Why a Live-Action Producer Took the Leap Into AI | Toby Walsham, Made By Humans

WEBVTT
1
00:00:00.608 --> 00:00:07.149
Hi, and welcome to Technically Creative, where we talk to creative leaders about how technology is changing their companies and their industries.
2
00:00:07.749 --> 00:00:11.850
This show is brought to you by Kubrick Labs, the AI and IT partner for creative companies.
3
00:00:12.730 --> 00:00:16.751
Today's guest is Toby Walsham, founder of Imagine This and Made by Humans.
4
00:00:17.271 --> 00:00:23.092
And what I really like about Toby's story is that it isn't actually a story about choosing AI over traditional production.
5
00:00:23.573 --> 00:00:25.033
It's about refusing to choose.
6
00:00:25.733 --> 00:00:33.998
And then when Made by Humans did well, he grew Imagine This into a full-service creative studio and again refused to choose between the two companies.
7
00:00:34.358 --> 00:00:37.500
And now he has two successful companies that feed each other.
8
00:00:38.261 --> 00:00:42.083
Toby came up through production at a time when making things felt like an adventure.
9
00:00:42.483 --> 00:00:43.524
You went somewhere.
10
00:00:43.624 --> 00:00:45.045
You tried something difficult.
11
00:00:45.345 --> 00:00:47.026
You worked with extraordinary people.
12
00:00:47.286 --> 00:00:51.228
And yes, you were responsible for enormous budgets and serious clients.
13
00:00:51.649 --> 00:00:53.250
But the whole thing was supposed to be fun.
14
00:00:54.223 --> 00:00:57.065
That idea became one of the principles behind Imagine This.
15
00:00:57.585 --> 00:01:01.748
Make production fun again, be multidisciplinary, and innovate.
16
00:01:02.329 --> 00:01:07.152
And eventually, that desire to innovate became made by humans, Toby's AI creative studio.
17
00:01:07.812 --> 00:01:11.755
What I find interesting is that he didn't launch it because he thought live action was finished.
18
00:01:12.095 --> 00:01:18.540
In fact, he's still a live action producer and very much hopes to spend as much time on set as he does making AI work.
19
00:01:19.120 --> 00:01:25.208
He built it because he saw a new creative tool arriving and wanted to find out what would happen if he actually engaged with it.
20
00:01:25.629 --> 00:01:27.852
He stayed curious and he stayed optimistic.
21
00:01:28.452 --> 00:01:32.338
And there's a line Toby says near the end of this conversation that I think sums up the whole thing.
22
00:01:32.938 --> 00:01:34.661
He would have regretted not trying.
23
00:01:34.961 --> 00:01:35.482
I love that.
24
00:01:35.762 --> 00:01:40.444
Because so much of the conversation around AI and production has been framed around prediction.
25
00:01:40.624 --> 00:01:41.685
Will it replace this?
26
00:01:42.025 --> 00:01:43.186
Will it destroy that?
27
00:01:43.606 --> 00:01:45.307
Toby took a much more practical approach.
28
00:01:45.387 --> 00:01:46.847
Just start building something.
29
00:01:47.168 --> 00:01:47.888
Experiment.
30
00:01:48.228 --> 00:01:50.289
Learn what the technology is genuinely good at.
31
00:01:50.729 --> 00:01:52.970
Understand where traditional production still matters.
32
00:01:53.330 --> 00:01:55.812
And then figure out what the company should become as a result.
33
00:01:56.552 --> 00:02:01.556
We talk about building an AI creative business while continuing to champion live action production.
34
00:02:02.076 --> 00:02:04.918
Why the conversation around AI attracts so much attention.
35
00:02:05.559 --> 00:02:09.562
And how creative companies can expand beyond the traditional production company model.
36
00:02:09.742 --> 00:02:15.226
And why curiosity might be a much more useful response to technological change than certainty.
37
00:02:15.646 --> 00:02:21.711
But more than anything, this is a conversation about having the courage to try the thing before you know exactly where it's going to lead.
38
00:02:22.171 --> 00:02:23.152
This is Toby Walsham.
39
00:02:37.022 --> 00:02:38.965
Toby Walsham, I'm so psyched to have you on.
40
00:02:39.005 --> 00:02:44.732
You and I have been connected for quite some time, and I've just been watching what you've been working on with great fascination.
41
00:02:45.112 --> 00:02:46.795
So I'm thrilled to have you on the podcast.
42
00:02:47.055 --> 00:02:48.477
Yeah, thank you for having me.
43
00:02:48.517 --> 00:02:52.342
Likewise, Orlando, it's been fascinating watching all these podcasts over the last kind of
44
00:02:52.842 --> 00:02:58.624
I kind of discovered the maybe six months ago and you know what you're doing Kubrick's labs is amazing as well.
45
00:02:58.644 --> 00:02:59.624
So yeah, thank you for having me.
46
00:02:59.864 --> 00:03:00.465
It's a pleasure.
47
00:03:00.965 --> 00:03:05.346
So listen, you're the CEO and founder of both Imagine This and Made by Humans.
48
00:03:05.686 --> 00:03:09.888
And they're, you know, it's a huge job running two separate but equal companies.
49
00:03:10.268 --> 00:03:13.749
But could you just explain the two companies just for anybody who doesn't know?
50
00:03:14.189 --> 00:03:14.509
Sure.
51
00:03:14.649 --> 00:03:18.832
I mean, to explain the reason there's two, I probably need to back up a bit.
52
00:03:19.612 --> 00:03:24.575
When I decided to launch Imagine This, it was out of necessity.
53
00:03:24.675 --> 00:03:35.080
I had left my previous company and I only know how to do one thing and I have a passion for only one thing and that's, you know, making films, commercials and I'm a producer and that's what I wanted to do.
54
00:03:35.420 --> 00:03:41.143
But when I launched Imagine This, I didn't want to just open another production company.
55
00:03:41.223 --> 00:03:41.744
You know, my...
56
00:03:42.404 --> 00:03:49.010
15 years prior to that working in the industry, I had noticed that production companies kind of look remarkably similar.
57
00:03:49.050 --> 00:03:52.853
You know, they have black backgrounds, white logos, similar websites.
58
00:03:52.874 --> 00:03:57.257
You know, really the thing that makes them stand apart from the next is...
59
00:03:57.958 --> 00:04:00.640
if they've got a better director than the other one.
60
00:04:01.360 --> 00:04:08.845
And I wanted to continue doing the production work that I loved, but I didn't want to just do another copy paste production company.
61
00:04:08.865 --> 00:04:10.966
The market's already so oversaturated.
62
00:04:11.066 --> 00:04:17.911
So you'll see that's why our websites and our branding is very colorful, is very intentional, but that's not enough to,
63
00:04:19.111 --> 00:04:21.372
to make yourself different from one another.
64
00:04:21.752 --> 00:04:26.574
For years, I'd always have clients call me up and go, hey, do you know an amazing animator?
65
00:04:26.734 --> 00:04:29.035
Or who's the best VFX person for this?
66
00:04:29.095 --> 00:04:30.916
Or what about composers and so on and so on.
67
00:04:30.936 --> 00:04:36.198
And I had this amazing little black book of incredible people that I've collaborated with.
68
00:04:36.218 --> 00:04:36.878
And I just figured,
69
00:04:37.418 --> 00:04:41.262
Why not represent all the people that I want to continue collaborating with?
70
00:04:41.502 --> 00:04:46.107
So we built a multidisciplinary roster and that was something I wanted to do.
71
00:04:46.127 --> 00:04:52.653
But something I'd also been playing with for about two years, much like yourself, you know, started playing in the AI space in about 2022 was AI.
72
00:04:52.673 --> 00:04:54.435
And I wanted to launch AI.
73
00:04:57.037 --> 00:05:07.324
imagine this with an AI roster, but I kind of got a little bit scared at the time AI, you know, there was like WhatsApp groups and Facebook groups protesting against AI.
74
00:05:07.464 --> 00:05:13.688
And it felt a little bit risky to launch a brand new business that no one had ever heard of with also an AI roster.
75
00:05:14.468 --> 00:05:15.169
Very wise.
76
00:05:15.229 --> 00:05:20.372
We lived through so much of the lynching of of AI companies and anybody who would talk about it.
77
00:05:21.032 --> 00:05:22.674
It was a it was a very tricky time.
78
00:05:23.074 --> 00:05:23.934
It quickly changed.
79
00:05:24.094 --> 00:05:30.058
And six months later, I was brave enough to to make that announcement.
80
00:05:30.098 --> 00:05:31.719
But it made sense at that time.
81
00:05:31.920 --> 00:05:33.761
And that was then made by humans.
82
00:05:34.121 --> 00:05:41.708
That's when Made by Humans was formed with Guy Soldby, our chief creative officer.
83
00:05:41.948 --> 00:05:45.471
So, you know, we started that business separately, but that's just the businesses.
84
00:05:45.711 --> 00:05:46.732
They complement each other.
85
00:05:46.792 --> 00:05:52.637
And certainly what we're finding is hybrid work is becoming more and more commonplace.
