Subscribe
Share
Share
Embed
Conversations with designers, founders, and builders behind some of the best work
Weber Wong Full Video Podcast.txt
English (US)
00:00:00.040 — 00:00:16.160 · Speaker 1
How do I make something that reflects my ideas and still gives me a lot of creative control, setting up repeatable systems and being like a creative system designer? After we launched, it was kind of like stepping outside of Plato's Cave and everyone was like, wow, they've been cooking. I see what they're doing.
Whoever Wong was supposed to.
00:00:16.200 — 00:02:03.930 · Speaker 2
Be a venture capitalist, he did the whole thing. Investment banking spent two years at Menlo Ventures looking at companies like Anthropic and Runway. And then one day he was looking at the founders that he was supposed to back, and he realized that he wouldn't back himself. And so he he quit. He moved to New York, he got a job at a coffee shop.
And then he enrolled into New York University's ITP program, a grad school where artists used technology to create weird, beautiful things that, frankly, nobody asked for. And that is where Flora came from. If you haven't seen it yet. Flora is this node based creative tool that lets you build these AI workflows for image and video generation.
So it's not just this prompt box, it's actually closer to what would happen if you merged a touch designer with Midjourney and designed it by a bunch of creative pros. Webber's background is poetry, like actual poetry. He started writing poems when a Berkeley professor sat down with him at a coffee shop and explained to him the metaphors of Cat's Cradle, and that obsession with language speaking ideas into existence is baked into how Flora works.
For the next hour, we get into why node based interfaces have had such a weird reputation, and how Flora is trying to change that. What anti-SLAPP actually means when you're building AI creative tools, the difference between creative technologists and agencies, and what happens when pentagram reaches out to you and you realize that you accidentally built something useful?
This is State of Play, a podcast where I'm trying to figure out what fuels some of the most interesting designers and builders. Let's get into it.
00:02:11.009 — 00:02:17.890 · Speaker 2
How are you doing, dude? Like, how are things going? What's the current, uh, what's the curtain like feeling on this Tuesday for you?
00:02:18.970 — 00:02:53.330 · Speaker 1
Good. Um, we've been growing a lot. That's been great. Um, we just moved offices twice in a month, which is crazy. Uh, so I think last time we chatted, we were in an office downstairs that was smaller, and we just moved up. Um, yeah, I think, like, it's interesting. The product is only, like, 30% done. You have a huge roadmap ahead of us and, um, both very exciting and then also hard to prioritize.
So we're thinking a lot about these different dependencies of building out one thing first and then another. Um, but yeah, it's been a really exciting time recently.
00:02:54.370 — 00:03:05.690 · Speaker 2
Like, how are you choosing how to prioritize right now. Like, you have to move fast and you're in completely uncharted territory, and then you have to pick where to plant the flag. How are you planting that flag?
00:03:06.250 — 00:03:49.770 · Speaker 1
Yeah, I think like we have a very strong view of what? Like we want Flora to be able to do in like five years. The hard part is what do we build first and along the way to get there, there's a lot more layers on top of what we're doing. Basically. Um, what we would go so far is purely the foundation. Um, and the foundation still needs work.
There's like a lot of untapped potential there. Quality of life features, better workflow features, more control that we're we're actively working on. Um, but yeah, also trying to push ahead with a lot of, like the crazy stuff that when we build it and we launch it, I think we'll make as big of a as an impact as our first launch.
00:03:49.810 — 00:04:22.740 · Speaker 2
A while back, I was talking to Ben Huffman, the founder of contra, and he had this really great metaphor, which he called the Pain Cave, which is kind of that like moment you're in when you've got this vision and other people don't seem to see it. Maybe they kind of discount you, and all your efforts haven't really started compounding yet, you know?
And I feel like I've been in the pain cave a lot of my career. Do you feel like you're in a pain cave right now, or do you feel like people. Are you surrounded by folks who really, like, believe in this? Where are you at right now?
