How the Creative Industry is Adopting AI with Miora, Tencent’s Creative Agent

This episode was a special collaboration project with Tencent’s Behind the Penguin podcast. AI Proem remains editorially independent but is open to collaborating with companies and individuals if that means they can bring better resources/ guests/ insights to the AI Proem audience.

I speak to Serrino Liu, the AI Product Expert at Miora, Tencent. He is responsible for product innovation across film, 3D, and the interactive formats built on top of them, bringing AI agents into real content-creation and workflows.

He tells me that Miora is not another image or video generator. Rather, it is a creative agent studio that runs entire workflows for filmmaking, e-commerce, game and 3D professionals.

What makes the conversation unusual is Miora’s stance on models: it deliberately routes across mainstream models from different providers: DeepSeek for documentation, Seedance for video, Hunyuan Image for graphics, and models from direct competitors like ByteDance and Kling. Serrino is refreshingly candid about why: Miora is not a foundation-model company, and every improvement in the underlying models flows back into the product. He also unpacks Miora’s relationship with Tencent’s office agent WorkBuddy (”two pieces on the same board” where one manages the workspace, the other the creative work), and how the newly launched Hy Image3.5 preview is “professional-grade” in the sense that matters: usable, consistent, and able to survive revision, not just prettier.

The heart of the episode is a nuanced take on taste and judgment. Serrino splits creative work into what can be verified, such as frame format, dialogue, and character consistency — and what only a human can judge: whether the emotion lands, whether the camera movement serves the story. Features like the new StoryWeave and side-by-side version comparison are built to keep the human in the loop, “a partner in making, not replacing.” The same philosophy extends to memory: Miora’s visible, editable memory system exists to kill tedious repetition — remembering your color palette, your story decisions, your style — so the next revision respects everything you’ve already decided, instead of forcing you to re-explain yourself.

The episode closes with two big-picture exchanges. On geographies of preference in AI adoption, Serrino argues the difference between Hollywood and Asia is less about values than stakes: Hollywood’s AI debate arrived as a contract question (consent, credit, compensation), while Asia’s arrived as a capacity question — enormous demand meeting thin mid-tier production, with both anxieties eventually converging. And in his contrarian take on a fully automatic creative AGI, he lands the episode’s most quotable line: in creative work, the most expensive thing was never execution — it’s knowing who to pass the ball to. Craft will always change, but the author and the user never do.

The AI Proem Podcast is part of the AI Proem newsletter, which has ~13k followers globally. To learn more about China AI, the business of AI, and how AI is impacting society, please check out the newsletter here and more insightful conversations here.

Chapters

00:00 Welcome from Macau: Meet Serrino Liu

03:35 What Is Miora? Agents, Memory & the Professional Harness

04:53 One Product, Many Models: Routing, Rivals & Token Economics

11:14 Building for Real Creators: The StoryWeave Approach

18:12 Taste, Judgment & Keeping the Human in the Loop

28:09 Memory as a Creative Asset & the Hy Image3.5 preview Launch

37:20 Asia vs Hollywood: Two Different AI Stories

44:23 Misconceptions, Copyright & a Contrarian View on AGI

AI generated transcript, for reference only

Grace Shao: hello everyone welcome back to AI Proem Podcast this is your host Grace Shao. Today’s episode is very special because we’re partning up with Behind the Penguin a newly launched Tencent backed podcast. We’re in Macau today because we are at the 2026 International Film Camp and joining us today is Serrino Liu, Tencent’s Miora AI product expert. He had just revealed and demonstrated Miora’s capabilities in film making to industry really excited to hear from him today about how the creative industry is reacting to AI, adopting AI and how he sees the future ai I’m really excited to hear from him not only just about the product itself but how the industry is embracing AI or maybe not.

Serrino Liu: Okay, thank you so much for introducing me and it’s great opportunity to meet with all kinds of the different directors today and I think I actually have already got some questions I wanted. I think this conversation about making film with AI is actually welcoming a different a gate for the young director to try AI with some film production and I should say i’m only speaking for the room and not industry. But in the room nobody was impressed with the film itself, though like the output anymore. It’s actually how they can steer it.

