0:04 A video clip of a dragon burning a village. If I don't return in a few hours, you must take her and flee this place. For it would mean that I have failed and they would be coming for you. Female elven warrior jumps from roof shooting a bow.

0:19 These visions in your head, songs or story, you can create them yourself. So, this is still not cheap to do. Stay tuned as we talk more about some of the advancements in AI when it comes to creating video. What does it do well?

0:36 What does it not do so well? What does it look like when it gets two battling dragons coming together? How do we create consistent characters? When is it better to be human?

0:47 When is it better to actually take the inspiration that we get from AI and use that to create something for ourselves? Something human made. What is the impact on other creatives who are making video or doing these things manually? What is it going to mean for them?

1:04 What is it going to mean for them with AI? How can they embrace that? How can they continue to have a job that they love doing? They can still work with what they love to do to be creative.

1:15 We can look at this and many other things in this episode as we continue to uncover what is better with humans. In this episode, I talk with my friend Glen Boswell. Glen and I met early in the year in a course related to film and TV. Glen is a digital creative, has worked on numerous computer games and TV shows,

1:33 and now runs the Astral Forge, a content production company that uses the latest AI advances to help businesses do more with what they have for less. He's also the author of Flight of Dragons, Wrath of the King. Glenn, welcome.

1:48 Welcome. So, Glenn, I think I'm excited about this conversation. I feel like there's a lot of great things that we can be talking about. And firstly, give us a bit of a background on your time working.

2:03 In TV and computer games. Yeah. So when I was at school, all I ever wanted to do really was to make computer games. And I'm talking back when we're looking at sort of pixel lemmings, you know, that that that kind of era in the original Nintendo,

2:22 Super Nintendo and Mega Drive. And I came out of school and went to college. But back then, there wasn't really any college courses that dealt with the creation of computer games. So a lot of it was kind of teaching myself from home and trying to get hold of the packages that I needed.

2:40 So deep paint and then moving on to 3D studio after that. And I came out of college and created my own sort of 30 minute cartoon and used that to go to interviews. And my first job was working on a Thomas the Tank Engine title. And alongside that was Noddy.

3:02 And. Starsky and Hutch, which was a fantastic time to be working on those games. Wow. Oh, it's when we when we had this discussion earlier about what we're talking about, the TV games, you'd like nothing amazing.

3:17 But actually, there's some special, special shows there. I mean, you know, we've all said we've all maybe not so much now, but a lot of people are growing up, growing up with Noddy. And my mum had this habit of what's a habit that she would buy, like to buy me. That can bend. It's compendium of something.

3:35 So to read to my to my two daughters when they were young. So I had the compendium of Thomas Tank Engine, compendium of Winnie the Pooh. So I read all of the all of those stories. Yeah.

3:45 I've played the Starsky and Hutch game as well. So that's. Did you read? OK, that's good. That's good. Yeah.

3:50 We nearly made a follow up, but the follow up never came to fruition. But yeah, I went on to there and multiple sort of games companies and then into the TV TV industry where I worked on. Act in Scarlet at Pinewood Studios. We did two series of that.

4:09 And the industry has changed so much because when I first started, there wasn't any kind of 3D 3D art as such until really sort of Toy Story came out and then it changed everything. And the way that I sort of look at AI now is just it's just an evolution. It's an evolution of technology that's always been there. An example of that.

4:35 And they were talking about AI taking over everything. And I said, well, when you think back to when you're lighting a scene in 3D studio, you put a sphere into the screen, into the scene, and then you put your light source in there. When you move that light around that sphere, you see the light moving around the sphere in real time. And the reason why it's doing that is because it's basic AI.