00:04:23.060 — 00:04:36.699 · Speaker 1
I think we were in a pain cave before we launched. After we launched. It was kind of like stepping outside of Plato's cave and everyone was like, wow, um, they've been cooking. I see what they're doing. I think we're in a very good position. I think we just need to
00:04:38.180 — 00:05:03.900 · Speaker 1
make sure that we're putting in the right inputs into our design process and your engineering process, and make the right decisions on what to build when listen to our customers. Um, and I think we'll be in a very good spot. I think we've proven that we're quite good at designing creative tools and building them, and we just need to get better and better at that to achieve what we're looking to achieve.
So I don't know. Overall pretty positive. Still working very, very hard. Um, but yeah, things are looking pretty good.
00:05:03.940 — 00:05:37.300 · Speaker 2
You have one of the most fascinating paths, man. Like, I love I love people who have these nontraditional trajectories getting into, like, creativity and tooling. I mean, you came from, like, banking and venture capital and then went to, like, interactive art school, and now you're creating the most, I think, ambitious new paradigm in the design world.
Did you feel like you were in a pain cave coming out of that? I mean, what was the what was kind of the impetus of like leaving that background and venturing into this seemingly completely unrelated area?
00:05:37.340 — 00:08:37.640 · Speaker 1
Yeah. Initially growing up, I was really into poetry. Um, and then I realized that I kind of have to make money for a living. So I decided to go into invest in banking and, um, partially because I knew I just naturally didn't gravitate towards that and finance in general. So I thought it'd be good to like, you know, eat your vegetables in some ways and like, learn the things that I'm not naturally good at.
Um, but the main reason was I wanted to work in a top investment bank. You would get a job with a top venture firm, uh, invest in the best startups, and then do my own. As I've stirred up so potentially somewhat of a creative act and the opportunity to own your own destiny and make something of value in theory.
Um, so I worked at this firm called Menlo Ventures for two years. They've invested in stuff like anthropic and whatnot. They're really great fun. I really enjoyed my time there. I did realize that after two years there that I wouldn't back myself. Uh, I wasn't a one of one founder. Um, I thought I had good ideas, but, you know, I didn't really believe in myself for any specific idea, so I decided to just quit, uh, move to New York, work at a coffee shop, and get into the art scene because I wanted to get back to doing what I was really passionate about.
I was interested in, like, conceptual art primarily, actually. um, but the week before I left, I looked at investing in runway ML. Um, and that was really interesting because, um, they're the only company I looked at that I would have wanted to start myself. And I asked the founders, like, hey, like, how did you guys get to doing this?
This is really different than the other enterprise software startups that I look at. And they're like, oh, we're from NYU, ITP, it's a graduate program at NYU Tisch where we use technology to make art. And honestly, I didn't know you could do that. I thought you could only use technology to make money, right?
Uh, I didn't know creative expression was on the table. So when I got to New York, um, before going into, like, the deep art world, I still did explore a little bit there. Um, I went to I got immersed in the art in tech scene in Brooklyn, and I met the head of ITP, uh, that program, and we really hit it off. He took me under his wing, and, uh, he gave me a full scholarship and let me join right away.
So in three months, I went from working in venture to, um, just working on creative technology. Art projects. Learn how to code. Learn how to do Arduino. Creative. Tech. All that stuff. Started working on these real time AI mirrors that, um, took real, real time webcam data. Texturing it with AI with an image image model running like 20 frames a second, and then outputted the result.
And the experience result was that you could see yourself in a Bettys painting lives, and if you moved around in the team, painting would change to reflect where you were and you could change it to any prompts. I also did a thing where, um, I put a GoPro on my head and then like, live stream my life for two weeks.
Um, and if you signed up for like the for Alpha waitlist, you could change the prompt and see my world in whatever prompt you wanted. Uh, people put in some pretty crazy prompts. Um, but yeah.
00:08:38.680 — 00:09:46.330 · Speaker 1
So, yeah, we kind of came out of doing stuff like that, and I started building a tool for myself to make it easier to make these different AI related art projects that were essentially AI creative systems, interactive systems. Um, at some point I thought about using that to do kind of creative work on an agency level.
And then I realized that, hey, like, I think people started using it a lot, even when it was just like and our project we were kind of working on. It actually started costing me some money. So, um, I realized like, hey, there's actually a big need here. If we build this out properly, a lot of people probably want it.