Grace Shao: It’s very cool. But before we get more into Miora and the Tencent’s ecosystem and just how AI is being adopted. I just found your personal background very interesting actually before we start recording we just had a brief chat and you’re actually a French/Russian/ Chinese descent, but now you just returned from the US to China to work for Tencent. You have a very interesting professional journey as well. You’re an artist by training, having studied design at Parsons, studied CS at NYU, studied at Harvard. Tell us a bit about yourself and what brought you to Miora and the vision that you have for Miora.

Serrino Liu: okay thank you. Though “artist” is a hat I‘m not sure I’ve earned — I‘m really a generalist. From computer science, design and business and machine learning and AI... I have touched all of it. But it sounds pretty scattered but I think there is a clear line behind it. It’s like I know the tech I know the product and the whole time i’ve been obsessed with is just one thing: How much idea in someone’s head can get real and that is actually the reason I joined Miora. I want to see how creative agent can provide some opportunities for customers for users to get some inspiration from their head to the real life and to see their inspiration come true.

Grace Shao: If you had to use one or two sentences to describe Miora or what it is today. How would you describe it?

Serrino Liu: I think it’s just agents and memory and the professional harness to serve as the assistant directing different area.

Grace Shao: so it’s not really a generic AI image or video generator in any sense.

Serrino Liu: No, it’s not. It’s more like a creative agent studio or like a system for you in personal so it can help you to do some professional work and jobs etc.

Grace Shao: Very cool. So what industries do you think Miora can be used in or is being used in right now?

Serrino Liu: I think we have some targeting area like the e-commerce, filmmaking, game and even 3D... Those different areas have some different professional harnesses with the agent to do the specific workflow to serve this kind of area.

Grace Shao: but I imagine for each of those use cases just mentioned or each of those industries just mentioned, their need for how to even use a creative tools actually very different right, so in that sense how are you guys optimizing which models you’re using because I understand that actually miora routes through quite a few different kinds of models from different providers and it’s not like solely committed to just offering Hunyuan.

Serrino Liu: Yeah, so Miora isn’t just tied to a single model we’re actually routing different mainstream models like the Seedance and the different agent model as well. So user can pick what fits their needs best and in particular Miora is actually one of the way (for) the hunyuan model to reach the overseas uses. We have the newly launched Hy Image3.5 preview releasing today and that is actually the agents own choices to inside the Miora and users can actually choose it.

Grace Shao: but so say in my workflow, I need to script some data I need to talk to them convey my visual vision, I need them to relay like an actual image and I need to help me finish a complete workflow. How does the agent then decide which model to use in that process?

Serrino Liu: so that’s harness work so we have some auto routes for choosing different models so for like the documentation work we can choose like DeepSeek, because it’s learning through the literature models. The Seedance model is more like the video generation model. If you have some work like graphic and some like the e-commerce operation, like the images you can choose Hy Image or Seedream; it’s more like the image generation models. If you have some style to generate, very synthetic pictures you want to generate so you can choose Mid-journey. We actually have this all under the cover, so user cannot sense. We just have you control like the routes in different models.

Grace Shao: so it’s a very smooth experience for user itself.

Serrino Liu: for user you can just switch from the different models and you can change your best value model combination like the agent you can select like the DeepSeek and for the picture you can choose like the image 2 and for the video generation you can choose Seedance rapidly and for the best value and the cost.

Grace Shao: so would you say most of your user are prosumers, consumers, enterprises...

Serrino Liu: I think it’s pretty general like we have all kinds of users from different areas like the filmmaking like the campaign marketing and even the UI designers yeah they can all use Miora.

Grace Shao: One thing that you know caught my ear just now while I was talking to you is that you guys actually route through quite a few different models and what’s interesting enough is some of them are frankly from your competitors right? um you know you name Seedance which is clearly one of the strongest video generation models coming out of Bytedance, also a chinese player as well. How does that play into your thinking and building up Miora, like it’s not very protectionist in that sense right? How does it work?

Serrino Liu: Miora is not a foundation model output; We just use these kinds of models in our product to serve as a servant to the users and to best serve their use. So we actually like to see improvements in these kind of models because it can back to our product with this kind of advantage.