86
00:05:52.697 --> 00:05:54.119
You know, something shot.
87
00:05:54.819 --> 00:05:57.580
and AI is used to assist that production.
88
00:05:59.101 --> 00:06:03.742
So that was why we kept them separate and why there's two entities.
89
00:06:04.362 --> 00:06:11.765
But I certainly didn't set out to make this network or collective or ecosystem, but that's certainly where my head is at now.
90
00:06:11.825 --> 00:06:23.109
You know, imagine this group, you know, working title, it is growing and there's new entities and businesses that are being acquired and built at the moment, which is all very exciting.
91
00:06:23.879 --> 00:06:33.305
So I imagine this was, for want of a better word, a sort of creative collective or more like agency style kind of collective situation.
92
00:06:33.685 --> 00:06:50.675
And then Made by Humans became, on the face of it, traditional looking in terms of branding and it being a quote unquote production company, but immediately leaning into AI workflows, AI tooling, and individual creators who were able to create...
93
00:06:51.355 --> 00:06:52.615
from their own platforms.
94
00:06:52.695 --> 00:06:56.676
And now it's merged into hybrid production pipelines as well.
95
00:06:56.716 --> 00:06:58.997
Is that a fair assessment of the situation?
96
00:06:59.337 --> 00:07:02.798
Yeah, imagine this in many ways is quite a traditional production company.
97
00:07:02.978 --> 00:07:10.159
You know, we just represent more than directors and photographers, which is what most of our peers kind of end at.
98
00:07:10.679 --> 00:07:14.680
And on the Made by Humans side, yeah, it's also a production company.
99
00:07:15.900 --> 00:07:17.381
We just happen to be
100
00:07:18.381 --> 00:07:20.022
our specialty is AI.
101
00:07:21.724 --> 00:07:24.146
Sorry, I forgot what your second question was on me.
102
00:07:24.186 --> 00:07:32.072
No, no, it was basically just saying, is that a fair assessment of what, you know, what made by humans and imagine this kind of are?
103
00:07:33.052 --> 00:07:33.613
But that's okay.
104
00:07:33.693 --> 00:07:35.354
You're talking about hybrid production.
105
00:07:35.434 --> 00:07:37.736
I'm very fascinated by a lot of the stuff that we've seen.
106
00:07:37.756 --> 00:07:40.758
I was just at AI on the lot two weeks ago.
107
00:07:41.419 --> 00:07:42.540
That was at MGM.
108
00:07:42.640 --> 00:07:44.902
They did a lot of talks on the volume stage.
109
00:07:45.382 --> 00:07:59.309
that AWS Luma Labs and Innovated Dreams did their House of David production on, which is a feature length production starring Ben Kingsley that took only two weeks to film on a hybrid stage.
110
00:07:59.909 --> 00:08:10.054
And I think that like what I found really interesting about that was that the hybrid production model for me is kind of the most interesting model there is.
111
00:08:10.134 --> 00:08:12.415
Like I never actually liked shoot days.
112
00:08:12.955 --> 00:08:16.577
I found that I really like bidding and figuring out how to do the project.
113
00:08:17.017 --> 00:08:19.418
I really like pre-production.
114
00:08:19.538 --> 00:08:21.239
I really like post-production.
115
00:08:22.200 --> 00:08:24.621
But the shoot days I found incredibly boring.
116
00:08:25.061 --> 00:08:34.986
And actually, when you're able to bring post-production, pre-production, production designers, working with post-production personnel all onto a volume stage, I was like, oh, this to me is...
117
00:08:35.506 --> 00:08:42.074
Instead of having the collaboration be so linear and the producer and director are really the only two that go throughout the whole thing.
118
00:08:42.434 --> 00:08:49.383
I love this idea of taking multiple departments from either side of the actual shoot day and cramming them into the shoot day itself.
119
00:08:50.500 --> 00:08:52.543
Well, I love the shoot day.
120
00:08:52.563 --> 00:08:54.365
You know, that's why I got into the business.
121
00:08:54.385 --> 00:08:56.548
So I don't agree there.
122
00:08:56.628 --> 00:09:07.201
But I do find virtual production utilizing Unreal AI physical set builds.
123
00:09:08.903 --> 00:09:13.806
you know, I do find that to be possibly what the future of filmmaking would be.
124
00:09:14.166 --> 00:09:20.930
Certainly, you know, we just came off the back of a project, which I think I can speak about, but maybe you beep it out if I find out later.
125
00:09:21.030 --> 00:09:21.870
That's totally fine.
126
00:09:22.150 --> 00:09:22.370
Yeah.
127
00:09:22.851 --> 00:09:28.494
We just finished a project for PwC with McCann and their script was
128
00:09:29.434 --> 00:09:51.404
very ambitious their budget wasn't small in in in no way small but you know the ideas are always pushing the budgets right it was a big idea and it was not possible to just do that as a live production their brief was also trying to position pwc as this innovative brand at the forefront of ai so
129
00:09:53.465 --> 00:09:59.150
It kind of also made logical sense to look at doing virtual production, which was AI assisted.
130
00:09:59.170 --> 00:10:01.652
You know, we made our animatics using AI.
131
00:10:01.692 --> 00:10:06.956
We created the backgrounds with Unreal, but there was AI integrated into that process too.
132
00:10:07.557 --> 00:10:18.166
We had an amazing art director of production designer who I've been working with for years, Mark Connell, fly over to Poland where we were shooting and design these amazing grand sets.
133
00:10:18.606 --> 00:10:28.610
But outside the windows of all these sets was the virtual production stage, which allowed us to have that control and it was all one continuous move.
134
00:10:28.970 --> 00:10:38.174
But there's no way that project would have worked with us flying around the world to these different locations or just budget constraints, which wouldn't have allowed for that.
135
00:10:38.655 --> 00:10:46.498
So I do think that's going to be more commonplace is the kind of virtual production, hybrid AI approach to productions.
136
00:10:47.206 --> 00:10:50.927
What sort of freedom did it give you on the shoot day that you wouldn't have had otherwise?
137
00:10:51.448 --> 00:11:05.653
In many ways it doesn't give you freedom because with virtual production and the level of planning we had to do, because it was all one continuous shot, everything had to be perfect in order to match the next shot.
138
00:11:05.813 --> 00:11:07.373
So we had no freedom, really.
139
00:11:07.393 --> 00:11:07.793
Right.
140
00:11:07.853 --> 00:11:09.414
The camera moves had to be what it was.
141
00:11:09.614 --> 00:11:32.713
everyone specifically dictates very little freedom and everything but you have the freedom you do have freedom in the pre-production you know you can you could look at a thousand we had one of our scenes for example was the top of a skyscraper building and what the question to the team and the creatives is what does the vista look like is it new york is it shanghai you know where in the world are we so we had that flexibility to
142
00:11:33.193 --> 00:11:57.142
design a new york inspired skyline hate it and then design something else whilst of course if you're scouting and you've you know purchased to be on the 55th floor of a yeah you know one world trade center that's where you're shooting yeah so there was flexibility in that regard uh but once the shoot day came about it was very much a yeah factory situation like this is the shot it's perfect let's move on
143
00:11:58.082 --> 00:12:00.824
Is there post-production collaboration on the shoot day?
144
00:12:00.944 --> 00:12:03.206
Are you getting into the volume stage ahead of time?
145
00:12:03.566 --> 00:12:07.528
Is it a multi-day pre-light?
146
00:12:07.828 --> 00:12:14.352
How does it actually fit in with the conception of the previous production model that you know very well?
147
00:12:14.392 --> 00:12:16.994
My background, my career started at Framestore.
148
00:12:17.314 --> 00:12:17.474
Yeah.
149
00:12:18.635 --> 00:12:24.199
almost every production I worked on was incredibly technically advanced for its time.
150
00:12:24.259 --> 00:12:36.248
I even was a young PA on Gravity, which if you know about Gravity, was really the foundations of virtual production, the amazing box that Alfonso Cuaron and Zulo designed.
151
00:12:37.970 --> 00:12:43.475
So from my perspective, this project we did was very, I felt very comfortable in.
152
00:12:43.515 --> 00:12:51.002
It's kind of commonplace, big VFX, puzzles that need to be solved is the type of producing I like.
153
00:12:51.422 --> 00:12:54.565
But you made a point about post-production and how much is involved.
154
00:12:54.585 --> 00:12:55.966
And I think it's a big mis-
155
00:12:56.927 --> 00:13:03.832
understanding or misconception with AI right now and that is that AI is being blamed for lots of jobs disappearing.
156
00:13:04.192 --> 00:13:16.960
I can say for sure that I hire far more people in the post-production world now than I ever did before and you know anyway isn't this magic button that you know is giving us these results that we can just put out straight away.
157
00:13:17.001 --> 00:13:26.347
We hire teams of retouchers when we're working on stills and VFX experts to fix you know there's still artifacts in that in that AI world.
158
00:13:27.408 --> 00:13:37.107
in fact you know since launching made by humans i feel that we have employed far more people than we were just running imagine this
159
00:13:38.008 --> 00:13:49.177
But just to get into the specificity of it, one of the things that I sort of always advocate for is I think it's the people with the hard-won experience of traditional production, traditional creation that need to be the ones picking up these tools.
