I I see a delay delay for you a little bit. Yeah, it's your your machine sometimes gets a

6:13

little tired. Yeah. Um, so while you reset your

6:18

machine, I'll I'll I'll share kind of something something that's

6:24

so a couple of things going on. Tomorrow, GPT5 comes out. I didn't know if you knew that, but tomorrow at 10 a.m. Pacific. And

6:32

it looks like there's going to be GPT5, GPT, GPT5 mini, and GPT5 Nano, which is

6:40

fascinating because that could be something that runs on a phone is is where my head goes with that, that they

6:46

might have a local a phone based model, which could be amazing. Um,

6:52

and a bunch of other things came out, including a thing from Google called Genie 3, which is you you invoke 3D

7:02

worlds into existence that you can then go interact with and walk around in. And like the one of the examples they showed

7:08

is you can like paint a wall with like a paintbrush and then you turn around and you turn back and the things you painted

7:13

on the wall are still there. like these are full-on worlds that you can walk through at 24 frames a second, interact

7:20

with the physics of the world. Um, it's not out yet, but but we're entering a

7:27

new era. And then, you know, combined with um last week I did the back to basics um AI crash course

7:36

and what struck me the most is how complicated it is to just know the

7:42

basics anymore. that that knowing just knowing what chatbt can do

7:49

y is kind of a herculean feat at this point. Just knowing how to how to how to

7:55

use different image models and which one has which personality like which one's good at text, which one's good at

8:01

images, which one's good at logos. That's all things that you have to learn, right?

8:08

and and and then as you learn them, the skill when you apply this in the real

8:14

world is not any one of those tools individually, but just having them as kind of a like a

8:21

pallet, like a painters's pallet, where your job is to maintain the the vision for what you're trying to

8:29

accomplish and then be able to like an like a conductor in an orchestra or, you

8:34

know, painter, pick the right colors, pull in the right tools at the right time and use them. And so you need to be

8:39

nimble enough with the individual tools to not get bogged down in the details, right? Yeah.

8:45

To be able to bring them in. And so for me, the importance of critical thinking and

8:52

of adaptability and kind of resourcefulness are the really important things. But but

8:59

the new the baseline for for what it takes to be able to be in that space is

9:04

you kind of have to know a fair amount of tools. So you have to get there. And it and it's it's not as simple as it was

9:11

when we started this game. No, it's not as simple as it was when we started this game. But along the way,

9:18

we've learned a lot of things. We've learned you don't have to know all of them, right?

9:24

We've learned. You can narrow down. You can narrow down. We've learned that

9:29

uh it's it's uh a trap to try to learn them. It's totally a trap.

9:34

We have learned that. You and I both we both learned that. Um one thing that

9:39

I've learned that I I bet other people have learned and if they haven't, this is a pep talk then is that the tool is

9:47

so much more complicated in your head than it is when you get your hand when you roll up your sleeves and get your

9:53

hands dirty. It is always like because it's so like facemeltingly

9:58

like advanced. You're like, "Oh, there's no that you must have to be just a mad

10:04

scientist to be able to make that work." Now, obviously, there are some things that are like that, like hugging face

10:10

and whatnot, the stuff that, you know, our genius friends use, but like you know, you open up V3, you're gonna

10:17

figure it out. It's not going to take a day and a half to figure it out. Yeah. It's not as hard as it seems. It

10:24

seems otherworldly, but it's really just like any other skill. Getting your hands

10:29

dirty. It's get your hands dirty. And each individual tool is relatively simple. There are some exceptions. You take

10:35

something like Leonardo or some of the video tools have so many bells and whistles that it can be overwhelming.

10:41

Um, but I think just the the general capability of being able to generate an image in in an image generation tool

10:48

relatively the same across tools, but you have to kind of play with them enough that you you start to internalize

10:58

which ones are better at what things, right? Some people really love the writing style of Claude, some love chat

11:05

GPT, some love the research of perplexity. And and that's something that every I think everyone's going to

11:12

have a different version of the mix, right? Yes. What are their go-to tools? And there's no wrong answer,

11:18

but I think the I think the art right now is getting competent enough at the individual tools that you can kind of

11:25

forget about them and just think about the problem at hand. Who are you doing this for? How are you trying to affect

11:31

them? What does success look like? What do you know? And then okay, let me start

11:36

with chat GPT. Okay, now I'm going to run over to Perplexity and do a little

11:41

research or Manis or Gen Spark and make some slides, whatever it might be. It's just that ability to kind of nimly

11:48

navigate those tools, but but maintain the blinders for the idea

11:54

that I think is the art and I and I don't think it's I think people who will be good at this are people who have

12:00

already done that, right? like creative directors have always had to wrangle multiple

12:07

resources to, you know, execute a vision, right? That that they hold

12:13

and they might have some artists that are stronger at one thing than the other. And they they sort of know who in their who in their bullpen that they can

12:20

deploy. And that's I think what we all get to be now. But if you don't have

12:26

those skills, I think that's something to develop. So I absolutely so

12:34

I haven't seen the data on this maybe you have but anecdotally

12:40

my sense is that lots of people who are adopting AI more quickly than others are

12:48

in some way neuro divergent and particularly with ADHD

12:55

AUHD um I see a lot of us getting so much

13:01

relief so fast and also so much dopamine that it's just we're just addicted to

13:08

it. Right. Sitting there crack, right? Exactly. But I think that our

13:14

ability to hold on to multiple things at one point in time is part of what makes

13:21

us really good at it. like we can live in that like chaotically three-dimensional space.