4:59 It's always been AI. And now we're just looking at an evolution. evolution of that technology so yeah exciting times yeah it's interesting we can we can talk about this a bit later on and it's something that's particularly close to my heart as well as

5:14 helping to change the narrative a little bit in terms of how we look at ai and how we approach it look at it as something that we have that can actually enable us to do more rather than actually do less and it's about how we use it right yeah if we use it in a way to replace people it will replace people we don't have to use it in that way a better way to do it is to use it to actually

5:35 grow and learn more have ai do the tasks that really are not really that valuable to us like you just described you know if we go back to a time when you had software that you know to get the light you know you'd have to maybe move the light then you'd have to determine how intense you have to do all these these all these things manually and i think and you can argue well okay

5:59 there is there's value in that when you couldn't do it any other way but if the value is in being able to create this creative piece then you won't have all the tools don't you you know if you're an artist and you've only got five paints you paint with the five paints but if you've got a palette of 10 you're going to want to use the palette of 10 because you've got the ultimate thing is the

6:18 creative output that's correct yeah yeah and i mean a perfect example is um i think a lot of people look at a game environment at the moment where you've got trees across the the environment you've got a river flying through the middle you've got the correct pebbles underneath the water and you've got little soft areas where maybe an animal's crossed

6:38 it is it looks fantastic but i think a lot of people think that an artist has actually gone in and manually placed all those trees and placed all of those rocks and and the various areas but they haven't it's all done by sliders you know how many trees do you want where you just do up and down the slider you can specify where you don't want trees you can specify where you don't

6:59 want trees and it will vary them as well so it'll vary the branches and the leaf density and the height so that they look different to each other and ai is basically an extension of what's been going on for years and years and i think it it's scary because a lot of people in the industry are fearful of it and it's affected me it's affected other my colleagues but you need to embrace it

7:26 and use it as a tool and the way that i do it is hopefully the right way if there is a right way yeah it reminds me of um as you're saying it's been around for a long time and we think about computer games i mean you mentioned about right yeah the first bit of coding i ever did was on a zx spectrum 48k yeah and it was copying

7:49 code out of a out of a magazine and even then i'd play a game uh there's a game i think it was called match day a football game and that was using some you know it was fairly primitive but he was using some form of artificial intelligence and we found a lot with the games back then because what we would find was you'd

8:10 find the patterns right a lot of the old the old sort of arcade atari games you there was patterns because it was limited to in terms of in terms of what it could do but it was still using a form of artificial intelligence and the people that we mentioned earlier on grand theft auto is actually grand theft auto all these games that we've been using for a long time and we've been using them

8:29 for a long time and we've been using them for a long time and we've been using them for a long time that they were using that there is some form of the game interacting with us in some way in and not necessarily a predictive pattern is is is using some form of ai or like you say it's also have been used ai to to generate it and it's i think that the biggest challenge and i've

8:50 been talking about this a lot recently is this as you've mentioned so it's the speed of change when we've seen things change before we've had a chance to go okay come out that's going to change things for me well you know it's it's a way off you know i've got i've got time to to to train i've got time to to go and do more things but actually what we're finding now

9:13 is there isn't necessarily the top well there is time but we're having to we're having to sort of chase up we're having to actually be much more quicker in terms of in terms of doing that yeah certainly i mean ai speeds up the speeds up the output vastly and it's one of the things that that we can we can use these new tools or we

9:37 can be fearful of them and put our head in the sand and and not use them but they're here and they're getting stronger and stronger over a very short period of time if we're looking at video as an example i would say maybe 12 months ago you had struggled to get somebody to walk in a straight line and now you can essentially you can put a storyboard on your computer and you can put a

9:59 storyboard in put your characters in and you can get that to follow that for 15 seconds of footage with camera cuts um effects sound and and it's just a it's a vast improvement yeah yeah and i think one of the things you mentioned there is the getting a character walking a straight line and and again something else we talk about the advancements that we've also come to the point

10:26 where we can get some character consistency i know i'm played around with this probably over 12 months ago now just in using something like chat gpt and trying to get it to reproduce a character a picture of a cartoon character in different poses but actually in a consistent way and what i found was i i could there were similarities

10:49 but every time there was something that wasn't quite right so so what is what does that process look like how do you how do you make sure that you get this consistency in your characters so i've got a little example here that i could actually show you so from an area i've been working on if i just go to share this a character sheet will essentially have all of your character

11:12 details on it so this is um for example this is a variety to warlock he's a bad guy you'll usually have different angles of the face and standard pose so that helps the ai to work out if you've got for example shots from behind the character any specific items that they would usually be carrying around you would include these into the character sheet along with your palette palette of colours