Grace Shao: and I think that’s’ another thing that was really interesting is when I was doing research on you guys is that the entry point for non china is through users of Miora they can access Miora through Workbuddy. That’s quite fascinating because Workbuddy is obviously now the darling of Tencent, it’s probably the leading office-use agent in China. What is your relationship with the team how do you feel about building within such a big ecosystem Tencent as well so two parts of the question.

Serrino Liu: So in their public positioning, they are actually two pieces on the same board. Miora is actually building on Workbuddy’s agents capabilities: it can understand the context and remember the user’s taste, style, and assets. Workbuddy actually handles the workspace jobs; Miora is actually working with creative works. Just same direction: get AI into the real work, but very different jobs.

Grace Shao: but what is it like building Miora within Tencent, just given how grand you know the ecosystem is. On one hand you have a lot of resources obviously you have a lot of talent, on the other hand I’m sure there’s also internal competition also you know internal KPIs that might not fully align with each of the different product managers right?

Serrino Liu: So let me just say this, Miora obviously is a standalone product (for overseas users). You mentioned the ecosystem, the ecosystem gives you the resources and conditions. A bit like I can sit here talking to you right now. It puts me in front of lots of different customers and users, so I can hear clearly and directly from them where the product actually needs to go. But conditions are conditions. The product still has to start from the specific person’s task, and an ecosystem cannot decide that for you and it just helps you listen very closely to the real target market and users.

Grace Shao: let’s shift and look at the product itself right So I find it quite interesting you are one of the key people building up your right now but you come from both AI and design which is very unique background. What did you see missing in the first generation of AI creative tools that made you now want to build on the product layer.

Serrino Liu: You mentioned the first generation. It’s like a decade ago but it’s actually just three years ago right? When DALL·E first launched, the first wave was actually very striking, very shocking. What actually keeps a creator coming back to everything after the generation is actually the results. The results are very stunning and amazing but then the creator ask, “Oh, just keep this character and change the action or like the conversation is right but the emotion is wrong... “ So the problem is never the generation; it’s actually the revisions that follow. So the production layer has to let people to say clearly to mention that strategy about revision adjustment, what to keep and what to miss. And so generation actually gets into everyday work, instead of sitting in a demo like a toy.

Grace Shao: so, another thing that was talked about was that when OpenAI closed off Sora, obviously people were talking about how expensive multimodality is. It’s very token hungry. Does that play a role in anything in consideration for you at all? How hungry for tokens multimodality can be like working with images and videos.

Serrino Liu: I think that’s a wonderful question for users and the young directors, as they have mentioned as well. So the token is actually a different marketing strategy for the Miora to serve as the product to the users so we have some best value combination for user to use. We have some strategies to help people to save costs and to have different strategies to have the token saved.

Grace Shao: Walk me through a potential real product say I want to make a mini drama right now or I want to use AI to generate an interview between us. What does it look like on Miora?

Serrino Liu: So we have a new launched feature called Miora StoryWeave. So we just decide how what to ask user first? So it’s never like though you have an idea that we generates a whole different whole short drama for you. We ask what this short drama is about, what you have to keep, what are the materials you have? so a person with the full script is different from the someone who has just an idea, just a shape, so this new feature is actually walking through the whole different workflow. If you have the script, you have an idea, so it generates a screenplay for you. We extract the characters and scenes and some assets for you, then a combination of shots, sounds, music. Then we combine all these kinds of elements and into one system and help you to generate and produce a film for you.

Grace Shao: Just a partner in the making.

Serrino Liu: It’s a partner in making, not replacing.

Grace Shao: I don’t work in a creative field but what I hear a lot from designers actually is that they use software like Figma because of the shareability of it because they’re shared context shared environment. much of right now when we work with agents such as like Claude and Claude Design, there’s less of that kind of shareability. I have my memory context, you have yours. I have my environment I built it you have yours How are you addressing this or is this something that you guys are even considering?