13:28

Even me, and you know, I mostly do boring stuff with AI. Even me, um, I'm

13:35

able to bring in things from lots of different areas and make it and weave it all in. And I think it's because I can

13:41

keep I I don't have to write anything down. Like, it's just it's all there. Yeah. Yeah. It's if if you've got a

13:48

brain that's good at synthesizing things, taking small amounts of information and being able to quickly

13:54

draw a conclusion to that, I think that kind of brain is really good at, oh,

13:59

okay, I need this tool, I need that tool, I need that tool. To the extent that you're more deterministic, you're

14:05

more like, I've got this thing that I do, and this is the thing I do, and you want to stay in a predictable lane, this

14:11

is this is likely going to feel more uncomfortable for you. It doesn't mean you can't do it. And it

14:16

doesn't mean you can't go deep on a subject, but it does mean that if you only focus on that subject, you you cut

14:23

yourself off from a bunch of other things that you might be really good at and not even realize it.

14:29

Yeah. Like Roger Scott in the AI salon, software engineer for 30, 35 years, like

14:36

deep deep into software engineering. Turns out that skill applies really

14:42

nicely to prompt engineering and midjourney. And he loved making art. And so he got really good at making art

14:48

because he took an analytical approach to prompting. Oh, right. And so and so I think that there

14:56

are gifts if if you're if you're in a deep vertical, there are gifts waiting

15:01

for you if you sort of come up out of it and go horizontal and play with some of those other tools.

15:07

Yeah. Yeah. Two two things uh I want to say. One is

15:13

that as you go through like this analysis of the tools, skills that are

15:20

going to be like the real skills of the future, I I'm utterly convinced that

15:27

basic everyday run-of-the-mill kindness is the is a ginormous strategic

15:34

advantage. If you can if you can hearken that kindness within yourself on a

15:40

regular basis while everybody else is melting completely down around us. I

15:46

mean they're going to be showing their ugliest selves, right? It's going to be

15:51

awful and and we we've seen it start on social media. All of us have been dragged recently,

15:57

right? One of my friends who is the most I mean she's a PhD. She's social justice

16:04

like expert. She's a neuro something. She's like the smartest person I know

16:10

and her her heart is like a 100% in the right place. Got lambasted by somebody

16:16

um over ei AI being evil. Yeah. And I was like, well, if she gets

16:22

lambasted, no wonder why I did. It's about to be a mean mean world.

16:27

Yeah. It Well, it's it people are in their amygdala takeover, right? they're in fight or flight and and they don't

16:34

want this. And I think I think what's happening is it's becoming more and more apparent

16:39

it's not going away. And so I think that has people in a bit of a panic. But to your point, I think

16:45

empathy and kindness I think are the way because when you get someone who's in

16:50

that space and actually hear them and say, "I understand why you're scared. I understand why you might be pissed off." Whatever it is. Can I show you

16:57

something? and you get them side of that thing and they're like, "Oh, I didn't know I could do that." Right? And all of

17:02

a sudden, they're in a different frame of mind. That's a remarkable moment. So, speaking of that, let me let me talk

17:09

about a thing that I think is is worth paying attention to this week. So, so today is Wednesday, August 6th

17:16

and: 2025

17:24

An artifact of the golden days. An artifact of the golden age of AI.

17:33

Tomorrow we get GPT5. Yesterday we got GPT OSS which is the open-source uh

17:38

version of chat GPT. Yesterday we also got an announce announcement from Google

17:45

um for this thing called Genie 3 which is you can you can prompt a 3D world

17:50

that you can then enter and interact with. I saw one video of a guy like walking down a dirt road with puddles in

17:57

it and he turned the camera down to look at his feet and you see his feet doing this and when his footsteps in the

18:02

puddle the water ripples like these are full-on 3D

18:07

worlds with physics that are being generated. This is not the the AI does

18:13

not generate a 3D world and then you walk through it. It's generating the 3D

18:19

world with persistence at 24 frames a second. Um,

18:26

we're entering a different phase. And and what I would encourage

18:32

you to do for this next week, this is the the thing to pay attention to is

18:39

that give yourself the grace that everything you've learned up to

18:45

this point may not be as relevant as it was last week.

18:51

and the skills you need to have to use the

18:56

stuff that's happening this week, you probably don't have because we've never

19:01

had this level of tool before. So, so I would say this is a good week

19:08

to get curious. Pay attention to to announcements on X and Sam Alman and

19:13

like that. Um, go tomorrow at: 10:00

19:20

announcement of GPT5. If you've got time, watch that or watch a YouTube replay of it. Not that you have to learn

19:26

anything from it, but just watch it so that you're aware of at least what's being claimed, right? Because

19:32

that's right. That's what I they're going to say sort of the most positive thing about it, right? And it may or may not do that as well,

19:38

whatever. but at least understand what's being claimed because if it's even, you

19:44

know, 75% of what they claim, it's probably going to be a very new way of

19:50

working with AI. And if you have yourself locked in or trapped back in,

19:57

you know, um, end of July: 2025

20:05

this is a good week just to remind yourself that adaptability is is probably one of the single biggest

20:12

um, aptitudes. Yeah. Um, above critical thinking. I think

20:17

critical thinking really for me remains the thing. And I'm super excited to talk to Leon because I know he's very passionate about that. Um, our guest

20:24

today. Um, but but yeah, I think this is a week to just remind yourself to be adaptable and and give yourself a little

20:31

grace that that maybe you're not if you feel like you're farther behind than ever. We all are this week.