11:34 um any other little accessories that they might have on their belt um their height their kind of attitude you know the way that they might walk etc etc and then you would then be able to use that to create a scene so this is just a very very short shot it's the same character ride the dragon master it so there we've got wow full emotion

11:59 we've got the minas scene that's got the rain throwing it down onto the character you know you can get a really a really nice sort of sense of what what's going on there there's there's different ways of doing this so what you can do for example with an environment if we look at this one here this is a forest environment an elven kind of forest with you've

12:22 got the magical creatures that you might see in that forest you've got sort of the styles of the trees um you've got a description about the trees they're very specific to the stories they're the nedi eye trees and then you've obviously got some ancient ruins in various areas throughout that that woodland if you run that into the system along with your characters and what you what you want

12:43 to see Glen is showing a generated video clip with an elven woman warrior walking in the forest alongside a young girl the large dragon walking behind so there you've got a really nice natural walk through your environment um you've got i'll just sort of scroll back a little bit there you go you got got your dragon walking with your

13:06 two characters they're adhering to the surface on the ground the lighting striking them in the right area the dragon has got real weight to his movements so if you look as his claw comes down here he's got spreads his claw out puts the weight down you know this kind of thing is even here as he comes across here with his other claw he pulls the claw up to avoid the root as you can see that yeah wow

13:33 you know it's just it's just mental yeah and it's this kind of thing that would have taken animators a long long time to get through now there are if i show you this one so this is a burning a burning village uh the village has been attacked by dragons and our main character who's a friendly dragon that goes into the village to try to rescue people so i did a small storyboard made

14:02 with chat gpt and then just made a little bit of a storyboard and then just made a little bit of a tweak afterwards and this is a test shot from that particular scene the clip is of two dragons squaring up the traitor he's here the one that spoke lunges at a smaller dragon dodges and throws the large dragon back he's far away yeah so that's incredible isn't it i think and and i i think it's

14:31 such a brilliant um such a brilliant representation of the in terms of getting the most out of ai is preparation yeah we see it so often you know you put a one line prompting or maybe even a sentence you might get something and you probably get something relatively good but is it really what you wanted and often we can see people get

14:54 the frustration with how to do that and because you've really shown a great example there of there's multiple layers in that there is there's actually you generating a storyboard of the scenes describing what you what is actually going on there and you're doing the same thing with the character sheets so there is a reference

15:11 for the ai to be able to continuously pull that from and that's yeah it's an amazing scene so what what is the inspiration behind this what is the you're creating these videos with these dragons what is there yeah well this this particular video is me actually trying to to create a series or a film from my book trying to sort of visualise it and it and when i first saw the dragon and

15:38 drakarim and uh neolandra who's the the elf that that saves him earlier on in the story on the screen and actually reacting with each other it's quite an emotional moment because it was there was a lot of work that went into getting neolandra to look correct and drakarim to look right and and then putting them in the scene and actually seeing them animate it's a feeling that

16:00 it's it's strange and then this this is a whole new kind of medium for people to present their and there's a lot of creative out there that for years wouldn't have had that option yeah um and now now they have if you've got these visions in your head songs or stories or you know just short stories scenes you can actually go now and you can you can create them yourself

16:29 if you just put they learn learn the methods and and put the time in you can you can get them to look fantastic um there's there's various options without there so you've got cling which is great at the moment that's that's a fantastic animation package you've got sea dance 2.0 which is quite controversial but it's um and it's expensive to use but it's really really good for particularly combat scenes

17:00 so do you want to see a little combat scene yeah just just before we jump into that just what you mentioned about sea dance by actually sea dance being controversial what is what is controversial so when they first showed the examples of sea dance um they had literally tom cruise fighting against i can't remember who it was now

17:20 i can't remember who it was but tom tom cruise was basically there having a having a fight and it looked superb and it was fantastic and there were more and more of these videos coming out and then really they had to reel it in because it was just it was too powerful so now when you when you use sea dance particularly in the uk if you try to upload a face it could it

17:45 could be your face on my face any face it won't let you do the footage so there's various ways that people are working around that at the moment but i think it's kind of like trying to find the balance where you're not encroaching on the full likeness of known actors and people that do it for a living you know you or encroaching on copyrights you have to get the right copyrights