At the time, I wasn't thinking about a startup, but I realized that, you know, I would back myself to go after this because I love building creative tools, and this came out of my own organic need. Um, so that's when I decided to go off and do the startup. And that was like a year and a half ago. Um, a lot of the team is from my grad school, and, um, the requirement to join the team is you have to have built a creative tool before would be proficient at least three.
So we've kind of kept that creative culture coming out of ITP. Um, and yeah, you've been kind of at it for, for the last year and a half roughly. And it's been a pretty fun time.
00:09:46.930 — 00:10:56.930 · Speaker 2
Man, I'll tell you what. I've spent most of my career needing to make money, and it was always top of mind. And I never got to explore creative technology as art. But I would come across friends who were just like, creating the wildest, uh, like, technology. And I'm not even sure really what was powering it, but it would be like just responding to sound or following some sort of mathematical equation to like, hit certain patterns.
And I'm like, that's wild. I don't know how that works, but I can't make money with it. And then NFTs took off. And it's unfortunate that NFTs suffered a lot of kind of the reputation that they did because, yeah, on one end, a lot of people make memes about like, oh, people are creating low quality things and turning it into an NFT.
Some of the most cracked creatives that I've ever seen are creating, uh, were part of that NFT scene and the things that were coming out on that end of quality. We're blowing my mind. Today they're insane. Uh. Were you. Were you, like, plugged into that network at all of the people creating these, like, wild visuals?
00:10:57.730 — 00:11:33.570 · Speaker 1
Yeah, I mean, in two different ways. One is like Nike's. They, um. Yeah, they they also participated in that wave. And there's an interesting, um, aspect of that of, um, because it's digitally native in theory, you can use that to track like interactive art as well, in addition to just like, you know, static art.
And the other side of it was, um, when I was at, when I was in venture, I, I covered crypto and, uh, looked at a lot of the NFT art stuff, and that was really fascinating to me on that end as well. So it definitely been around that for a while and yeah, haven't kept up as much recently. But curious what's kind of going on in that world right now.
00:11:34.130 — 00:11:58.260 · Speaker 2
Well, and I've got to imagine, like Flora, our team has been using it to experiment with thumbnail generation and a couple of other things, like just recently. and it's it's kind of a new workflow. But but they're starting to just play. There's like this attitude of play on our team with tools like Flora.
What are the types of people that are like digging into Flora and immediately finding use? Is it like marketing? Is it product design? Like, what are you seeing?
00:11:58.460 — 00:13:25.860 · Speaker 1
It's honestly quite broad. I would say. Um, there's one big split that I would make of, like functional pixels and nonfunctional pixels. Functional pixels, like the whole like idea to interface design to code pixels that get interacted with. And the nonfunctional pixels is like branding and marketing, you know, just media filmmaking, stuff like that, where it's just pretty pixels to some extent.
We primarily focused on on that end, of course, that's also really large. Um, on that end, we've seen usage from, you know, design agencies. So like one of our first, uh, designers that we iterated with was from pentagram, even, like, folks, like, mischievous, uh, the founder of that is also an angel investor.
They use it to, like, many different things. Product photography. Uh, marketing agencies. Um, one very specific example I remember is, um, this Japanese guy who owned the sofa brand out in Japan, and he was like, I have 99 bottles of sake with different labels, but I only have one picture with the model holding my bottle.
Um, so he used to have every single image of a model holding the bottle and just change out the label for each one. Um, so it's been interesting to see all the different things people do, from storyboarding to product photography to just using text to come up with ideas. Um, just because, um, I guess there's a lot of ways to combine text, image and video together to create kind of creative process or workflow.
00:13:26.540 — 00:15:24.470 · Speaker 2
You just give me an idea. So we have this like private discord community of all these incredible designers. And, and we kind of came up with this unspoken rule that we're not trying to grow this thing into a giant community. We like it being this like hidey hole of dope talent, but we do want to expose it to some degree to other people who would really get a lot of value out of it.
And so we decided we were going to do like monthly or quarterly, like cipher puzzles on the internet and through my media release, little codes here and there. And if people care enough, if they're willing to put the work in, then they can go ahead. And, you know, if they crack the code, they get the discord invite.