20:37

Well, here's the good news. We're all going to get this at Wait, are we all getting it or is it rolling out over and

20:44

it's just one of those fakey fake things where get two months? I think they're going to

20:49

launch it. Okay. So, I mean they might they might start with

20:54

um they might start with the pro tier just as a slow roll out. Yeah. But I think something will launch

21:00

tomorrow. Like I think we will know tomorrow what it is. Okay. And if it's good and like how far up on

21:06

the benchmarks it is and does it actually unify all the different models and you don't have to pick models anymore. There's a bunch of things that

21:12

are promised that that are sort of uh speculated and and we'll find out

21:17

tomorrow what they are. And then, you know, we get to spend probably the next six months figuring figuring out how to

21:23

use this thing. Yeah. I wonder if they're going to do the same thing to us. Just treat us just

21:28

like fullon guinea pigs. Not tell us anything about how to use it and just see what happens when we bang our heads

21:33

against it enough. Yes. Promise. I promise. I promise you that.

21:41

Well, in fact, in fact, um, for the open- source model, they teamed up with

21:47

Kaggle, which is a data source company, and there's a half a million dollar, um,

21:52

bug bounty for people to kick the tires. So, so if you want to if you want to do testing on the 20 billion parameter

22:00

open-source thing that they launched yesterday, there's prize money if you find a bug. So, but yeah, they are

22:07

absolutely putting these things into the world. Well, I mean, here to be to be

22:12

fair to them, they've got a company full of the smartest engineers in the world

22:19

inventing like reinventing how computers even can work, right? Yeah.

22:25

They think in a very particular way. Yeah. It's a very mathy, very deep, dense,

22:30

right, systems framework kind of way. They're not going to think about it like the way an animator looks at this tool,

22:37

the way a you know the way a business person looks at this tool, the way an entrepreneur looks at this tool, the way

22:43

an educator looks at this tool, right? So they cannot just by definition they

22:49

cannot anticipate how people are going to use this, what's going to work or not. So it is on us to be the guinea

22:54

pigs. Like that's that's part of the design. We should all put it on our LinkedIn. We

23:00

we are uh open AI R&D Exactly.

23:06

Cranky Gen Xer in the open. Cranky Genexer R&D for Open AI. Okay. I

23:11

have a question for you. Were you the person who asked me how we're how we process our grief over

23:19

when the tools that we've all, you know, stayed up all night around the clock for the past two and a half years when it

23:25

doesn't are those skills don't matter anymore. Excuse you or somebody else. But I I

23:31

I've definitely talked in that area because I think it is a grieving thing like you or especially if you've built

23:36

something you've built a custom or you've built a thing an automation and you're like oh this is great and then

23:42

something just comes along and now it just does that. Well like what if it was your company

23:47

for example? Well that's there's going to be a lot of companies like that right? A lot of companies it goes back to the the word of the of

23:54

the Yeah. Like like Yeah. I think I and I think I

24:00

think how you put it is right like you've I think you do have to mourn the loss of

24:07

when that thing had value, right? Whatever it is. And and then I think you need to move on quickly because I because I don't think

24:14

any of us have the luxury of kind of wallowing in those losses because I

24:19

think they're going to come fast and furious. And like this is not a it's not a judgment like like if if if

24:27

if you spent a bunch of years building a business and then all of a sudden it's

24:32

just not needed anymore. That's sad. You're not going to get over that overnight. Um but I think even small

24:37

things, I learned to use this tool. Like you said, I learned to use this tool. I learned all this prompting framework. You're telling me that's useless now? Uh

24:44

yeah. No. Oh no.

24:55

No, no, we we all need to spray Teflon on our

25:00

back so the water just moves off it, you know. Oh my goodness. No, I still tell people

25:07

they need to know how to prompt, but but knowing how to prompt actually is a a universally helpful skill. It makes you

25:14

a better communicator. prompting in general you learn. No, I think I think prompting is still valuable, but I don't

25:20

think prompting is a requirement anymore. Yeah, I think there are lots of different ways that you can get these. If if you look

25:27

at the Grock um image gen tool and movie gen tools, there's very little prompt

25:33

engineering you need to do in that thing. You can just sort of ask it for what you want and it does it. And I think a lot of people's first experience

25:40

with AI might be a tool where you don't have to prompt all that much. You just get kind of magical stuff out of it and

25:46

that's fine too, right? You know. Yep. Yep. Yeah. If they have especially if they

25:51

have a good experience that first those first couple of times if it actually works then it'll be golden.

25:58

Exactly. So why don't you tell the good folks about this organization you have?

26:05

It's so cool. Um so our mission for in she leads AI is

26:11

to unite accomplished women to advance AI for global prosperity and we get to do lots of things in the quest for that

26:18

mission. Um one of which I was I was you know sharing earlier in the show. This is our first cohort in our certified AI

26:26

educator program that is graduating. And that means that there are now, you know,

26:32

a handful of women who teach, you know, responsible AI the way give or take,

26:38

right, with their own personal flare, but the frameworks that we use. Um, and

26:44

they are capable of pitching themselves and writing scopes of work and getting

26:51

paid and and serving all sorts of communities, including through their

26:56

internship, which is they get to teach uh, young women in lowincome countries

27:01

how to use AI. So, that's been a really rewarding experience. It's just an example of the kind of thing we get to

27:07

do. We have she leads reads. We have a consulting agency. We've got our AI

27:14

academy that Michelle Muny Silva runs and that's external facing. We've got a conference. It just it went from being

27:21

a a few posts on LinkedIn in: 2024

27:28

to this. Yeah. It started really started in nove in December at Festivus at AF Festivis,