18:08 right level and they're trying to find that right now but your thing is you want i mean the the aim is isn't it that you want creatives to be able to be creative you want those that had that their living is is around being creative to continue to to make a living from that continue to you know what's the ultimate aim for a lot of people is to do a job you love and if you're in that space you

18:30 want to continue to do that and yet there's also the element of you spoken about as well how without to get the funding i remember many years ago i don't think the film ever happened but i remember being approached about it was it was an early crowdfunding days and about his film um i think it was it might have been jason donovan was going to star in it or something but it was it was

18:53 never happened you know because you know as much as i didn't part with any money but they were wanting to raise these funds and it never happened and we look at what you've been able to do already then it's like wow that's incredible to bring your vision to life and you shared that the emotional emotion of seeing that of seeing something that you you have envisaged

19:15 that i'm sure you have dreamt i'm sure you have visualized these characters coming together and being able to see that it's something like you say you may never ever ever kind of uh see so so i think that's an incredible an incredible blessing as well and it's but like you see you mentioned that the balance is the balance it's kind of going we shouldn't use ai

19:38 to not pay people for people to do stuff you know if if we have the opportunity to be paying people to do stuff let's pay them for what we need them to do but let's not need but we don't need to pay them for the things that really they don't need to be doing you know it's it's the it's the creative piece right you for example you've come up with the story you've come up with these

19:57 ideas and you're basically outputting that creativity and using ai to build you to visualize that for you but you still had to come with a story and you still you still you still had to figure out how it was all going to pan out how were you going to and you still are that this is the thing i've i um kind of turned recently and i call it i call it ace and it is this this approach

20:20 to ai where we say okay we need the human to be asking the human prompts the human makes the request and then the um the c is the ai creates it the eye creates it creates something and then then we have the evaluation so as a as the end user will evaluate it and we'll say okay we're going to evaluate and now we're going to feed back and we're going to continue that loop

20:43 until we end up with what we want and then i had one other thing onto that which is d you know that we aced it to say well actually the d is to develop because the d is where we take all these learnings so the next time we go and do the ask again or a different ask then we're more prepared with that so we can actually have this growth but we're growing with ai as opposed to either

21:06 ignoring it or actually using it in a different way so i think that's a really good point and i think that's a really good point and i think that's a very inefficient way which is not great for sustainability but also it's it's not great for it for anyone if we're using it for things that we could and are better off using a human

21:19 to do yeah certainly i mean well the the examples are getting stronger and stronger of of the change that we're looking at here it is hard because there is a lot of kickback from various areas against ai and rightly so in a lot of ways but also it's so you know we it is i'll show you i'll show you something else very quickly so this is different

21:45 from animation this is uh i started out working with ai through mid journey and what i did there was um portraits so i was i wanted to to do portraits for my characters um but also for customers particularly people that like dnds if you like dungeons and dragons then um yeah this is this is for you so basically i wanted to be able to visualize their character

22:14 all the little details that they have and and there's loads of little bits and bobs in details and in the dungeons and dragons character a scar in the right place um a little item in the right pocket all sorts and i wanted to be able to do that at a decent price where you know it wouldn't take me months and months to do it so let me just show you this quickly yeah

22:38 i'm looking forward to this actually um you now got me thinking maybe i should i should do your character sheet for my old dnd character maliki who was a half elf sorcerer a perfect example is this character right here so hopefully you can see her so the the customer wanted to have she described the character to me you know sort of a sassy type bard rogue sort of character

23:04 and they wanted to have them gambling in a casino so i put them in a casino and i put them in a casino and i put them in a casino and i put them in a casino put the the kind of the description that i had everything else into a mid journey and each one that you get out it's okay you know you could essentially you could you could use that if you

23:20 wanted to as a as a but these items here they make no sense whatsoever you know her hands here make no sense the cards if you look at them are complete gibberish um and run through to the next one same thing you know that it's not right and if you zoom in you can tell it's ai you can just see that from the the way that the

23:43 the image is structured so what i would then do is say well i like how she's posed i like the the character there i kind of like the environment so i then take that and this is my final painting and this is all done by hand so if you zoom in you can see these are hand strokes this is my this is me painting