Serrino Liu: So I think creator or designers are not the blank space right? They have tools, ways of the working they have been familiar with for over years. And honestly they‘re actually attached to this kind of tools..So we never say “what can Miora replace.” We ask “where exactly can Miora help” — developing a direction, or turning a piece of feedback straight into the next version... And I never think someone should move their whole workflow just to try AI. That’s just a misunderstanding of AI.. I judge how well we fit by three questions: can they keep using the assets? Can they keep editing the resul? And is it clear when they hand it to a colleague? Pass those three, and you’ve actually fit in. That’s the questions Miora have to solve to help this kind of creators to move the workflow into the Miora and we have a memory system for that.

Grace Shao: so I think one interesting question I have for you is that you know there’s been some push back from the creative industry saying you know if we’re working with AI, what about the creative taste and creative judgment right? So my question for you is how much of that taste decision-making is still in the hands of people, in the hands of people, like yourself, who actually spent years studying design the best practices right versus in the hands of now a machine?

Serrino Liu: so i’ll actually still use the film making as an example:. I split “what we can check” from “what the person has to judge.” Frame format, dialogue, an agreed character reference — those can be verified fairly clearly. But whether the emotion lands or the camera movements, pacing grips, that only makes sense against the intention of the piece. What a product can do is make comparison easier— or happen earlier. So we have two versions of the clips side by side, show exactly what changed, so feedback goes from “”it feels off” to “right there is wrong.”

Grace Shao: What does that mean?

Serrino Liu: so we have two versions of the generated videos right side by side and this one is original this one is AI enhanced so you can see the original one you can see I just want to change that emotion I just want to change the facial expression right there and you have the AI enhanced one and you said right there I just have that generation back.

Grace Shao: so it’s really just like a tool to tweet...

Serrino Liu: just tweaks the emotion... or just tweaks the expression. I think the you mention the taste and the subjective. But we don’t have an aesthetic score for that, we don’t have an emotional score for that— instead, We just have some adjustment from the people that use it to adjust to sense that kind of the moment that kind of expression from the original form from the video you want to change and we just have the clip script to nailed it.

Grace Shao: so I think this example is quite interesting because it’s really about so keeping the human in the loop really as a core process part of the process but what versus like what about all these AI short mini dramas that’s being turned out frankly they’re not sophisticated it’s not like a lot of the motions are you know that detailed uh maybe some of the actors someone may argue are not as skilled you know as the big screen actors so thus are is that completely just giving away like the judgment and the direction to the machine to decide then? it said you feed them a script they turn out a video is that understanding correct?

Serrino Liu: I think that’s different directions. Like the short drama you mentioned about is targeting a different area. We just have the momentary excitements from watching it so the production or idea is very different from making a film. So for film, you have to have to address some very minor very different micro expressions and very narrow emotions. In terms of the short drama, it needs very dramatic twists, very exciting moments to watch and I think that’s not about the taste it’s about the marketing strategy here.

Grace Shao: so Serrino I think that makes sense so essentially what you’re saying is for different uses use cases how much you use AI how much you delegate to them is different right and how much you want to keep human and loop is different that’s all fair. But my question really is how do we understand this philosophically. So for creative writers artists so much of what their value added is or even frankly what they’re getting paid for is their creativity, now when you add AI in it. is it so fair to claim that as your own? and I will use a personal example and I always disclose to my readers as well on AI Proem newsletter, I sell insights essentially I do a lot of deep dive analysis spent hours and hours thinking about business strategies and business models. I don’t see myself as a literally writer. I’m not sure in fact I don’t call myself a brilliant writer at all in fact I thought I was one of the worst writers at journalism school but I think you know I don’t get paid essentially for my writing no one’s coming here to say I was touched by her beautiful writing so I also am not ashamed to say I used ai but I think many people will feel quite differently about this right some people will say if you are a writer then therefore you have to be humanly writing everything um which again it really depends on what you’re doing so when I write for traditional publications such Bloomberg, Fortune, I actually do my writing in the traditional way. I may use AI for research and assistant, now I throw this back to you, how do you view the usage of AI in the creative field in general

Serrino Liu: I actually have a similar talking well discussion in this talk today and what this about is actually how much creative work involved in AI and it is valuable. For me it’s actually two different things: so while using AI is actually to expand your knowledge because a person’s knowledge is very limited but AI can help you to know different things and to expand your knowledge basis and to open your eyes, so that’s one thing and AI can actually enhance your capability of using different tools and to help you to level up your ability, the quality of your work so that’s different things but in the value itself actually You can never judge a person to use AI to enhance the script, to have a scene generated and to have the similar real life environment setting, so I think it’s very hard to judge how people use AI right now.