27:34

right? Yeah. And it's all happening. Amazing. And it's all people. It's all because of

27:40

people. Not I'm not doing this by myself. Yeah. You you cannot discount your vision and

27:48

your contribution, little Missy. There's there's a little bit of that,

27:54

but you are doing some remarkable work. You do you do have a really good team around you. So So congrats to all of you. It's

28:00

it's really really good. Um let me tell you about the AI salon. So the the salon um I put together with

28:05

uh Leah Fasten who's a photographer out of Boston. She's actually moving to San Francisco this month. So so she's she is

28:13

putting things in boxes and literally driving across the country. Um uh but

28:20

but the salon is a community of about 3,000 AI optimists and people that are

28:27

um playing first uh with AI. Um, and and one of the one of the roles I think of

28:33

the AI salon is is to really explore, you know, the humanity of AI, right? How

28:39

does how do you use AI to amplify who you are? How do you make sure that it's serving people, not, you know,

28:45

exploiting people, things like that? So, there's all sorts of really good dialogues. It's people experimenting.

28:50

It's people building things. And more importantly, it's people, we call it generously leading. Um, it's people

28:56

learning out loud and sharing what they know and inviting, you know, newcomers in. A lot of times with communities as

29:03

they level up, they'll get kind of elitist and it gets, if you come into it, it's kind of intimidating. Um, one

29:08

of the things I think is powerful about the salon is people are very, very generous and they're very welcoming and there's no judgment that you're new

29:15

here, right? Because I think I think we all recognize that, you know, to my point this week is new tools are coming

29:21

out. We're all effectively starting over. So, so if you feel like you're just on the ground level, to a great

29:26

degree, we all are. We all are. And that's what the salon is. And then within the salon, we've got a thing called the AI salon mastermind, which is

29:33

a subscription area of the salon for people that want to step up their game and want to get to know each other better and really support one another.

29:40

And that's that's been quite powerful. So, really love that stuff. And then the other thing that we we should tell you

29:46

about is um our AI readiness training program. We've got new new branding, a brand new look and feel, which we're

29:52

super excited about. Um, and if you go to are you readyforai.com,

29:57

this is a comprehensive course. I I would call this like a a m an AI mindset

30:03

training. Yep. Right. Tool that this isn't about learning any specific tool. You won't

30:09

learn that here. You won't learn how to use midjourney or this or that. You'll hear people talk about it, but what you'll learn is how to think about it,

30:16

how to apply critical thinking, all that sort of stuff. And so, go check that out. that's now available and super

30:22

excited about it. And you get a certificate at the end and it's not a gimme because Vicki made this

30:28

course. I mean it's you can't you can't not watch the things. You have to watch

30:33

them. Vicki is a she she's detail oriented person. The questions are hard.

30:40

It's true. It's very true. Um great. So, with that, uh, I am excited, uh, as I

30:48

know you are, um, to bring up our very special guest, Leon Furs. I think it's

30:53

furs. Do you know furs? Yeah, he said furs. One syllable like furs. Leon furs.

30:58

Not furs. Not fur. Not fury. No, there's no there's there's no drama here.

31:05

Uh, no, I'm super excited to bring him up. Um, he he uh he Why don't you tell the good people about Leon? Well, you

31:12

know, Leon is going to enlighten us quite a bit, Kyle, because he he his

31:18

work overlaps with some of the stuff that we do, but it also goes way farther

31:23

out, and he's nerdier about it than we are and more writerly about it than we

31:29

are, and we'll be able to break it down like an academic does. Um, but I I

31:34

encountered Leon in one of the many many conversations about AI and education and

31:42

pedagogy and assessment. And I was particularly lured in by he has created

31:49

I think with a collaborator but maybe himself this AI assessment

31:55

tool and the idea is that the way that we've been doing assessments forever is wrong, right? and that it's that now we

32:04

get to break the broken model completely. Wow. Um and then the other thing is that he

32:12

is very um involved with how creative uh

32:17

critical thinking and AI go hand in hand. So I'm really happy to have a

32:23

chance to introduce him and get to know him uh better. Hello Leon.

32:29

Hello. Thank you. Good morning or good evening. I'm sort of broadcasting live from the

32:35

farm in southwest Victoria, Australia. So, it's uh 8:30 a.m. over here. 8:30 a.m. What do you have on the farm?

32:43

Uh, lots and lots of sheep. Like thousands of sheep. Thanks. That's good. So, nothing to do

32:50

with AI. I was going to say, are they AI literate? Your sheep?

32:55

Uh, well, I don't know. Maybe. You never know. Would you maybe we could get create an AI that can understand sheep

33:02

language? We started to do that for cows and dolphins. That's awesome. Why don't you why don't

33:08

you introduce yourself? Tell us who you are. What What should we know about you? Um so I'm a I'm a farmer. No, I'm a I'm

33:15

a PhD candidate. Um I am a former educator. So I was working in secondary

33:22

ary education and uh November: 2022

33:29

the start I began my PhD in generative AI um broadly digital texts which is

33:36

kind of my um my area and then a couple of weeks later chat GPT came out and

33:42

sort of you know blew everything up. So for the past few years I've been uh consulting uh advisory positions with

33:49

schools with sector organizations over here some of the independent schools um

33:54

you know everyone really because the whole world seems to have been thrown into chaos by things like chat GPT. So

34:02

yeah, just a real sort of wild ride out of the classroom and then into this role of uh

34:08

supporting people to understand artificial intelligence. And um I mean

34:14

you mentioned uh before I was listening to the show from behind the scenes and we were talking about uh the way that