24:08 oh wow yeah so that is that is the final result that you get out of it and you you can use that in in multiple ways another one would would be this character here so we so this is what they provided me with was it like a little sketch of how they wanted the character to look um they they sent me over an

24:29 image of the weapon and that's her weapon that she uses putting that through into you know into um i think it's a little bit of a picture of her this was actually gemini and this is what gemini came out with which is okay once again it's a it's a decent image but it's not personal it's not got yeah you know you're missing out on a lot there so

24:52 i just restructured it and painted it and as you can see that this is all hand painted but it's using the ai to kind of okay you know there's a pose there i've got an idea of how i want to do this yeah that's that's that's a great example you know actually a great example of again that thing of inspiration in episode five

25:16 the uh in the best with humans podcast the conversation with a screenwriter and and and he's using he's creating it and using ai to create a tool to help inspire screenwriters to help them improve their screenwriting not to do the screenwriting format it's a great example of where we can we can go well maybe we can't we can do something get that extra inspiration

25:40 that allows us to do that to to grow and this is the thing isn't it what we want to grow we don't nobody wants to stagnate nobody wants to have ai do everything for them and just you know put their feet up and you know uh just uh drink pina coladas all day right you know at the beach right people might say that but they're going to get bored after a little while it's nice to go for a week

26:03 or two on holiday isn't it but you don't want to go for a you know a permanent holiday and people and we want to grow and be creative we're we're born to create we were created to create and so that's a great example and i don't know i do want to um one of those listening are probably we're thinking about oh yeah you just described um co-dance and we didn't actually look at that

26:26 but if you could just show us that show us what you did with that briefly as well would be uh i think would be cool just so there's not some people that are going oh you never showed us that thing what was that sorry with was it co-dance was that what you see dance see dance sorry yeah what c-dance is really good at is

26:40 it's good at combat so the the dragon scene for example that was through c-dance 2.0 so because it doesn't seem to be afraid of actually throwing the uh the characters against each other and having conflict between them whereas other ais tend to to struggle with that so if i sort of share this one here this is a an orc fight test that i ran through and i'll take you through it

27:05 a female elven warrior steps off from the start looks back at the second orc spins the spell in the chest of first orc and brings up her sword to block the second orc's attack a lot of that is fantastic right so if we go through slowly we have the block here where she actually twists her arm back to to hold off the attack then we actually get a wound on the

27:32 orc which is you know you get a bit of blood spray you get the reaction she looks satisfied with the with the uh attack yeah and then she brings it out we get a bit of spray as she pulls her sword back out again now she turns around she's aware of this person behind her and it looks like this is going to work out well to start with but this is where c-dance went

27:54 a bit wonky so the orc could have swung around with its sword straight into the character but instead he switches it and it goes from the other side and she takes her eye off of him and goes back to the other the other orc which is a strange thing to do right yeah but then she follows it up with what is actually quite a good move which is a

28:19 she docks spins backwards to give herself space and then eventually releases a spell which the spell looks great but i would like to see more impact in the in the spell there yeah and then and then here finally she brings her sword up and she blocks this attack here which looks great to start with but if you look into two frames time her sword's now in the middle of the attack and

28:43 the opposite side oh yeah so what what you'd probably need to do here is is um you can either individually paint the frames or do a do a cut to a different um shot there's all sorts of tricks you can get away with to to adjust that and this is another c-dance male warrior dunks an attack from north and slices orc's stomach with sword from this point here i think this is fantastic

29:14 because you've got this real kind of look in his face that he wants to to deliver the blow and then obviously you've got the blow impacting on the orc and again you've got all the the results of that but this is what this is what interests me about this particular shot if you look at the belt around the waist of the orc as he goes forward from the impact

29:37 the belt loosens so the weight of the belt is coming forward which you wouldn't normally see if somebody was was animating that you'd probably just have the belt near enough attached to the orc there yeah and that there's real kind of weight and impact that you can get from an ai sequence that you might not even get particularly out of some of the better animators uh in the