Grace Shao: I think one thing you mentioned earlier is that you guys create these characters and they’re actually part of the memory and I want to bring the conversation to memory something you’ve talked about time after time within our conversation. is actually that Miora has a very strong memory now what does that really mean in terms of actually for users when they’re using Miora and how is memory really useful when we’re thinking about creativity or in a way does it not actually restrict you because your agent has such a strong memory pull me back to what you previous thought?

Serrino Liu: So memory system exists because the creative work is actually continuou. so the next revision has to respect why you‘ve already decided, or the creator ends up re-explaining the same things over and over. Memory exists to kill that most tedious kind of repetition. I split it into two things: the decisions of this project — which character, which story version; and preferences that might last — like a brand’s visual conventions. And beneath that, there’s project memory versus personal memory. project memory is actually the project you have been working on like the taste and the campaign and the graphic, your character...you don’t have to repeat I have to use this character as my starring role. so, memory has ultimately remembers that. and for the personal memory is more like your taste your style. Like your color palette your choice of colors like the bag a blue or red the agents will just automatically remember that you don’t have to repeat I have to use this color again and you have just say like uh what is my favorite color and just use that in my video generation, in my image generation.

Grace Shao: Is that not something all the other models can already do? Is it something so unique to you guys?

Serrino Liu: yeah, it is very unique. we just make memory system visible for user to add it to delete or to adjust and so people can use memory as an asset or a database for them to use again.

Grace Shao: help me understand technicalities is this memory system within your orchestration layer then?

Serrino Liu: it’s not orchestration; it just automatically remembers everything you produced in the canvas.

Grace Shao: I see okay. So you know we alluded to the fact that like this space is quite crowded there’s a lot of players beyond even just the raw models of multimodality models we’re talking about even just multimodality agents or the creative studios per se you know there’s Higgsfield, Tapnow, Seedream from Bytedance, Pippet AI under Capcut. then I think yesterday Byteplus just pushed out something new as well. there’s just so many players like I can keep going. How do you view your competitors or even just a general landscape right now?

Serrino Liu: I think from the perspective of the user, creators will compare any tool that helps them finish the work, so the competition over those use cases is real, and I won’t judge it. Miora’s advantages are in three things, agent, memory and professional harness. For us as an agent product, what we have to work on is: how a creator sets the direction, compares results, and turns feedback into the next version. Like Higgsfield, the new launched “ByMagic”, they are actually not a agent product they are workflow product, a tool for people to use. Miora is actually an agent product that will actually help users to adjust to control their needs.

Grace Shao: I see so look let’s talk about what’s underneath at all. You mentioned Tencent Huyuan is releasing a new professional grade image model. Tell us about Hy Image3.5 preview. Why is Miora the main entry point for many users and what is decopeted edge for that being embedded into Miora?

Serrino Liu: Hy Image3.5 is “professional-grade” here means usable, not just prettier — it holds a spec and survives revision. That’s what creators actually need. And for creators outside China, Miora is where you can get your hands on it. The model raises the ceiling; Miora is how you reach it. You said token is expensive. 3.5 is much cheaper and much usable than like the other different image generation models.

Grace Shao: for yourself as a creative i’m sure you use all these different models tell us which models good for what like if I had to name a few, like you know this is not representative Tencent, it’s representative of Serrino Liu’s own usage so let’s say Minimax they’re quite famous for the multimodality models and then let’s say Seedance let’s look at Hunyuan the various other ones on the market how would you actually rank them for what usage?

Serrino Liu: My personal preference is actually for cinematic pictures to generate I would choose Midjourney because it has the best taste, like the composition of the picture and the lighting shattering everything just right on the time. If I have to make these images consistent like the character and like the environment I would choose Image2 because it can hold your consistency, and for video generation models I would choose Seedance2.5, because it has best control of the prompt you wrote. a full triple choice you can choose the minimax h3 definitely, it has some best motion effects so it’s best for the e-commerce users, customers and I think that’s it. and so my combination is actually I will use general agent model I actually use DeepSeek pro and I Seedance2.5 for video, for the picture I actually use Image2.