34:22

this seems to be really accelerating out of control now. And I think a lot of people, particularly in my world, a lot

34:27

of teachers have been feeling that pretty much since the release of Chat GPT, but it never seems to stop. And uh

34:36

you know, every every time I turn it around, there's some new feature or some new thing or or whatever. So

34:42

yeah, um my role is really to try to kind of float across the top of all of that and

34:48

and cherry pick out the really important things for educators and for students and just just help them to say, "Okay,

34:54

you know, you don't need to be across everything, but here are some things that are be really helpful if you do

34:59

understand and you know, get your head around these right now." Yeah. How do you how do you square the

35:05

circle where you've got particular I I find it

35:12

completely baffling in education quite honestly you've got you've got the the world's most profound knowledge tool has

35:18

been invented and handed to us and a lot of educational institutions are like never shall we play with the twix the

35:26

the evil tool. Um and and and yet you've got people like you that are thinking

35:31

about this so so thoughtfully. How you you know how do you explain you know

35:38

that gap and and where are you seeing where are you seeing people who were

35:43

resisting it before kind of accepting it or embracing it even ideally? Mhm. Mhm.

35:49

Um I think so I wrote a post on my blog um I don't know a couple of weeks ago

35:54

now about this idea of resistance and refusal within education. And I think in

36:00

certain corners of education and for the right reasons there are really good reasons to resist the technology. Uh but

36:08

the reason is not I don't know what it is. It's kind of terrifying. Um right it's moving really quickly. I'm going to

36:14

go over here instead. I think the reason that a good reason the reason for resistance in education for me is that

36:20

like so my my main method I'm an English and literature teacher mostly and then a

36:26

media drama teacher. So the media side of me super interested in, you know,

36:31

this like real kind of plastic quality of AI content and what it means now that

36:36

the internet is filling up with this stuff and what it means for content creators who can now, you know, create

36:42

videos or Google genie that you mentioned earlier, Kyle, around, you know, building 3D worlds, all of that

36:47

super interesting to me as a media teacher. As a literature teacher, I could probably teach 90% of my entire

36:53

course without a computer, let alone without AI. you know, I want students getting into books. I want them

36:58

discussing. I want them working together, being really critical. So, like I think it's highly contextual and

37:06

I think what we're seeing with some of the informed resistance from educators, we're seeing people say actually like

37:13

this, you know, I don't really need this really technodriven narrative in this

37:19

classroom because it's not super relevant to what I'm trying to teach over here. Mhm. And so what I would say is that so if I

37:26

was a literature teacher, if I was in the classroom right now, you know, 80 90% of my stuff I could probably do

37:32

offline and do a better job. But then when it comes to teaching, you know, adaptations, digital text, multimodal

37:39

texts, I'm probably going to have to work with artificial intelligence in that context and teach it really, really

37:45

well because artificial intelligence is going to be part of the future of literature

37:50

and literature studies. Yeah. whether whether we want it to or not. So, it's about it's about that balance

37:55

and it's about that context for me is is way more important than, you know, whether we we ban it or whether we, you

38:02

know, jump in with both feet. Leon, this is related, but I I cannot

38:09

resist. I need to hear you talk a little bit about your um

38:17

the question marks in your mind around the motivation of Open AI to roll out

38:22

study mode. And the reason why I ask is because my first take was, "Oh, nice

38:30

try. What a bunch of greenwashing. This is lame. This is like some

38:36

custom instructions. Uh why are they doing this? What are they? What is the play? So, you tell me what do you think

38:43

the play is? Yeah. I mean, I'm I'm really cynical. Okay. if I start there. Um, particularly

38:50

of of Open AI, which is interesting because I mean, you know, there's great reason to be cynical of Microsoft and

38:56

Google and every other tech company, but Open AI in particular, I think they're like they're like this, you know, this

39:01

yappy little dog that's kind of nipping at the heels of the big tech companies, and they're going for market capture,

39:07

you know, they're they're going for as many um dollars in the bank as possible,

39:13

you know, which is the business model. That's fine. That's them. Yeah. But in education, I think what they've done is very um very canny, very savvy.

39:23

They've they've gone at every level of education. Okay. So, they know that students are their biggest user base,

39:29

but students don't have any money. Okay. So, you know, students aren't going to bring in the profits. So, so actually

39:34

where in education holds holds the purse? Well, government. Okay. So, let's go to government as well. Okay. Who else

39:40

holds the purse? EdTE. So, let's go to Canvas. Let's go to Instructure. Let's get our technology embedded in the

39:46

learning management system that millions of students are using because then that'll probably, you know, add a

39:52

premium onto those instructional licenses. Universities will start paying that bill and then OpenAI's got a

39:57

revenue source there. Yeah. So, when OpenAI released study mode, I

40:03

saw it. A lot of people were saying, oh, you know, this is OpenAI responding to this problem that all of these students

40:09

are using it and um you know, they're offering a solution. It's okay. First of all, it's a problem that they created.

40:16

So creating a problem and then selling the solution is is pretty classic tech industry territory. That's nothing new.