160
00:13:49.357 --> 00:14:01.926
Now working in this new AI hybrid environment or fully AI environment, what are the primary differences that you're noticing in terms of how the workflow works, how collaboration works?
161
00:14:02.126 --> 00:14:04.048
What are you noticing are the primary differences?
162
00:14:04.068 --> 00:14:04.388
Yeah.
163
00:14:04.488 --> 00:14:08.871
I wish there was like a real standout thing that feels incredibly different.
164
00:14:08.891 --> 00:14:13.575
But the way we try to approach things is still with craft at the very forefront of things.
165
00:14:14.135 --> 00:14:27.745
We're rebranding our AI artists as AI directors because we think artists maybe is more connected with people that are given a task to do a thing rather than someone that's given a spark of an idea and execute and embellish on that.
166
00:14:28.965 --> 00:14:34.989
I think there are some fundamental differences, absolutely, certainly on our fully AI productions.
167
00:14:35.109 --> 00:14:38.471
You know, we're not on set, we're not working with...
168
00:14:38.491 --> 00:14:40.212
The collaboration does shrink.
169
00:14:40.232 --> 00:14:43.854
You're not hiring a DP or working with an electrical team.
170
00:14:44.494 --> 00:14:46.035
It's crafted in a different way.
171
00:14:46.055 --> 00:14:48.677
But I think fundamentally,
172
00:14:49.701 --> 00:14:55.004
there is still this, well, certainly we try to make it feel very familiar for our clients.
173
00:14:55.044 --> 00:15:03.088
And we have this pipeline that feels very much like a traditional pipeline where we start with location scouting, but everything's AI.
174
00:15:03.188 --> 00:15:04.769
So what is that background?
175
00:15:04.829 --> 00:15:09.611
Using the PwC, for example, what does a New York skyline look like?
176
00:15:09.651 --> 00:15:10.812
Or what if it was Shanghai?
177
00:15:10.832 --> 00:15:11.992
And then we do that fully in AI.
178
00:15:12.072 --> 00:15:15.834
And then we do castings to people, if there's people or characters.
179
00:15:16.334 --> 00:15:18.916
And again, fully AI, but we do casting documents.
180
00:15:19.636 --> 00:15:25.559
So we follow this traditional pipeline to, I guess, to make it feel quite familiar for our clients.
181
00:15:26.560 --> 00:15:29.761
But fundamentally, a director is leading those decisions.
182
00:15:30.182 --> 00:15:34.304
A team of AI artists, AI directors are supporting that person's vision.
183
00:15:34.924 --> 00:15:36.785
So it sounds like it's more about sequencing.
184
00:15:36.825 --> 00:15:42.708
Like instead of waiting until post to do your plates, you're doing your plates in pre-production.
185
00:15:42.748 --> 00:15:47.811
You're just front-loading that work instead of doing it in post or instead of shooting your place elsewhere.
186
00:15:47.871 --> 00:15:48.051
Yeah.
187
00:15:48.371 --> 00:15:53.692
Yeah, definitely when it comes to the virtual production side of things, it's all about the pre-production.
188
00:15:53.972 --> 00:16:12.336
I mean, those shoots have generally always been about pre-production, but what's almost always happened on virtual production jobs I've done before, and hopefully producers that are listening also agree, and it's not just me, is that we are roto-ing and comping those screens anyway, because the client changes something, there's an update to what needs to go on that screen.
189
00:16:12.596 --> 00:16:36.053
change it but we were very clear as to our process on this one so we have to have this all signed off in advance because the way it interacts with people's faces is really important that light spill and everything else but i think that's the vfx producer head that i have you know i want i do do a lot of prep up front so the post is easy the the pwc job that you just did it sounds like a sort of fairly traditional
190
00:16:36.934 --> 00:16:44.116
You know, it sounds like a fairly traditional studio shoot just with Unreal Engine enabled green screen backdrop.
191
00:16:44.136 --> 00:16:45.556
You're building sets inside.
192
00:16:46.116 --> 00:16:50.337
You're shooting it the way you would if you were on a stage normally.
193
00:16:51.278 --> 00:16:53.598
It doesn't actually change anything about the shoot day.
194
00:16:54.058 --> 00:16:55.138
Is that a fair assessment?
195
00:16:55.439 --> 00:16:56.579
I think the difference...
196
00:16:57.339 --> 00:17:00.062
was that we were able to be really efficient.
197
00:17:00.282 --> 00:17:02.364
Utilising AI allowed us to be more efficient.
198
00:17:02.424 --> 00:17:15.116
We managed to do what would have needed four or five shoot days in three because of all the speed at which we could go through these shots because of all the planning and all the AI assistance that we had in place.
199
00:17:15.736 --> 00:17:15.996
Yeah.
200
00:17:16.676 --> 00:17:22.258
And then I suppose that the fully AI production pipelines look very similar to an animation production pipeline.
201
00:17:22.458 --> 00:17:31.222
Yeah, I think the pipeline and the way that certainly we run those productions feels closer to a animation or post-production pipeline.
202
00:17:31.482 --> 00:17:32.002
Yeah.
203
00:17:32.622 --> 00:17:39.945
But for our clients, we give that experience of a live action thing with the casting and the locations.
204
00:17:40.825 --> 00:17:41.025
Yeah.
205
00:17:41.425 --> 00:17:51.948
Because often, Orlando, certainly in the early years of launching Made by Humans is we would get a phone call going, hey, can you look at this from an AI point of view, but can you also quote it live action?
206
00:17:52.408 --> 00:17:57.030
Because people were unsure if they wanted to explore a fully AI or hybrid thing.
207
00:17:57.710 --> 00:18:09.964
We were bidding everything three ways, which was a lot of work, but actually quite interesting for us because we were seeing lots of stats out there in the world of, oh, AI is 20% of the cost of live action.
208
00:18:10.044 --> 00:18:12.467
It absolutely is not.
209
00:18:13.288 --> 00:18:17.573
So we were able to look at those comparisons because we were quoting every job every different way.
210
00:18:18.040 --> 00:18:19.021
That brings up a good point.
211
00:18:19.041 --> 00:18:21.364
Like, what are clients asking for right now?
212
00:18:21.424 --> 00:18:35.720
Well, I can instantly see when I receive a script from brand or agency whether or not they understand AI and if they're educating that space based on the creative ambition and the budget they assign to it.
213
00:18:36.501 --> 00:18:43.508
Because some people, I mean, certainly in the early days of Made by Humans, we'd get phone calls or emails going, hey, I've got 10 grand, we want to make a TV commercial.
214
00:18:43.528 --> 00:18:45.570
So, well, you know, not with us.
215
00:18:46.311 --> 00:18:53.579
But I think that education is getting better because I'm seeing budgets in the AI space become much more realistic.
216
00:18:53.919 --> 00:18:54.119
The...
217
00:18:54.780 --> 00:18:59.203
Creative ambitions are always getting bigger and bigger, and I love that about this industry.
218
00:18:59.243 --> 00:19:08.068
Even with our, you know, forget AI for a minute, budgets have been decreasing over the last decade, and ideas still are getting bigger and bigger, and that's very exciting.
219
00:19:08.408 --> 00:19:17.834
And I think that's what I find really exciting about AI, is not that it's this shortcut, it's unlocking the ability to scale again.
220
00:19:18.334 --> 00:19:21.896
You know, the client's budget is what the client's budget is.
221
00:19:22.196 --> 00:19:26.838
The agency's idea that they've sold in almost always isn't possible for the budget.
222
00:19:27.158 --> 00:19:30.260
And that's where we come in to come up with solutions as production companies.
223
00:19:30.720 --> 00:19:44.547
And I think having the AI side of the business has allowed us to go, okay, we can do that because we can do shot three and six in AI, which are the big drum shots that don't require us to fly over there or do something ridiculously expensive.
224
00:19:45.227 --> 00:19:48.509
And that's allowed us to unlock scale like never before.
225
00:19:48.869 --> 00:19:50.470
What does that look like right now?
226
00:19:50.610 --> 00:19:53.812
Are you finding it all still a little bit lazy?
227
00:19:53.913 --> 00:19:59.056
Or are you finding that there are certain creatives, creative directors and clients that are getting the medium?
228
00:19:59.156 --> 00:20:03.859
Are you starting to see better, more natively smart AI things?
229
00:20:04.019 --> 00:20:07.401
Or are you still mainly just seeing cost savings requests?
230
00:20:07.521 --> 00:20:10.243
No, we definitely see scripts come that have been...
231
00:20:12.064 --> 00:20:18.947
scoped with AI in mind and that they're definitely, agencies and brands, there's definitely something to understand the capabilities more there.
232
00:20:19.427 --> 00:20:25.390
But, you know, maybe I shouldn't say this, but I think there's slops on both sides of the aisle.
233
00:20:25.470 --> 00:20:29.813
You know, of course there's bad eye work, but there's one in 10- It's a really implied action.