Grace Shao: so how did each of these labs then develop their own uniqueness like to your point on Midjourney right why is it that they have better taste why is that they’re better with the composition and everything?

Serrino Liu: I think it comes from the training data used. They have different direction more specific or the style and the community they have is actually help them retraining their data so each different models have the focus. so Midjourney is more focused on the taste, authentical pictures and the images.

Grace Shao: but they’re very expensive

Serrino Liu: Not really, Image2 is actually more expensive than Midjourney. Minimax is actually very cheap, much cheaper than Seedance 2.5, which gives you more real, more cinematical, more consistent content.

Grace Shao: so let me take a step back you actually just moved back to Asia five months ago and I think right now the kind of discussions around AI adoption is quite different in Asia versus the US or it’s China versus the US, and you know Hollywood has a stands in AI right now but it seems like whether it’s in Hong Kong film festival or today, where Macau the industries are much more let’s say normalizing embracing optimistic about AI. Would you say that’s a fair statement or do you think the industry concerns are quite similar?

Serrino Liu: The difference is less about values, more about what’s at stake right now. In Hollywood, AI arrived as a contract question — consent, credit, compensation. That’s rational: there’s a mature, well-organised craft economy to protect. In much of Asia, it arrived as a capacity question — huge demand, thinner mid-tier production. The anxiety is framed differently — the craft need is identical. And they’ll converge: Asia will hit the rights question as it matures, Hollywood will hit the access question as budgets compress. Asia’ isn’t one film philosophy — Japan, Korea, China and Singapore differ hugely.

Grace Shao: that’s very interesting so you’re seeing it’s not really a philosophical debate rather jua pragmatic decision.

Serrino Liu: I think so.

Grace Shao: look I love to kind of touch on just filmmaking a little bit more um so actually I spoke to Kuaishou before - Kling were partnering up with a lot of film festivals held in like can Hong Kong etcetera. How do you see the space playing out do you think we’re gonna have many different players like the Seedance of the world the Kling of the world and you guys or do you think you know one day this will be more consolidated. different studios even right now will have different taste and different it will serve different kind of clientele

Serrino Liu: I think that’s right, but let me just clarify so Kling and Seedance are the models. so Miora is like the different approach than them. so the Kling and Seedance is more focused on the generation itself but Miora is actually helping the creator to do their own work and to more like a different professional approach. Miora is actually serves as an assistant. we have like e-commerce users, we have the flim making users and we have some games users in Miora. Kling and Seedance is more like the beta generation but Miora is actually helping different users to take their demands and into the real production workflow while Kling and Seedance is more focused on general generation.

Grace Shao: I see you know one thing I heard from the Kling team was really interesting and kind of at close to what you’re saying that AI right now enables or empowers a lot of the different players in the industry say for example you’re a actor you also had always had a vision to direct your own film you don’t actually have frankly the know-how to do so or the resources do so. So now you can work with an AI agent to direct your film, because you could tell them you know how this whole works from actors perspective or you’re a director but you don’t have money to hire actors and you have a whole vision of how a film can play out so you may use AI generated characters to help you build your vision. How do you see that trend?

Serrino Liu: I actually think is connect with the talk today i’m with the differential directors they have the vision they have the inspiration they have the idea but they don’t have the budgets right but AI actually can solve the kind of problems to help them get that kind of ideas from their head and into the AI filmmaking production process to help them actually to ah determine which direction is a good fit for them to move forward and I think AI actually can help different people to to think differently and to work differently and in a finite budget situation.

Grace Shao: no that’s interesting and I think but my question I guess is do we need certain capital to technically maintain certain quality like is that so necessary or do you think what is scarce and valuable is really just a human that has a judgment and with AI they can just produce it or do we actually still need that resource for the real production like a Christopher Nolan project is not something AI can generate. How do you view that quality?

Serrino Liu: it can never do like a hundred percent perfect film from AI. it cost a lot but well, what AI can do is actually mak a eighty point production for you and we can lift it to a 90 and that’s it. That’s a different approach like the human way as the watcher to see a perfect view versus some AI generated view. and that’s a different style, different aesthetical approach.