40:23

But then then also I mean it doesn't work very well. Okay. So it doesn't actually solve the problem. Any student

40:28

who wants to actually get an answer will just turn it off. Never bo never bother

40:34

turning it on in the first place. So I started to think so what's the bigger play here? And I think that it's to to

40:42

kind of persuade students, teachers, universities, you know, oh, we're we're doing the right thing. You can invest in

40:48

our product. We've got your best interests in mind. We empathize. It's it's classic kind of marketing

40:54

territory. Yeah. Yeah. Yeah. Now, over here in Australia, we just had um Open AI in in CRA with our

41:00

politicians behind closed doors, and they've written a blueprint for AI in

41:06

Australia where they've made a, you know, a two-page spread recommendation of policy changes for education. And

41:11

that's where the real game is, you know, it's it's not this sort of halfass study mode

41:17

crap. It's it's what's happening behind closed doors with politicians in countries like Australia and the UK and

41:23

the US. That's where the interesting stuff's happening. Can I can I ask you Leon for for me the

41:31

thing I find most fascinating if if the tactical skills of the tasks of work are

41:38

increasingly automated with this? What do you think is

41:44

what do you think is the most important thing to be teaching right now? like like what should we be preparing our

41:51

students to leave the workforce? I like I imagine if you if you look at what's

41:57

happened in the past two and a half years, I can't imagine being a freshman right now, like going into my freshman

42:03

year and imagining what's the world going to look like four years from now. Are you kidding me? So, so, so, well, I

42:10

know that, but but but given that that this thing's accelerating, you know, from your point of view, Leon, what do

42:16

you think are the most important skills to be teaching right now? I mean, I'm I'm I'm pretty

42:22

old-fashioned, I guess, um, at this point, which is interesting in my in my position as someone who's also like right on the front end of artificial

42:29

intelligence. But I think that there's a lot of a lot of narrative around like, oh, you know, AI is going to replace X,

42:35

Y, and Zed skill and, you know, nobody needs to do this anymore because AI can do it. I think there's a lot to be said

42:41

for um valuing some of the the really important skills that we that we already

42:46

have and that we already teach. I'm a big fan of teaching AI through existing disciplines. For example, like you know,

42:52

you can in mathematics teach about algorithms, in English and media studies, teach about critical literacy.

42:58

So you know reinforcing some of those skills which may seem to have been

43:03

offloaded into AI already like yeah okay chat GPT even you know can handle really

43:08

advanced mathematics by now um certainly up to a senior secondary level that doesn't mean we shouldn't teach the

43:15

foundations so that when students leave when they start to use these technologies or as they're using them

43:21

throughout schooling they're using them better and I I I want us to really

43:26

resist falling into this trap of saying because a AI can do a certain skill, that skill is not important anymore.

43:33

Mhm. Because those skills are still super important. But I think it really benefits the technology companies for us

43:39

to think that those skills aren't important anymore and they can just do it for us. Okay. But that's I mean that's breeding

43:44

a whole generation of idiots, isn't it? It's it's not creating a society where people can actually do things. It's

43:51

creating a society where people consume things that someone else does for them. So yeah, like you know that's that's a

43:58

pretty probably conservative answer and maybe a bit of an oldfashioned answer. That's my

44:06

answer is my answer is definitely not everybody needs to learn AI literacy though because I mean you look at the

44:12

people who are profiting the most from this tech. Sam Alman, Mark Zuckerberg, Elon Musk, all all of the other white

44:20

bloss, you know, none of them had AI literacy

44:25

courses when they were at school, right? And they they seem to be doing okay, so I'm told. Like, you know, Elon Musk

44:31

wasn't sitting sitting in front of a computer in grade five learning AI literacy, but he seems

44:38

to be doing all right from the AI industry. So if we pitch forward a few years, what are the current generation

44:44

of students, what do they need to be secure in a world that we we can't predict? I I don't think it's, you know,

44:51

how to prompt chat GPT or any of those kinds of skills. Yeah. Yeah, that makes sense.

44:56

Um, I am going to ask you my first question, but first I have to just say for the for the nice people that I think

45:04

what you just demonstrated is exactly what Kyle and I were talking about, which is the skill of critical thinking.

45:11

So, you know, when study mode hits the scene, like it doesn't have to mean

45:16

everybody go start using study mode. It could mean why is this thing here? Is it

45:22

actually good? Do they think we are idiots? and then and then make

45:30

Well, yes, they do. Yes, they do. If they give us enough pellets, we will

45:36

stay on Chad GDT for all the hours of the day and then they can start feeding us ads right into right into our veins.

45:44

Um, okay, Leon, what was the tipping point when you knew you had to go all in

45:49

on AI and what happened after that? Yeah, I mean, so I've been playing

45:55

from, you know,: 2021

46:01

an eye on this stuff. And this this was back when I was an English teacher. Um, and I was seeing students start to use

46:08

these, you remember like paragraph spinners and write sonic and Jasper,

46:13

these third party applications, these sort of rapper apps like back back in the day when uh GPT2

46:18

and 3 were open source and so people were building this tech on top of that and I was keeping an eye on that stuff

46:24

or my PhD in sort of May June: 2022

46:30

a lot of reading around the ethical concerns and around, you know, the the kind of where this stuff was heading.

46:36

So, I was maybe a little bit less surprised than most when chat GPT came out in November.

46:41

But at the moment that chat GPT came out, I thought, "Right, this is it. We're done." Like my my horizon for my

46:47

PhD was in three to five years time, everybody's going to be using this. And I called my supervisor, I think, and

46:53

and said, you know, sort of 24 hours after Chat GPT was released, I said, "It's not three years." anymore. It's

46:59

It's three months. Yeah. And then we saw over that three months that 100 million users passion for about

47:06

I think we're up to 700 million active users now. Yep. It's crazy, isn't it? Yeah. 100

47:12

million in six weeks. And the worldwide web, it took about six years to get to 100 million users.

47:18

Yeah. So look, I mean ear early early in that process is when I clocked it and I

47:24

think that that positioned me quite well to to help schools in those first few months.