234
00:20:29.893 --> 00:20:31.814
Pieces of commercial that are actually good.
235
00:20:32.414 --> 00:20:41.897
And I'm guilty as I'm sure the rest of us are making some very average commercials because the script was and the budget didn't allow for certain things or whatever it was.
236
00:20:41.937 --> 00:20:46.718
But I also think there is a, you know, people hating on AI work right now.
237
00:20:47.038 --> 00:20:47.979
Actually, there's a brilliant example.
238
00:20:47.999 --> 00:20:59.022
I don't know if you ever saw this, Orlando, but someone posted on Twitter or X a kind of a quote that said, hey, I took a bunch of Monet paintings and...
239
00:21:00.002 --> 00:21:28.012
you see all that and made a made a new uh painting and posted it and then all the comments were like the brush strokes are crap it definitely this definitely isn't a real one the color choices are awful the compositions rub it basically just calling it ai stop and then later he revealed he just picked a slightly more obscure monet and posted it just to show that everyone is just hating on it uh and trying to find fault when actually that was an iconic masterpiece by one
240
00:21:29.777 --> 00:21:33.458
Yeah, I think that's a really good point is that there's slop on both sides of the aisle.
241
00:21:33.478 --> 00:21:36.518
And I've always kind of looked at it through an economic lens.
242
00:21:36.558 --> 00:21:43.240
I'm like, listen, there are loads of things I've made that were probably never going to be Can Lion winners.
243
00:21:43.300 --> 00:21:45.280
And we just went, oh, that's a nice piece of work.
244
00:21:45.340 --> 00:21:47.121
Like, oh, that's a funny little joke.
245
00:21:47.181 --> 00:21:52.722
And that job will keep the lights on and it'll pay our staff.
246
00:21:52.802 --> 00:21:55.522
Not all work is intended to reach from the stars.
247
00:21:55.602 --> 00:21:57.183
Sometimes you need a retail ad.
248
00:21:57.463 --> 00:21:57.703
Yeah.
249
00:21:57.783 --> 00:21:58.023
Right.
250
00:21:58.625 --> 00:21:58.905
Right.
251
00:21:59.546 --> 00:22:01.387
Sometimes you do just do a retake on.
252
00:22:01.968 --> 00:22:08.333
When you get a brief in, are you able to articulate what you think makes a good AI brief?
253
00:22:08.353 --> 00:22:11.095
Creativity remains incredibly difficult.
254
00:22:11.795 --> 00:22:15.679
That, I would say to creative directors, write something brilliant.
255
00:22:15.919 --> 00:22:17.440
Don't write with AI in mind.
256
00:22:18.080 --> 00:22:21.523
We look at scripts going, how can we make this?
257
00:22:21.743 --> 00:22:23.665
Not, let's use AI.
258
00:22:23.785 --> 00:22:27.068
You know, if AI can assist us in making it and be...
259
00:22:28.057 --> 00:22:31.118
help with the execution being more efficient and cost-saving than brilliant.
260
00:22:31.218 --> 00:22:36.240
But I would just love scripts to send to us, going, hey, we've got this script.
261
00:22:36.381 --> 00:22:38.261
How do you want to make it?
262
00:22:38.281 --> 00:22:40.722
What's the best route to success?
263
00:22:42.825 --> 00:22:49.287
But, you know, what I would love agencies and brands to do more often if they're thinking of using AI is contact us early.
264
00:22:49.307 --> 00:23:01.892
In fact, we're working on a project right now where they've contacted us about a month before they plan to even put it out to pitch just to do R&D and work out if the script that they have put together is feasible.
265
00:23:02.592 --> 00:23:07.074
I think that's a really nice thing to do, just bring people in early.
266
00:23:07.494 --> 00:23:10.615
Do you think it's two different skills a little bit?
267
00:23:13.228 --> 00:23:24.178
feel about AI is that when theoretically anything is possible, I think it's a different mindset that finds its way through that.
268
00:23:24.579 --> 00:23:27.481
It's a different, it's more of an auteur.
269
00:23:27.902 --> 00:23:36.629
Whereas I know a lot of people, a lot of directors especially, they thrive with the problems that production throws up to them.
270
00:23:37.310 --> 00:23:42.711
They almost become better directors because of the friction of live action production.
271
00:23:43.111 --> 00:23:47.752
I'm not sure that's the same skill set of going blank page.
272
00:23:48.652 --> 00:23:54.154
I can do anything I want and I can tell any story and make it look any way.
273
00:23:54.594 --> 00:23:56.854
And I can do that with specificity.
274
00:23:56.914 --> 00:23:57.854
That's interesting.
275
00:23:58.454 --> 00:24:00.215
I think those are two different skill sets.
276
00:24:00.895 --> 00:24:04.796
I wish that was the case that we receive scripts that say you can do anything.
277
00:24:05.096 --> 00:24:05.536
Right.
278
00:24:06.176 --> 00:24:08.056
That would be amazing.
279
00:24:08.317 --> 00:24:09.297
But no, you're absolutely right.
280
00:24:09.337 --> 00:24:20.599
And, you know, I think you're correct that these restrictions that creatives are given, sometimes the narrower those restrictions, the more creative they have to be to solve those problems.
281
00:24:20.639 --> 00:24:22.040
And I think there's a lot of truth to that.
282
00:24:22.600 --> 00:24:28.181
I think in the AI space, we're still working to scripts and there are still absolutely restrictions.
283
00:24:28.221 --> 00:24:30.662
Whilst, you know, from a free...
284
00:24:30.782 --> 00:24:32.142
I can't speak now.
285
00:24:32.842 --> 00:24:35.383
You know, you're absolutely right that...
286
00:24:36.383 --> 00:24:41.347
But from an AI perspective, I think there are still restrictions.
287
00:24:42.107 --> 00:24:52.795
And whilst you can, you know, make anything with AI, the reality is when you actually start using the tools, it's not a magic button that you click and it gives you the script exactly how you saw it.
288
00:24:53.415 --> 00:25:01.701
Our artists, our directors are hitting their heads against brick walls every single day, trying to get the tools to work exactly as they see fit.
289
00:25:01.721 --> 00:25:01.841
Yeah.
290
00:25:02.461 --> 00:25:04.162
and we're learning something on every project.
291
00:25:04.182 --> 00:25:07.323
So there are absolutely times where we're on a project, we go, right, we've got to pivot.
292
00:25:07.664 --> 00:25:15.147
We've got to pitch a different route, or we've got to, more often than not, we've got to build this in multiple layers, and we've got to bring in VFX, and we've got to do this.
293
00:25:16.068 --> 00:25:24.672
I would say 0% of the AI shots that we put out have come out exactly 100% perfect from the tool.
294
00:25:25.492 --> 00:25:28.415
Every single one has been built up through layers.
295
00:25:28.935 --> 00:25:33.579
And I think that's something that people have not quite understood about AI.
296
00:25:33.619 --> 00:25:38.884
They think it's write script, put script into AI, spit out final film.
297
00:25:38.924 --> 00:25:40.005
It is absolutely not.
298
00:25:40.025 --> 00:25:44.890
There's a lot of craft and pivoting and solutions that have been found.
299
00:25:45.190 --> 00:25:48.353
How much does it look like just a regular post-production workflow?
300
00:25:48.743 --> 00:25:54.987
I think on the surface it might look like that, but I don't know how much you use a Weavey or a Comfy type tool.
301
00:25:56.188 --> 00:25:57.569
They're like node-based tools, right?
302
00:25:57.589 --> 00:26:03.113
So you can make connections between all the various prompts and rules that you set.
303
00:26:04.034 --> 00:26:07.596
The web of mess becomes quite complicated quite quickly.
304
00:26:08.037 --> 00:26:10.979
It certainly depends on the complex of the show, of course.
305
00:26:11.499 --> 00:26:14.321
And that's what feels to me like a regular post-production pipeline.
306
00:26:14.341 --> 00:26:17.664
I mean, any regular post-production pipeline is kind of a fucking mess.
307
00:26:18.644 --> 00:26:19.305
Yeah, it is.
308
00:26:19.805 --> 00:26:20.145
It is.
309
00:26:20.805 --> 00:26:31.070
But you know, on that point, I think when we launched Made by Humans, we just signed a roster of every cool artist we found via almost like Instagram and LinkedIn was the way to find artists back then.
310
00:26:31.730 --> 00:26:35.332
And we just signed this roster of I think it was 10 at the time of cool people.
311
00:26:35.812 --> 00:26:37.393
And then when we started winning work,
312
00:26:38.153 --> 00:26:42.778
we started to realise that actually not everyone in this roster is capable of delivering that work.
313
00:26:43.439 --> 00:26:48.605
They're all brilliant artists that have a point of view, a specific look.
314
00:26:49.046 --> 00:26:54.812
But a big part of what we do as filmmakers in the commercial side of the business is we take a...
315
00:26:55.493 --> 00:27:00.237
a thing from an agency, a script, we interpret it and we then deal with feedback.
316
00:27:00.277 --> 00:27:06.662
And it's like a therapy session that we were on for a few months with a team and we're finding the solutions.