And so the idea was to get these and I bought a few, um, regular Nintendo cartridges that were blank. And so we're running that into Florida and we're going to generate we have certain images we want to put on those that'll help us with this code. Like at scale. This isn't something I would have had time to do.
It's like, okay, cool. Ciphers are awesome. We're a small little team. We don't have time to like, put all that love and attention into detail on everything we do. But something like this becomes possible. And the question I have for you is, I heard somebody refer to say something called software as a gift, and you were talking about like nonfunctional pixels.
We're talking about like pixels as art. And I'm seeing people like Lee black designing these like audio Midi players in tools like framer or I'm seeing like Sam Pates, who's just like releasing like a pet rock app. And he has no intention to monetize this stuff. And I'm starting to recognize, like, there's this growing space of people designing like software as arts or these like, or a, you know, a cipher challenge right out of thin air.
Or do you think that that's going to become a larger group of people as this tooling continues to grow this way?
00:15:24.510 — 00:16:21.920 · Speaker 1
So in my grad program, that's like literally all we do. Like that's what the entire program is for. We just like no functional requirements. We're not trying to scale. We're just trying to make something sick. And yeah, it's really cool to see that the barrier to entry to making those things is really going down, I think.
Like at my program ITP, the main thing that people have is a high pain tolerance to like learn a new creative tool or learn creative coding so that they can do the creative vision that they want. And it requires both sides of the brain, like in a rational desire to make a gift or an artwork to, you know, and then the very rational ability to sit down and understand how to build it or learn the tool first, or learn the code first to go build it.
I think, um, that very entry goes down with like vide coding, hopefully with stuff like Flora where it makes a lot easier too. So yeah, I think it'll definitely spread even more. And it's cool to see more folks just kind of doing things for the sake of it, just because it's funny. Uh, that's always great to see.
00:16:21.960 — 00:17:05.800 · Speaker 2
What's interesting to me is when I meet people like you, and I don't feel like I meet a lot of people like you where you have this, like, intrinsic passion for things like poetry, but you also then go into like investment banking or venture capital. Right. And it's like typically and in my mind, commerce and ardor kind of at odds.
Like they're really hard to balance. And I know some people actually don't think they need to be. But Ben Blumen Rose doesn't, uh, think that needs to be the case. But I'm curious from you, how do you find the balance between, like, the necessity of money and then creating something beautiful that doesn't have to make money?
00:17:05.800 — 00:18:54.930 · Speaker 1
I don't know why I'm like this. I think, um, it is true that typically, like, you know, people that are into art are very oriented one way, um, and then people that are oriented towards, I don't know, going to invest in banking are very much one way. And I never felt really at home in either circle. Um, because I have an interest in both.
Flora is my way of getting, uh, both in the same place as much as possible. I do find it fun to to build a business. And to some extent, I find it fun to win. I also remember my co-founder because we were both getting recruited for college volleyball growing up. Um, but and even if I were to fully pursue art, I would want to be great at it.
But I'm also just very I intuitively driven, like, I just like doing things because they're cool and art is pretty, pretty sick. And, um, poetry was my initial interest, actually, because I don't think in the environment I grew up in, it was looked upon for me as an Asian male to, um, pursue art or a drawing class.
I was supposed to be nailing my SATs. So I think, like poetry to me, felt like an academic way to have creative expression and look like it was like, you know, high brow or whatever. Um, but, um, I'm glad I started there because, like, it's a very minimalist place to start and a weird coincidental overlap to, you know, creative with generative media, where a lot of it does start with the words.
Um, I think I, you know, Me starting to explore poetry like 810 years ago has given me a lot of experience and thoughts on what it's like to speak your idea into existence. Um, and was the reason why I probably got interested in generative techniques in the first place. The picture's worth a thousand words, and I wanted to explore how I could do that.
And floral was kind of an extension of that.
00:18:54.970 — 00:19:11.970 · Speaker 2
So coming up in an environment where, you know, poetry wasn't necessarily like on the the top list of things that you felt, you know, were socially acceptable. What was your exposure to poetry? How did you come across that and realize, like, this is really this is really something interesting to me?