47:29

Yeah, for sure. Have there been what have what have been your biggest um we we call them Kevin

47:36

Mallister moments which probably won't be a reference for you. There's a movie called Home Alone where the kid goes

47:43

um so like what have what have been those moments for you where you've used the tool and you've gone oh wow this is

47:50

this this changes things. Do you have any do any of those stick in your mind of moments of just super clarity that

47:56

ething shifted for you? March: 2023

48:02

uh saw the release of GPT4. Um so the the upgrade from 3.5 to 4.

48:08

Yeah. And um also the release of Claude, right? And and for me that was two things that

48:14

was oh hey look it's not just open AI and there are other companies out there

48:20

capable of releasing really powerful models which made me think everyone's going to

48:25

do it. Okay, this is this is going to be every single tech company that has access to data, hardware, and money is

48:31

going to produce one of these things. And the step up from GPT 3.5 to GPT4 I

48:37

felt was was a step change. You know, it was huge. The multimodality, the image recognition

48:43

with GPT4V, um the the addition of the internet connection that was like full. Everything since

48:50

then's been kind of lukewarm, little incremental. I know that like GPT5's touted to come

48:55

out soon, maybe in the next 48 hours or whatever. But, you know, I'm kind of, you know, as as we know,

49:02

I'm cynical about Open AI, but I am hopeful that it's a similar kind of, you

49:07

know, 3.5 to four kind of step change rather than a four to 4.5 or four mini

49:13

high pro megatastic mega. Exactly.

49:19

If they can I' to be honest, if they can just get their naming conventions right, that might be significant enough.

49:25

Yeah. Frankly, what they need what they need to do, I mean, the genius of chat GPT was they buried all the interface of

49:32

like temperature and choosing models and all that sort of stuff and then they basically just let it turn back into

49:37

that in the you know, in the public interface. So hopefully I mean this is because this is a technology designed by software nodes.

49:44

Exactly. Software devs for other software devs. It's not technology designed for mass consumption and so

49:50

they don't it's like they don't know what it is. You know that nobody really knows what this product is for.

49:56

Well, they inadvertently became a consumer software company overnight. So,

50:01

I give them a little bit of grace, but at this point it's been two and a half years. So, it's like hire some marketing people, hire a couple liberal arts

50:08

majors, you know. So, so anyway, um so so let me let me jump in and ask your

50:13

second question. um based on your expertise and especially given you you've got such a clear focus, it's

50:19

great. Um what are trends? What's a trend or if there are more than one, but

50:24

what's a trend that you're paying attention to right now and why? I um I have been writing a post a sort

50:35

of a series of posts which I call like the near future of AI and I do it about every six months or so because that's I

50:42

sort of refuse to commit to anything further than six months and um it's it's even it was a chapter

50:49

at the back of the the book that I wrote um on artificial intelligence which I'm just going to have to do this cuz it's

50:55

right here next to me. Nice product placement. Look at that.

51:00

Recordings and courses and things earlier. So that's why that's sitting there. But um the

51:07

chapter in that book started to talk about increasingly multimodal artificial intelligence

51:13

and that sort of multimodal aspect is part of my my PhD as well. So that not

51:18

just you know text to image but image to text with image recognition um text to

51:25

speech, speech to text, speech to speech now with open eyes speech models

51:30

and I'm and I'm seeing you know an increasing trend towards multimodality from all of these organizations

51:36

particularly Google um and open AI. So, we're going to see, I think, a confluence, and I think this

51:42

might happen with GPT5, a confluence of the, you know, the video, audio generation, speech-to-pech model,

51:49

everything all in one place. Yeah. And, you know, from my perspective as an educator, that's content on demand

51:56

territory. That's teacher standing at their desk and says to their computer, "Oh, hey, I just need a quick video on,

52:03

you know, cellular structure for year 8 science." And ChatGpt responds, "Yes,

52:08

sure. I'll uh I'll generate that for you right now. And then 3 minutes later, you've got a you know a little short

52:14

fully um narrated video. I think that's where this stuff is ultimately heading.

52:20

Um and and you know there's there's a few reasons for that. One is that that's probably where the market wants it to go

52:27

and the other is that that kind of content is really engaging and persuasive as well. Um but that would

52:33

probably be the most useful version of the technology I think. Yeah. Um do do you think we go to a world of

52:42

hyperpersonalized education because of that? Like what what do you what are the implications of the multimodality for

52:48

you? Um that I mean the implication is sort of twofold. I wrote this post a few

52:55

years ago in a in a few sense called um digital plastic where I kind of compared this technology

53:00

to real world plastic. I said like like real world plastics like it's you know it's incredibly ubiquitous and growing

53:07

in volume. Um and some of it will be good like you know plastics are great. Plastics are sterile. They're great for

53:13

medical purposes. You know you can make prosthetic limbs out of plastics. Uh they're also horrible for the

53:19

environment and look at the Pacific Ocean. Yeah. Yeah. And and so like the online world's going to be the same. You know

53:24

it's going to fill up with all of this garbage. And we're already seeing it on YouTube on Tik Tok and all of these

53:29

social media channels. AI generated slop filling up everything. So that will be

53:35

one outcome. Okay. And then the tech companies they love to talk about personalized learning

53:40

but I think what a tech CEO thinks personalized learning is um is a bit

53:45

different to what a teacher thinks personalized learning is. you know, they one child sitting in front of one

53:51

computer interacting with a chatbot whilst a, you know, a teacher/babsitter kind of negotiates the room like this,

53:59

you know, alpha school thing that you've got in the US. Like that's not personalized learning. That's that's isolated learning. That's

54:05

lonely learning, you know. Um, so so I don't want to see it go down that way.