317
00:27:07.402 --> 00:27:12.687
And we have, there's a couple of, I won't name them, but we have a couple of amazing artists that we signed early on.
318
00:27:13.287 --> 00:27:30.130
that their day jobs were, funnily enough, one of them was a plastic surgeon by trade, but just dived into the AI world early, made these incredible works of art, but wasn't the right person to manage a creative team of widens, right?
319
00:27:30.490 --> 00:27:34.351
It is a skill set in itself, a director, just on that side of things.
320
00:27:34.991 --> 00:27:38.692
What are the skill sets that are emerging to you when you look at AI directors?
321
00:27:39.413 --> 00:27:44.634
I mean, I think everyone can access Mid Journey, everyone can access Runway.
322
00:27:44.734 --> 00:27:46.595
It's like $19 a month, right?
323
00:27:46.615 --> 00:27:50.956
But not everyone has taste or everyone can tell a story.
324
00:27:51.996 --> 00:27:53.576
I would be a dreadful AI director.
325
00:27:53.596 --> 00:27:54.896
I'd be a dreadful director.
326
00:27:55.256 --> 00:27:56.417
That's just not what I do.
327
00:27:57.397 --> 00:28:01.978
And I think that is still the key skill set of everyone that we
328
00:28:02.558 --> 00:28:10.204
you know, have on our roster and work with is that they have good taste and they're storytellers before they started exploring the AI tools.
329
00:28:10.385 --> 00:28:18.471
And actually, everyone on our roster was, you know, most people on our roster was a traditional live action director or photographer or creative directors.
330
00:28:18.751 --> 00:28:27.599
They all did that storytelling, creative taste-making stuff before they started playing with the AI tools.
331
00:28:28.099 --> 00:28:37.883
You've written about the gap between good and great widening, and that feels counterintuitive because obviously if these tools are supposed to democratize production, why is that gap widening then?
332
00:28:38.663 --> 00:28:39.764
And is that taste?
333
00:28:39.784 --> 00:28:41.064
Is that what it boils down to?
334
00:28:41.945 --> 00:28:44.025
I think it honestly just comes down to taste.
335
00:28:46.563 --> 00:28:56.952
What I loved about AI in the early days, and I spoke about it quite a lot in articles and interviews, was that it gave everyone access to be a director.
336
00:28:57.412 --> 00:29:06.580
If you wanted to break into the live action world, to go shoot your first short film, you had to still pull together a team of people willing to work for free or find some funding to do it.
337
00:29:07.141 --> 00:29:11.684
And the cost to entry in the live action world is still expensive.
338
00:29:12.583 --> 00:29:16.246
And that's why it's not a very diverse space.
339
00:29:17.086 --> 00:29:26.774
What we found in the early days of AI, and still true now, is that there were tons of female AI artists.
340
00:29:27.354 --> 00:29:34.019
People were accessing these tools anywhere in the world with no background necessarily in the creative fields.
341
00:29:34.039 --> 00:29:35.400
I thought that was really interesting.
342
00:29:37.302 --> 00:29:50.338
But like I said, when we start actually booking paid work, someone that understands what we do in the advertising space and has that deep learning and history of creativity is very important still.
343
00:29:50.638 --> 00:29:50.939
Yeah.
344
00:29:51.920 --> 00:30:00.463
I'm a director, producer, creative, VFX artist or stylist or production designer, and I'm watching this and I have taste, but I'm not AI fluent yet.
345
00:30:00.883 --> 00:30:02.944
Where do you suggest starting?
346
00:30:03.004 --> 00:30:09.767
I mean, what is it that you want to see from somebody who's producing AI or directing AI filmmaking?
347
00:30:10.067 --> 00:30:11.768
Like, where do they start on their own, you think?
348
00:30:12.069 --> 00:30:14.691
If someone wants to get into AI, is that what you're saying?
349
00:30:15.071 --> 00:30:15.271
Yeah.
350
00:30:15.371 --> 00:30:17.073
I don't think everyone needs to get on the tools.
351
00:30:17.613 --> 00:30:19.274
I don't think that's required.
352
00:30:19.374 --> 00:30:25.779
And I think what we'll find eventually is that AI is just another tool in our arsenal, much like...
353
00:30:26.722 --> 00:30:29.703
filming with practical effects or using CGI.
354
00:30:30.043 --> 00:30:31.003
It's just another tool.
355
00:30:31.023 --> 00:30:43.666
So if you're an amazing director, I don't think you necessarily have to start picking up the tools, but being able to work with AI artists and communicating the correct way to get the right results is a learning curve.
356
00:30:44.386 --> 00:30:49.007
It's why when you have a very VFX-heavy script, you search for VFX,
357
00:30:50.067 --> 00:30:53.008
directors that understand that pipeline.
358
00:30:53.188 --> 00:30:56.710
And like you said earlier, you call up a comedy director when you've got a comedy script.
359
00:30:56.750 --> 00:30:57.470
It's the same thing.
360
00:30:57.490 --> 00:31:06.295
I think we'll see more and more directors of every genre playing in the AI space, but it doesn't necessarily mean they need to get all the tools themselves.
361
00:31:06.315 --> 00:31:10.256
They just need to understand how to communicate what they're after.
362
00:31:10.657 --> 00:31:11.957
But what are you looking for then?
363
00:31:12.037 --> 00:31:12.738
I mean, what...
364
00:31:14.023 --> 00:31:22.069
You know, are you looking for traditional people who are picking this up so that they have that pacing and understanding of how to work in a client-oriented way?
365
00:31:22.089 --> 00:31:26.532
Or are you looking for talent that's coming right out of Central St. Martins?
366
00:31:26.572 --> 00:31:27.853
What are you looking for at the moment?
367
00:31:28.073 --> 00:31:32.896
I don't think I ever set out with Made by Humans to just make an AI company.
368
00:31:32.916 --> 00:31:37.400
I think it was always, let's make a creative company that uses those tools.
369
00:31:37.420 --> 00:31:40.002
So what we're always looking out for is still...
370
00:31:40.822 --> 00:31:43.023
is that a really good piece of creative?
371
00:31:43.463 --> 00:31:51.167
Obviously with Made By Humours, like hopefully it's a really good quality AI work as well without any artifacts or issues.
372
00:31:51.767 --> 00:31:53.768
But what are we looking for?
373
00:31:53.928 --> 00:32:04.412
If I'm putting a pitch out for people I'm trying to hire right now, we're looking for more female talent right now for sure.
374
00:32:04.993 --> 00:32:07.134
There seems to be, certainly on,
375
00:32:08.334 --> 00:32:15.577
from the research we've done, the majority of female AI artists are in the photography side of the aisle.
376
00:32:15.617 --> 00:32:16.037
That's interesting.
377
00:32:16.057 --> 00:32:17.358
And not the directing side.
378
00:32:17.678 --> 00:32:17.918
Yeah.
379
00:32:18.278 --> 00:32:21.960
And so we're always on the lookout for brilliant female talent.
380
00:32:22.500 --> 00:32:34.065
But I think young people that have been in, like you said, Central Saint Martins or, you know, Romansbourne or one of those creative colleges, that AI has been part of their curriculum
381
00:32:34.845 --> 00:32:55.600
since you know went into sixth form where it is um i would be really interested to see what those sorts of people are capable of you know the the ai native generation coming up now that's going to be fascinating what are the briefs that you're seeing that i mean are they more comedy oriented are they more graphic and visual are they just crazy scopes i mean are they all over the place
382
00:32:55.920 --> 00:32:57.301
They're so all over the place.
383
00:32:57.321 --> 00:33:03.446
And that's what I love about this industry, is one day I'm quoting to do a talking dog.
384
00:33:03.566 --> 00:33:08.330
The next it's a really serious pharmaceutical ad.
385
00:33:08.370 --> 00:33:16.597
The next we're trying to figure out the technical way of stitching all these shots together on a big virtual production stage in Poland.
386
00:33:17.118 --> 00:33:20.180
That's what excites me as a producer.
387
00:33:20.881 --> 00:33:22.923
Not that I get to produce an awful lot anymore,
388
00:33:23.704 --> 00:33:28.411
that's why I got into the business was to, I love puzzles.
389
00:33:28.471 --> 00:33:32.417
I love reading a script and going, how the hell are we going to do that?
390
00:33:32.557 --> 00:33:33.298
And cracking that.
391
00:33:33.318 --> 00:33:34.180
That's the fun for me.
392
00:33:34.777 --> 00:33:35.677
Yeah.
393
00:33:35.697 --> 00:33:43.281
You've mentioned a couple of times that it's a tool and you're not really focused on just over-indexing in AI necessarily.
394
00:33:43.381 --> 00:33:52.546
Is that an indication that over the next five years, we're going to see a normalizing to the middle and everything will have AI in it somehow?
395
00:33:53.266 --> 00:33:54.467
Everything will be...
396
00:33:55.287 --> 00:34:00.630
you know, some sort of hybrid or on some spectrum of live action to AI.