Grace Shao: Serrino if execution gets much cheaper and more abundant then what becomes more valuable for the filmmakers?

Serrino Liu: The judgement of what’s worth making, and why it should be made this way. When you can generate more images, the director’s responsibility to choose doesn’t disappear — it gets heavier. Is this shot necessary? Is the performance right? Does the film have its own point of view? And one more thing goes up in value: experience, and the feel of real life. A tool can help you try an expression, but your understanding of a person, a relationship, a place — that decides what you choose, and which detail you’re willing to fight for. That’s director’s choices. AI cannot give you that, and it can never do it for you.

Grace Shao: like the depth and the understanding and going through the emotions in life. I i love that beautiful answer however I feel like i’m repeating myself in my episodes at this point but what about the juniors how do they gain that I mean you know not saying people experience less when they’re younger but in general let’s say you know I feel like i’m a different person now in my thirties i’ve become a mother moved around the world how I see the world is very different how I saw the world when I was in my twenties I believe when i’m in my forties it will feel different as well um but if you say so much of that what makes you unique and valuable actually requires time and experience can you say it’s a bit unfair to juniors or do you think frankly that was always the case

Serrino Liu: no that’s not always the case actually juniors gets the whole different scenarios than that juniors have the inspiration that imagination rights like the child have the big different colorful imagination than the adults adults have limited the imagination and actually AI can help juniors a child to expand the inspiration and their imagination, then adults have a shelled in their imagination of the real life but the fixated world yeah sometimes yes and but the children or the juniors they have their own imagination they have the colorful world they have they have this opportunity to expand in bloom to release this kind of imagination in from the AI production like they can generate the in video images and just combine them and to see the actual world. that’s beautiful

Grace Shao: That’s beautiful. no I love that because I think you’re one of the few people that actually address this like you know you’ve really addressed this concern because some people like people are always concerned that what’s gonna happen juniors of AI is gonna take their jobs and execution becomes cheaper but really you’re saying young people have young people’s advantage seniors of seniors advantage and AI really end the day in your view is a partner and a tool to kind of progress and bring out their visions to life

Serrino Liu: I think the crafting can change always change but the author and the user they didn’t. we just used the kind of tools to expand our imagination our creativities to help us in a different scenario.

Grace Shao: and to even create things that used to be impossible right? When CGA came about we never even thought we can create these things

Serrino Liu: that’s right before yeah thank you so much thank you

Grace Shao: what is something you think people are still maybe misunderstanding or underappreciating the AI revolution right now, especially when it comes to a creative industry?

Serrino Liu: I think a lot of people misunderstand commercial videos. like when new features coming out they think oh why you can do that and why you cannot make that and they actually forgot what is underneath that is a lot of reviews, a lot of cards, and a lot of efforts behind. I think people are too easy to accept that AI can do or cannot do and they always worship AI’s advantage not so the effort or that the tedious work behind it.

Grace Shao: Then, actually on that I have to pick this pickle which is then how do you view the copyright issue?

Serrino Liu: I think some actor already sold their faces to the some modeling companies right I think it’s like a trade-off situation. I think I think that’s different from copyright in creative works, such as paintings. Here, we’re talking about someone’s face or voice. but in the future that is definitely a difficult problem we have to face.

Grace Shao: one last question I have for you which is a question I ask everyone that comes in the show what is one differentiative view you hold

Serrino Liu: oh I think I will say the fully automatic agent AGI. I think it’s one thing i’ve been circling the whole time: in creative work, the most expensive thing was never execution. It’s knowing who to pass the ball to. It comes from the experience and judgement you’ve built up as a person. but it actually comes from the experiences as the judgment you as a people and in the real life and the relationship of the people is actually deciding who you are and what is your relationship with AI so I think AGI is just a very misunderstanding knowledge in this conversation.

Grace Shao: thank you so much Serrino for your time today. It was lovely meeting you in Macao and we had a lovely conversation.

Serrino Liu: Thank you so much.

AI Proem is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

Get full access to AI Proem at aiproem.substack.com/subscribe