54:10

But but from the media teacher part of me says, "Oh, this is great." like content content on demand, the ability

54:17

to, you know, build a v a video game in virtual reality whilst you're, you know,

54:22

whilst you're in there kind of negotiating this environment. All of this stuff's really exciting to me. So,

54:27

you know, I'm I'm I'm torn, but it's it's this is kind of the job that I signed up for. I I love that you said that you're torn.

54:34

I I I often find myself like this is the greatest thing ever. Oh my god, this is terrifying. Like, and you kind of have

54:39

to hold both of those things as reality, right? Yeah. Yeah.

54:45

Um, all right. So, you know, this is the AI readiness project podcast. We have

54:51

the AI readiness program and Kyle and I chose the word readiness really um

54:57

really intentionally because we felt like it's a lot more than AI literacy.

55:05

That implies there's some kind of an end point, a check mark, a point you've crossed over where now we get to say

55:10

you're AI literate or who or who gets to say that? Like

55:15

who's in charge of that? Whereas readiness feels like an existential kind

55:20

of a a way to be. Um what does AI readiness mean to you? And what would

55:28

you advise somebody who's just now getting off the sidelines?

55:34

It's it's about being able to engage without immediately feeling overwhelmed and running away.

55:41

Okay. So, it's this ability to kind of, you know, meet the technology where it is right

55:47

now, which is which is overwhelming and then to just kind of sit with that discomfort for a little while and let it

55:53

be overwhelming and then say, "Okay, I'm overwhelmed, but I'm I'm going to explore anyway. I'm

56:00

going to kind of I'm going to step into the the more of this beast and and I'm

56:05

going to see what happens next. Just kind of one step at a time. So that that for me would be, you know, I'm ready I'm

56:13

ready to start engaging with this with this technology. And you know, nobody needs to be across every single thing.

56:20

Um you know, Kyle, when you mentioned Google's Genie 3, I I knew about Genie 2, but I didn't know that they'd

56:26

released an update until I saw you talk about it backstage. Yeah. Um, and I can see that was the 5th of August, so you know, 48 hours ago for

56:33

me here in Australia. Yeah. Um, so that's understandable. It's fine. And if somebody like me who, you know,

56:38

it's my full-time job to work with these technologies. Well, no, no. This I I I talk about this

56:44

five nights a week and it's I like reg last night on my live someone goes, "Did you hear about this thing?" I'm like, "No, I just told you six things.

56:51

Apparently that wasn't enough. There was a seventh one I missed, you know." Yeah. Let me ask you something. This this was

56:57

something I struggled with last week in my back to basics course is do you have advice on

57:03

is there a specific tool you think someone should start with if started h

57:10

having just slammed open AI for the entire session um just chat GBT

57:16

not because it's particularly you know an ethical technology or it's whatever but it's probably best in show

57:23

you know it's um until further not the best on the market for what it does. There may be certain other models which

57:29

do X Y and Zed better. You know, Claude might be slightly better at writing JavaScript or like but who cares like

57:35

right use the most common one. Yeah, it's it's common. It's it's got

57:41

high utility. It just generally sort of works nine times out of 10 now.

57:46

Yeah. You know, so unlike like you know Microsoft Copilot which

57:51

um Yeah. So, so go go go for that modern main platform. Have a bit of a feel

57:56

around start getting like comfortable with what this technology looks like. Beautiful. Love it.

58:03

Love it. Love it. Love it. Um, well, this has been this has been awesome. So, this is this is where the good people

58:09

can go to find find you and what you're up to. Yeah. Yeah. Just leon furs.com. Pretty

58:14

straightforward. I do have uh practical strategies.com which is um the online courses for

58:20

educators primarily and some for school parents. Um and are are there links to that at at

58:26

leonfirst.com? Uh uh yeah, look I mean this is my internet life. There's links for me for

58:33

everything for everywhere. I lose I lose track of them. You'll find it somehow. You'll find it somewhere. Just find find

58:40

clicking on links, right? Just stumble across the internet. Chat GPT actually recommends my website quite a lot now. I get a huge volume of

58:47

traffic via um via chatbt.com apparently. Did you did you optimize it for AEO or

58:54

GEO or really? No, I I don't have any time. I never even bothered with SEO. So yeah,

59:01

the game SEO AI AIO that's like a Exactly.

59:06

Just vowels. Just lists of vowels. Yeah. Yeah, I mean if someone is lucky enough to

59:12

find me on the internet, good for them. I've got a really weird name, so that's

59:17

it really. I just I leverage the fact that I've got an unusual name. No one that helps. Yeah.

59:23

Leon, it was so nice to meet you and to learn more about you and your your

59:28

perspective. I find it utterly refreshing that you can hold on to both,

59:35

you know, both ends of the spectrum so so capably and also your writing is so

59:41

fun to read and these are heavy topics and it's really nice when it's just a delightful um that you know sit it's

59:49

like your blogs are like sit down, get a cup of coffee, enjoy Leon's blogs for a little while. So nice job.

59:57

Thank you. Thank you very much. No, it was a lot of fun. Thanks for the Yeah, beautiful. And thank Yeah, thanks for getting up early. I know it's bright

:: 00:03

and bright and early where you are and I I don't My kids will wake me up at 5:00 a.m.

:: 00:10

This is awesome. Thank you so much. And thanks, Leon. Thanks. Thanks, guys. Thanks.

:: 00:15

Bye. Bye.