30:02 industry which is crazy yeah thank you for sharing that i mean i think this is just such a rich episode because people are really able to see some of this what is achievable but also some of the challenges right you're not saying this is simple and yeah also one of the things you briefly touched upon was was uh was cost yeah so this is still not cheap to do so

30:30 we're not talking we're not talking sort of 20 pound a month you know subscription wise are we on this this is a little bit more cost to these yeah especially for something like i said earlier sea dance it's expensive i don't know if if it's going to be a big deal i don't know if it's going to i can't work out myself i don't know if you would have a view on this but whether that's

30:50 going to get cheaper over time or more expensive as it becomes more advanced i'd like to think cheaper well i think yeah i think there's a number of factors that are going to come into play one is so many of these ai organisations are relying on funding and at some point stakeholders are going to want money back so that there has to they're going to have to have some plan of well how do we actually

31:14 make money from these things now i think maybe one thing will depend upon how much uptake is there in terms of in the film and tv industry and for some of these these types of applications because you say if they can make their money through the higher end space then maybe it does become cheaper i think there's also going to be potential legislation around sustainability use

31:39 of ai so our organisation is going to have to start to report on the i think it will happen i think it'll still be a few years away but i think that's something else that is going to have an impact and make it more expensive as i mentioned briefly earlier that we've got claude who just brought out their latest model fable 5 in the last couple of days and that's

32:01 free until june the 22nd and then it's going to be an extra cost for that so we're starting to see these are these companies wanting to squeeze a little bit more money out of us as well so i think it's uh i think it's going to be an extra cost for that so i think it's going to be an extra cost for i think it's likely to get more expensive i think over time so i think in some respects if you want

32:19 to try some of these things out i think my recommendation to people is do what you can with it now and if we think about when i talk about the the ace acronym we ultimately want to not just you use ai and develop from that we want to be developing things that require us to use ai less so i'm using it to develop code and to create some tools that i can use that don't

32:44 use ai so where i know i can i i'm on a predictable result and i can do that it doesn't need some intelligence it doesn't need much any thinking other than some you know some um some basic options but where i can build those things and it means i can be rerunning these applications that i'm building and the cost is my own you know sort of uh either the hosting or the local power

33:09 as opposed to using ai so that's the other thing is i would say yeah use ai while you can to build what you can but if you can use it to replicate the things that you're doing in a predictable way then in terms of building a tool to do that without using ai i think that's also a yeah an interesting thing to look at as well as using open source tools as well that you

33:31 can run on your computer because then at least then you know how much power you're using you know you're kind of almost limited to what you're using whereas when you throw something up i mean to be honest i have no idea i mean fable 5 is the most powerful and it's fastest the fastest one they've done so far so i have no idea what sort of um

33:51 what that's doing in terms of and i haven't even looked at the size of the model but i mean i imagine it's huge no i need i need to look into fable 5 i think yeah so i think that's that's going to be the challenge i think unfortunately it is going to i think it is going to get more expensive and the thing that i talking to businesses about is actually making sure again

34:12 this principle of of sustainable ai to me is using it in the in using the minimum you why you can so if you can use a lower model use the lower model you know use the more advanced stuff when you need to use it but again if you can build things out of it but like what you've done you know where you've used it

34:31 for inspiration and then you've drawn so there's an output that's been created with the ai and the human together and so then for therefore then you can use the output on an ongoing basis the danger is that if we become too much too reliant upon it we build all our processes around using ai then you know the stakeholders or these ai companies want that wants some return on the

34:55 investment yeah the price shoots up and we're all stuck there with well okay there's businesses particularly that we are now reliant on our cost basis shooting up because we've sat on that so as much as i'm an advocate of ai i am also on the other side of it an advocate of humans and wanted to make sure that we were able to do that and i think that's a big part of it and i think

35:14 that's reveal that we do things that are sustainable not just for the environment for ourselves as well and an interesting fact for you as well i don't know if you remember when when we were on the s2 course together when i looked into it the cost of a traditional tv series compared to the cost of specifically doing it all in ai is cheaper but the the impact on the

35:42 environment for using ai is quite a bit more than the cost of doing it all in ai i think it's quite the cost of doing it all in ai i think it's quite a bit more than the cost of doing it all traditionally for using ai for an entire series is way less than the traditional methods which is crazy we all know that ai is obviously using up a lot of um resources and it's it's individually it's not got a good