397
00:34:00.690 --> 00:34:08.074
And actually it'll probably skew towards AI the same way that when digital cameras came out, things started skewing digitally.
398
00:34:08.234 --> 00:34:08.634
Yeah.
399
00:34:08.974 --> 00:34:09.955
I don't, I don't know.
400
00:34:09.975 --> 00:34:16.098
I, from a, I really hope that live action continues to, to flourish.
401
00:34:16.118 --> 00:34:21.841
You know, I love being on sets and, but I do think AI, um,
402
00:34:23.301 --> 00:34:27.124
a big thing I was trying to campaign for early on is stop calling it AI.
403
00:34:27.304 --> 00:34:30.487
You know, we've actually been using AI for decades.
404
00:34:30.587 --> 00:34:36.071
We've just been calling them extensions and plugins in the VFX space, you know, or pushing updates.
405
00:34:36.091 --> 00:34:39.233
But AI has been prevalent in that space for a long time.
406
00:34:39.594 --> 00:34:40.314
It's just become...
407
00:34:41.395 --> 00:34:46.800
far more accessible and far more easy to use and clearly much, much smarter.
408
00:34:47.220 --> 00:34:49.562
But I do think it is still a tool.
409
00:34:49.763 --> 00:34:53.326
And it's just a choice that we can make in the production process.
410
00:34:54.106 --> 00:34:59.531
And, you know, there's many, many scripts we receive that clients go, hey, we want to do this in AI.
411
00:34:59.571 --> 00:35:01.673
And we go, you should not do this in AI.
412
00:35:01.753 --> 00:35:05.537
This is absolutely a shoot with a CG animal.
413
00:35:05.597 --> 00:35:06.818
You know, this is... Yeah.
414
00:35:07.258 --> 00:35:07.998
It's not AI.
415
00:35:08.819 --> 00:35:08.999
Yeah.
416
00:35:09.439 --> 00:35:18.582
There's many live action productions where we're sitting down going, hey, this shot, if we do this AI, we can move that budget to our department and so on.
417
00:35:19.062 --> 00:35:35.508
So I think that for me is more likely where this world is going is it's more hybrid thing where AI is a tool that creatives can lean on, but not necessarily something that drives the decision to make it in AI in the first place, just because it's there.
418
00:35:35.997 --> 00:35:43.263
You've spoken a lot about the production pressure that was in the economic environment prior to AI coming out.
419
00:35:43.443 --> 00:35:48.848
Everybody kind of acts like everything was rosy and then AI showed up and AI ruined everything.
420
00:35:49.068 --> 00:35:51.130
I know that that wasn't my experience.
421
00:35:51.150 --> 00:35:58.756
I mean, can you talk a little bit about that, about what the production environment kind of has felt like before AI came out?
422
00:35:58.936 --> 00:36:00.198
Yeah, you know what's funny?
423
00:36:00.218 --> 00:36:07.488
I wrote something down about this earlier and I was just trying to find it because I knew this would come up.
424
00:36:07.588 --> 00:36:10.712
You know, I don't think AI started the fire.
425
00:36:11.662 --> 00:36:18.184
I think we've been experiencing some decline in our industry for a while.
426
00:36:18.204 --> 00:36:22.665
We also went through COVID, which was a really tough time for our business.
427
00:36:23.225 --> 00:36:30.407
But we hear about CEOs making mass redundancies because of AI, but CEOs have been making mass redundancies forever.
428
00:36:30.647 --> 00:36:34.468
It is just part of the parcel of running big businesses.
429
00:36:36.029 --> 00:36:43.018
So I do think there is obviously AI is a little bit of a threat to filmmaking for certain people.
430
00:36:43.038 --> 00:36:49.085
You know, if you're a gaffer or a grip, AI is absolutely this threat.
431
00:36:49.125 --> 00:36:49.666
But I think...
432
00:36:50.807 --> 00:36:56.828
there is many times that technological advances have threatened creativity.
433
00:36:56.848 --> 00:37:05.090
You know, you mentioned earlier digital, you know, photography, uh, sorry, digital, uh, photography or video, you know, didn't kill film.
434
00:37:05.470 --> 00:37:06.130
It kind of did.
435
00:37:06.150 --> 00:37:08.931
And then it's come back, but photography didn't kill paintings.
436
00:37:08.951 --> 00:37:12.711
There's still many people that paint and film didn't kill theater.
437
00:37:12.731 --> 00:37:17.072
You know, theater is still absolutely something that is certainly someone that lives in New York.
438
00:37:17.112 --> 00:37:19.713
It's, you know, it's, uh, very busy on Broadway.
439
00:37:20.453 --> 00:37:20.573
Um,
440
00:37:21.113 --> 00:37:24.114
And, you know, CG didn't kill practical effects.
441
00:37:24.154 --> 00:37:35.737
In fact, someone told me this the other day, and I'm going to butcher the quote now, but Weta, the team behind... New Zealand, yeah, post-production company from Peter Jackson.
442
00:37:35.817 --> 00:37:37.838
They did amazing CG in Lord of the Rings.
443
00:37:38.738 --> 00:37:51.009
At the time, when they were doing the amazing crowd work of all the orcs in Lord of the Rings, no one was protesting outside their office because of all the extras that lost their jobs for playing orcs.
444
00:37:51.129 --> 00:37:54.012
It was an innovation that's amazing.
445
00:37:54.612 --> 00:37:59.016
And I think we're just seeing another innovation that we need to adapt to.
446
00:37:59.456 --> 00:38:04.981
One of the things that I was so encouraged by at AI on the Lot
447
00:38:05.802 --> 00:38:10.807
two weeks ago was we were on the volume stage speaking with AWS, um, Luma labs.
448
00:38:11.548 --> 00:38:11.708
Yeah.
449
00:38:12.329 --> 00:38:14.752
And they were talking about this production house of David.
450
00:38:15.873 --> 00:38:18.213
Again, done in two weeks, Ben Kingsley starring.
451
00:38:18.754 --> 00:38:27.095
And Ben Kingsley talked about it creatively, saying that, you know, when I was on the beach, I was seeing the ocean come in at me.
452
00:38:27.255 --> 00:38:29.595
I expected to feel my feet getting wet.
453
00:38:29.975 --> 00:38:40.397
You know, like that's how immersive it was for me, which, you know, anybody, Ewan McGregor, when he was talking about being on Star Wars, the prequels, he was like, I was so excited to be on Star Wars.
454
00:38:40.517 --> 00:38:44.238
And then I was acting across my tennis ball on green screen.
455
00:38:44.818 --> 00:38:48.819
And, you know, and that's a frustrating creative experience.
456
00:38:48.979 --> 00:38:59.322
And I think doing the math, we've got Warner Brothers and Paramount and all these studios, they have upwards of 40 sound stages each.
457
00:39:00.122 --> 00:39:02.222
Most of them are sitting empty.
458
00:39:02.822 --> 00:39:12.345
If you were doing a film in two weeks, in a two-week block, that means for 40 stages, you have 23 two-week blocks a year about.
459
00:39:12.365 --> 00:39:13.165
Yeah.
460
00:39:13.325 --> 00:39:16.927
That's over a thousand films, theoretically, a year.
461
00:39:18.047 --> 00:39:33.195
If we were to engage production on that level, and it would certainly take some years to get there, the sheer amount of jobs that would be required to staff those productions would be a gargantuan number.
462
00:39:34.014 --> 00:39:34.534
I agree.
463
00:39:34.574 --> 00:39:37.398
I'd love to know what feature films are being achieved in two weeks.
464
00:39:37.558 --> 00:39:39.240
Orlando, we're going to fix those numbers.
465
00:39:41.222 --> 00:39:44.066
Well, I mean, this is the thing to look up.
466
00:39:44.106 --> 00:39:48.070
This is John Owen's film, The House of David.
467
00:39:48.391 --> 00:39:49.092
I see what you're saying.
468
00:39:49.132 --> 00:39:51.134
If you were to do the virtual production approach...
469
00:39:51.274 --> 00:39:52.835
Like, they did virtual production.
470
00:39:53.115 --> 00:39:54.356
They did it in two weeks.
471
00:39:54.416 --> 00:39:57.218
They had Sir Ben Kingsley, an Oscar winner starring.
472
00:39:57.498 --> 00:39:58.139
They had O.T.
473
00:39:58.179 --> 00:40:00.981
Fagman, who's an up-and-coming star.
474
00:40:01.401 --> 00:40:05.904
They shot the whole thing in two weeks on the soundstage at MGM.
475
00:40:06.565 --> 00:40:11.368
I'm just saying that, like, the legacy production industry has a volume problem.
476
00:40:11.729 --> 00:40:15.151
Not the volume stage, but, like, a volume of output problem.
477
00:40:15.471 --> 00:40:16.312
There's, like...
478
00:40:17.599 --> 00:40:32.818
films lunch has been being eaten by social media and the ability for social media to crowdsource production on such a level that you can get so much newness that there's not really much point in watching a longer form production because I can just be swiping like this.
479
00:40:33.118 --> 00:40:43.367
If studios and legacy industries want to compete against the social mediafication of media, they're going to have to figure out a way to do it faster and cheaper.