36:00 impact on the environment but if you look across the whole of a series or a film the film industry is doing far more harm with traditional methods than ai would do which is crazy yeah yeah and it's interesting isn't it that that's a really good observation because it does so much depend on what you're using it for you know so what is the alternative as you say you know we

36:24 a uh the initiatives they use within film and tv albert where they have to report the sustainability the reason it's in that industry isn't it it's because of the fact that they know there's an issue so they know there's an issue to issue to do that and addressing that with making more local productions not having to travel so so much and

36:43 all those different elements but as you say yeah we shouldn't just assume that you know take the line of ai is is you know less sustainable it does depend on what we're doing if i'm sitting here and i'm going i i want to search out some information just about i don't know um what movies are coming up in 2026 or 2027 and i go well i'm going to use fable five and i'm going to type that prompt

37:13 in that's that is a waste of resources overkill i could probably do it in one of the lower models like haiku even or sonnet so and for those that don't know we're talking about the the different models uh within different or the different ai um i think what is like chat gpt like claude and gemini have different levels you can choose and you can also in claude you can choose the model

37:38 and the effort they'll use so like low medium high for example so haiku is a low is a small model and now we've got term fable five and they'll keep on building other ones as well but we can reduce you know if we find the minimum minimum model we can use or the minimum effort that we can set our ai tools to to achieve the

38:03 result we want then obviously that is going to make it more effective but i think that's a really that is a great insight and we shouldn't just assume one that ai doesn't cost anything because it's easy for us to just type it in it's only costing us or 20 pound a month or whatever on the other hand we also shouldn't assume that it's always

38:23 going to be less sustainable than the alternatives because it's going to depend on the situation yeah definitely well i think yeah it's been a it's been a it's been an excellent and i'm looking forward to the um yeah to when this this one comes out because i i think you brought some real value out i think to some level of ceremony and to where you can just see the will see some real things in

38:48 action and also just if you've been able to describe the processes you know as as an artist as a as a 3d animator to walk us through and to notice those things that we wouldn't necessarily notice we wouldn't notice the bell's moving i mean would we even know you know might necessarily notice the swords on the wrong side yeah might notice the fact that the uh you know the warrior turns she

39:11 turns away from this i mean i think anybody would think that's that's the only thing that's so necessary a normal thing you're going to turn however it's but that's the thing isn't it we i think you know as human beings we're using ai again this principle of evaluation like we have to evaluate what we see to determine you know do we think it's real or not you know whether that's

39:32 news whether it's images or video and if we're not sure or even if we are sure maybe we should do our own research as well you know this gone are the days really well even some people would argue these never were these days that you could just take what you you know what you see or what you're shown as that being the truth some people would argue that's never been the case but

39:55 certainly now more and more we need to be careful of where we get information from you know we're in this we're in the doing this podcast and we're not we're not saying we're saying this is what we see and you're demonstrating your stuff but we're not actually advocating what is true what is not true you know we're having a conversation to help people navigate

40:12 that space because anything that we say on here if you want to know it for yourself go and check it for yourself go and make sure that that is the case it's the case don't just assume and just digest what is what it sent you in six months from the date of us recording this the world of ai will be completely different yeah we'll have to we'll have to do another

40:35 recording and see yeah see how far you've got on with your uh your movie of the book and uh and what you've learned from that because as you say we've already spoken about and and it's really interesting and you probably deliberately chose six months because what we're seeing is yeah we've seen a 12 month jump and then we're going to see a similar if not

40:54 bigger jump in six and then it's going to probably be an even bigger jump in three in the the speed of advancement because obviously now ai is being used to in to develop where it's using that model of saying right okay we can do this now how can we use ai to help us make ai better but ultimately we always need to have the uh

41:16 a human in there because if we need the human creativity to drive creativity inside of ai once we if we lose human creativity then we'll lose the ai's ability to be creative because it's only able to be creative through us and uh it's always better with humans isn't it glen yeah i could have said it better myself

41:39 yeah thanks again yes yeah thank you very much for having me it's been a pleasure