480
00:40:43.507 --> 00:40:45.508
Empty stages are just sitting there.
481
00:40:45.809 --> 00:40:48.571
And like, you know, Universal is paying for those stages.
482
00:40:48.591 --> 00:40:53.575
They're paying the ground rent or the mortgage or the whatever the fuck else for these stages.
483
00:40:53.615 --> 00:40:56.697
They're sitting there, these facilities absolutely empty.
484
00:40:57.358 --> 00:40:58.540
Put a volume in there.
485
00:40:58.960 --> 00:41:00.823
See what you can make in two weeks.
486
00:41:01.243 --> 00:41:02.245
You know what I'm just saying?
487
00:41:02.705 --> 00:41:03.426
Start with...
488
00:41:03.446 --> 00:41:13.921
I don't disagree at all, but you'd think with all these volume stages, all these stages empty, and clearly lots of film, you know, live action film crew...
489
00:41:15.062 --> 00:41:41.028
not working as much as they would want that we would make a decision to reduce our costs to be able to make live action more feasible but actually what we're experiencing is rates increasing as a way to make themselves billable more when they're working less you know before i launched our ai company before i even launched you know imagine this and i was living back in london i think in
490
00:41:44.989 --> 00:41:47.492
because it became an unaffordable place to shoot.
491
00:41:47.512 --> 00:41:54.821
I mean, I now live on the border of New York and I've lived here for two years and I maybe shot here one time as well because it's the most expensive place in the world to shoot.
492
00:41:55.341 --> 00:41:59.326
But I don't think AI is the reason we're shooting this.
493
00:41:59.466 --> 00:42:03.110
I think the cost of shooting has become such a problem.
494
00:42:03.130 --> 00:42:04.091
It's economics.
495
00:42:04.191 --> 00:42:04.792
It's economics.
496
00:42:05.172 --> 00:42:12.775
I shoot more in Eastern Europe, South Africa and Mexico than I shoot anywhere else because I need to go to low cost centers to get my work done.
497
00:42:13.135 --> 00:42:16.177
And I think AI is a convenient scapegoat for that.
498
00:42:16.897 --> 00:42:21.899
But actually there's a much more bigger structural... Are they a scapegoat and potential savior?
499
00:42:22.299 --> 00:42:27.383
You know, I think that that's the thing that so many people are sticking their fingers in there isn't humming.
500
00:42:27.843 --> 00:42:37.391
And it's kind of like, all right, not every type of production is going to be poised or relevant or useful in a hybrid production environment.
501
00:42:37.731 --> 00:42:40.273
But I guarantee you, if you start digging into the background.
502
00:42:40.633 --> 00:42:43.977
millions upon millions of scripts that have not been made in Hollywood.
503
00:42:43.997 --> 00:42:52.906
I mean, Lord and Miller had, I read a script of theirs the other day that they wrote five years ago or six years ago and nobody produced it.
504
00:42:52.926 --> 00:42:55.388
And they're two of the best writer directors in town.
505
00:42:56.029 --> 00:43:00.033
You know, great people have good scripts that have still not been made.
506
00:43:00.334 --> 00:43:01.655
Let's go pick them up and make them.
507
00:43:02.216 --> 00:43:10.865
And I think that like, that's the thing that I'm at least encouraged by is I don't think that that necessarily replaces big budget films.
508
00:43:10.945 --> 00:43:13.087
I just think it creates a new class of film.
509
00:43:13.768 --> 00:43:17.091
that can get people working and get people moving.
510
00:43:17.111 --> 00:43:26.520
We were working with director James DeMonaco on some artwork character design for a new feature film that he's looking to put out there.
511
00:43:26.540 --> 00:43:30.224
James DeMonaco, famously the director and writer of The Purge movies.
512
00:43:31.065 --> 00:43:33.487
And we just got talking more after that.
513
00:43:33.507 --> 00:43:33.988
It was like, hey...
514
00:43:34.828 --> 00:44:01.194
you're a writer you must have thousands of scripts that never got made just not because you're a bad writer because yeah you like you said one percent of every script you know will actually ultimately get made and then we you know we started that conversation of like can we make it in ai can we make it for a fraction of the price can we just do it without studio funding can we just make this brilliant idea because of this new tool and we're exploring it and the answer is yes you absolutely
515
00:44:05.055 --> 00:44:17.300
money that it's still expensive i mean it's still you know like it's not free yeah you know absolutely but it does allow for far more opportunities for these scripts to be made
516
00:44:17.718 --> 00:44:20.320
And I think that that's the exciting part.
517
00:44:20.440 --> 00:44:25.605
AI has the potential, especially hybrid production, to make us both cheaper and also faster.
518
00:44:25.665 --> 00:44:28.767
And I think faster is another thing we're not talking about enough.
519
00:44:29.428 --> 00:44:36.814
Look, it's definitely faster, but I think the thing I keep coming back to and try to remind people of with AI is the scale and ambition that it locks.
520
00:44:36.854 --> 00:44:40.658
I think without AI, you have to, you know, assuming you don't have...
521
00:44:41.338 --> 00:44:58.775
unlimited pounds or dollars you have to compromise on ambition reduce the scope shrink the idea whatever you want to call it but with ai you can keep that ambition alive you can you can go for that scale that the script demands yeah and i think that's you know that's the thing that i keep trying to remind people of is
522
00:44:59.476 --> 00:45:26.772
you your client has 200k or whatever the number is but you can write an idea that's you know would normally cost 500k and that's not to say we're going to make it completely in ai we might go and do a one-day shoot i don't know i i really think the future is is hybrid production i really think that's where we're going yeah for sure i mean so what's your you know we're we need to wrap up i've kept you too long already but like let's say you know what's your hope for made by humans in the next kind of five years
523
00:45:27.252 --> 00:45:31.513
Well, my hopes for Made By Humans is that we've had such a brilliant first year.
524
00:45:31.593 --> 00:45:36.995
You know, we launched just shy of two years ago, and the first six months were absolute crickets.
525
00:45:37.015 --> 00:45:41.036
No one picked up the phone, no one called us, and we were certain we were going to be closing the business.
526
00:45:41.676 --> 00:45:45.077
And then slowly but surely, people, there are...
527
00:45:45.657 --> 00:45:47.358
their mindset towards AI changed.
528
00:45:47.938 --> 00:45:50.860
And this last 12 months have been absolutely amazing.
529
00:45:50.940 --> 00:45:58.663
For every live action shoot we've done it, I'll imagine this live action side, we've done five AI productions.
530
00:45:59.243 --> 00:46:02.245
So there is definitely a really big shift there.
531
00:46:02.265 --> 00:46:09.048
But I think my hope for Made By Humans is not just to keep making any old shit that comes through the door.
532
00:46:09.128 --> 00:46:12.670
Like I said earlier, I didn't set out to make an AI company.
533
00:46:12.710 --> 00:46:14.591
I set up to make a creative company
534
00:46:15.171 --> 00:46:16.712
that utilizes AI.
535
00:46:17.312 --> 00:46:18.993
So that is the ambition really.
536
00:46:19.073 --> 00:46:27.457
And kind of the mission for myself and Guy is to make the best possible creative work in the next 12 months.
537
00:46:27.477 --> 00:46:35.581
The last 12 months have been all about working out our pipelines and fine tuning those skills and finding the most amazing talent.
538
00:46:35.621 --> 00:46:41.523
And now it's about being more selective on the work, making it even better than the script that came in.
539
00:46:41.563 --> 00:46:42.184
And like really,
540
00:46:43.004 --> 00:46:43.685
adding value.
541
00:46:43.705 --> 00:46:46.689
I was to say, don't be more selective about the work.
542
00:46:46.769 --> 00:46:48.751
Just, you just put less of it on the website.
543
00:46:48.951 --> 00:46:52.436
That's not a bad idea.
544
00:46:54.323 --> 00:46:58.427
Obviously, with Made by Humans, that's one thing, but you've got this broader ecosystem.
545
00:46:58.527 --> 00:47:00.870
Tell me about what your hopes and goals are for that.
546
00:47:01.090 --> 00:47:01.811
You're absolutely right.
547
00:47:02.051 --> 00:47:15.264
You know, we, whilst Made by Humans is this really exciting new business that we launched, imagine this was the first business that we started, and it was, I was never setting out to build a network or multiple businesses.
548
00:47:15.324 --> 00:47:16.766
It just started that way.
549
00:47:18.047 --> 00:47:23.510
what has slowly become apparent is that there's lots of brilliant things that we want to do as a team.
550
00:47:23.550 --> 00:47:25.250
And the team grew from four people.
551
00:47:25.311 --> 00:47:30.753
We're now 10 partners around the world of staff and the rest of it.
552
00:47:30.853 --> 00:47:34.295
And we founded a match of this kind of on three main pillars.
553
00:47:34.315 --> 00:47:36.616
And that was, God, now I'm going to get my pillars wrong.
554
00:47:37.736 --> 00:47:38.897
Make production fun.
555
00:47:40.958 --> 00:47:42.719
Be multidisciplinary and innovate.
556
00:47:42.779 --> 00:47:47.161
And when we launched, we weren't able to innovate straight away because we were worried about innovation.
557
00:47:47.518 --> 00:47:53.799
the fallout of launching with an AI roster, but we did launch an amazing multidisciplinary roster which has served us really well.
558
00:47:54.299 --> 00:48:02.881
And making production fun again, like I told you I started my career at Framestore and I was doing crazy productions back then.
559
00:48:02.901 --> 00:48:08.482
Just flying on zero gravity flights in Texas, filming at the top of the andies.
560
00:48:08.622 --> 00:48:09.543
Every production was,
561
00:48:11.243 --> 00:48:35.037
an adventure and of course like it's a serious business and you know that was absolutely important we're being entrusted with millions of dollars of clients money but every single production we went on was just so much fun and so enjoyable i think that started to change a little while ago and and became a little too serious so something that we really wanted to inject into imagine this was this uh
562
00:48:36.518 --> 00:48:37.439
making it fun again.
563
00:48:37.459 --> 00:48:41.744
That's why we do these silly... Did I tell you before about our fries to survive thing?
564
00:48:42.425 --> 00:48:42.545
No.
565
00:48:42.565 --> 00:48:44.507
We go around the world ranking French fries.
566
00:48:44.547 --> 00:48:48.111
It's a totally ridiculous thing to do, but something that...
567
00:48:48.835 --> 00:48:51.776
you know, is rather fun for our clients on set.
568
00:48:52.076 --> 00:48:53.156
And- That's great.
569
00:48:53.856 --> 00:49:01.939
From there, the group is growing, the ecosystem is growing, and there's some very exciting announcements to come.
570
00:49:02.439 --> 00:49:02.779
Great.
571
00:49:03.099 --> 00:49:05.960
And I mean, you know, what does it feel like?
572
00:49:06.000 --> 00:49:11.742
Because it feels like Made by Humans is now bigger than, I mean, it's sort of eclipsed in size, imagine this.
573
00:49:11.882 --> 00:49:13.082
Is that a fair assessment?
574
00:49:13.862 --> 00:49:16.223
Maybe not on paper, you know, maybe not on the P&L.
575
00:49:17.363 --> 00:49:20.166
Uh, but certainly it's a more talked about thing.
576
00:49:20.186 --> 00:49:24.409
I think everyone wants to talk about AI, even if they want to talk about it in a negative way.
577
00:49:24.549 --> 00:49:33.737
So in every time I'm getting a call from LBB shots or campaign or whatever, it's all about always to talk about made by humans because it's, it's the emperor's new clothes, right?
578
00:49:33.757 --> 00:49:35.499
It's a thing.
579
00:49:36.339 --> 00:49:36.520
Um,
580
00:49:37.060 --> 00:49:38.401
But I hope it doesn't eclipse it.
581
00:49:38.441 --> 00:49:46.848
And as I said, as a live action producer, I really hope that I spend as much time on set as I do in this very office making AI work with the team.
582
00:49:46.948 --> 00:49:47.168
Yeah.
583
00:49:47.429 --> 00:49:48.990
Well, listen, good luck to you, Toby.
584
00:49:49.030 --> 00:49:51.672
I mean, I'm so glad it's going so well.
585
00:49:51.752 --> 00:49:54.515
I'm so glad that, you know, that the...
586
00:49:55.335 --> 00:50:07.147
uh, the, the veil of iniquity has lowered on AI a little bit enough that, um, that it can be both lucrative and fun and, and, you know, being able to grow and, and also educate people about what's possible.
587
00:50:07.227 --> 00:50:11.151
I think, um, I think you're doing really good work and, and I just hope it continues.
588
00:50:11.611 --> 00:50:12.673
I really think that's it.
589
00:50:12.733 --> 00:50:16.476
Orlando is I would have regretted not trying, um,
590
00:50:17.237 --> 00:50:21.159
You know, I had an idea of building an AI creative studio.
591
00:50:21.740 --> 00:50:24.902
It could have absolutely failed, but I would regret it not trying.
592
00:50:24.942 --> 00:50:28.224
And I'm really happy that I have explored and it's been very fun and exciting.
593
00:50:28.284 --> 00:50:31.626
And yeah, we had so much fun along the way.
594
00:50:31.686 --> 00:50:36.209
And hopefully we actually do have five more years of things ahead of us.
595
00:50:36.229 --> 00:50:36.449
Yeah.
596
00:50:36.609 --> 00:50:37.590
Well, great luck.
597
00:50:37.650 --> 00:50:38.550
I'm sure you will.
598
00:50:38.790 --> 00:50:39.931
And congratulations.
599
00:50:40.031 --> 00:50:42.873
And yeah, just keep us posted on whatever you have coming next.
600
00:50:43.373 --> 00:50:43.733
Brilliant.
601
00:50:43.793 --> 00:50:44.194
Sounds good.
602
00:50:44.214 --> 00:50:44.774
Thanks, Orlando.
603
00:50:45.154 --> 00:50:45.535
See you soon.
604
00:50:45.595 --> 00:50:45.975
Pleasure.
605
00:50:47.521 --> 00:50:48.101
Well, that was great.
606
00:50:48.201 --> 00:50:57.083
Other than having a really dreamy voice, I think that the other thing I really like about Toby is that he just has a real love for production.
607
00:50:57.103 --> 00:51:00.144
He has a real love for being on set, being on shoots.
608
00:51:00.364 --> 00:51:01.345
It's the one thing I never had.
609
00:51:01.525 --> 00:51:03.465
I love the notion of production.
610
00:51:04.165 --> 00:51:08.026
But to be honest, when everything's going well on a production, it can be frightfully boring.
611
00:51:08.866 --> 00:51:12.927
I always said that the more interesting something was to bid, the less interesting it was to shoot.
612
00:51:13.688 --> 00:51:15.028
A car chase is brilliant.
613
00:51:15.208 --> 00:51:16.529
bone crushingly boring.
614
00:51:17.069 --> 00:51:39.322
Um, but I think that his love for production, I think is what makes, uh, it's such an easy sell to hire him and his company for, um, AI production because it's somebody who is not just running into it because he has no other choice or didn't get the chance to produce at a high level, uh, prior to AI coming out, which is a lot of people.
615
00:51:39.342 --> 00:51:40.002
Um,
616
00:51:40.722 --> 00:51:46.544
He has a deep love for and enjoyment of the process of production.
617
00:51:47.344 --> 00:51:51.445
And if given the choice, I think he'd shoot live action.
618
00:51:51.505 --> 00:52:00.528
But the opportunity to create hybrid work, and I think that we talked a little bit about how interesting hybrid work is.
619
00:52:01.735 --> 00:52:03.656
you know, on volume stages and things like that.
620
00:52:04.076 --> 00:52:05.516
Things are about to get quite exciting.
621
00:52:05.536 --> 00:52:08.597
And I think that Toby's very much on the front lines of that.
622
00:52:09.478 --> 00:52:15.019
And so, listen, I applaud him for, you know, straddling both sides of that line.
623
00:52:15.039 --> 00:52:22.222
I think that's very much what we try to do here in our conversations on this podcast is have a real respect for the industries that have come
624
00:52:23.322 --> 00:52:36.253
This far and making sure that anybody who makes stuff with AI does it with respect for the wonderful work of the industries prior to AI coming out.
625
00:52:36.473 --> 00:52:40.637
And so listen, I think that that torch is well placed in Toby's hands.
626
00:52:40.977 --> 00:52:41.978
So I think he does a great job.
627
00:52:42.558 --> 00:52:45.500
Um, next week, we've got Keith Kupfer Schmidt.
628
00:52:46.100 --> 00:52:52.324
Uh, Keith is the head of the Copyright Alliance, and, uh, it's one of our, um, Future of Life episodes.
629
00:52:52.524 --> 00:52:54.705
We'll be digging into the depths of copyright.
630
00:52:54.925 --> 00:52:57.787
We've dug in a little bit, but I don't think we've done a deep dive.
631
00:52:58.167 --> 00:53:00.168
Um, and, you know, what we're trying to do...
632
00:53:00.488 --> 00:53:23.752
with some of these Future of Life episodes is take some complex ideas surrounding copyright provenance and some of the copyright concerns kicking around about AI and really dig into them and really subdivide them and make sure that we're giving you a good framework with which to understand the moment we find ourselves in, especially when it comes to copyright specifically.
633
00:53:24.293 --> 00:53:26.596
Um, but anyway, it's a great episode.
634
00:53:26.796 --> 00:53:28.559
He's actually a treat to chat to.
635
00:53:28.659 --> 00:53:30.522
It's, it's one of the interviews I've enjoyed the most.
636
00:53:30.562 --> 00:53:31.404
He's, he's great.
637
00:53:31.824 --> 00:53:33.147
So please tune in for that.
638
00:53:33.607 --> 00:53:34.629
Uh, stick around next week.
639
00:53:34.789 --> 00:53:36.892
Thank you so much for listening and we'll see you next week.
640
00:53:37.073 --> 00:53:37.213
Bye.