00:19:12.410 — 00:20:06.140 · Speaker 1
Yeah, I was like 14 and I was sitting at UC Berkeley. I grew up in the Bay area. I was there for some reason, so I was like coffee shop reading the book Cat's Cradle by Kurt Vonnegut. It's a very chaotic book. It's really flat. And then this Berkeley professor walked up to his. She was like, you have a very confused look on your face.
Um, is there something I can help you help explain to you about what's going on in that book? And then she sat down and explained to me for an hour, like the metaphor is of beans within cat's cradle. And she kind of became like my literary mentor. Um, she exposed to me how literature is interesting and, uh, why language and communicating?
Um, it's really cool and beautiful in itself and got me into poetry and saw a lot of potential with me, even though I thought I was like a map person or whatever. She helped me realize that I'm really into the humanities, literature, poetry. Uh, and yeah, that's kind of how I got into it all.
00:20:07.340 — 00:20:25.220 · Speaker 2
That's pretty. That's pretty cool to see an interaction or to have an interaction kind of serendipitously like that, to kick off what is now a really interesting path of creating tooling around this. What was your initial reaction when you like? What was the first experience you had with a large language model?
00:20:25.260 — 00:21:57.790 · Speaker 1
That's a great question. It's kind of like a big moment where like You realize that there is something, if not as intelligent as you, more intelligent than you thought possible. And then you think about what that means for you. I was still in venture at the time, um, and I remember seeing all the hype around it, especially because I was in SF, and then I tried it out myself and I was like, wow, it's kind of crazy.
And at first I was thinking about like, okay, like, there's probably investment opportunities here. But the second thing I thought about, I was in the shower and I was just thinking like, wow, like, this is going to change society. Like, wow. Like, this is big. Um, what does this mean for how are we thinking about ourselves?
I think there's been a lot of questions about what is intelligence? What is consciousness? Um, not because of the large language models, but because of how close they seem to us where they're making this question like, okay, what is our role in all this? Um, and that's been really interesting. And then the third thing I did was I tried training it on my poetry, and then I tried, like, seeing how good it was at that.
Uh, and it was it was interesting. Oh, I did a small sidequest where, um, I was. Yeah, like trying to get ChatGPT to, like, ingest all that poetry and upload it just for me. And it was like, uh, it was kind of all right. And then when I went to grad school, they taught me how to find to my own model. So I got GPT two, like an older version and then fine print out of my poetry, and it was better.
And the reason why is because it just drew these associations that, like, almost didn't fully make sense but were beautiful. And, um, that that was really interesting experience. Definitely.
00:21:57.950 — 00:22:05.870 · Speaker 2
Would you use an LLM to produce poetry? Is that something you've done? And if so, why? And if so, why not?
00:22:05.910 — 00:23:29.920 · Speaker 1
I stopped, like, trying to publish poetry. Uh, what I went into wasn't begging, but, uh, I still read a lot for myself. So I haven't grappled with the question of, am I doing it for an audience and am I signing off on it? Um, so when I write for myself, it's for fun when I fine tune that thing. It was also for fun, but I didn't really post it anywhere.
I basically like I had trained it on my Apple notes, and Apple notes is just like a great representation of the most intrusive thoughts and like random things that have happened. And it would refer to these things all throughout my life that were just super, super random that I never thought of and draw associations between them.
So to me, I felt like the consumer of of that poetry. They felt like it was doing like Spotify wrapped in a very, very poetic way for other stuff, like all of our copy. We try not to generate stuff. Um, I'm not, like, the best copywriter in the world. I do pay more efficient words than most, I think. And I can I can kind of, I can quite easily tell.
It feels like to me when something is written with AI or not, sometimes we have internal manuals being written. You can also tell when it's AI and when it's not. Generally when it's not. The quality of thought per word or sentence seems to be higher. I think the best combination is probably a little bit of both of like starting to write it out yourself first and the quality of thought initially improves the quality of thought of like GPT or whatever.
00:23:29.920 — 00:23:52.880 · Speaker 2
So how do you think about it? When Flora, which you know, I heard when we last spoke, you mentioned that you're trying to build kind of the world's most powerful creative operating system. And so, like, how would you define the most ideal scenario in that operating system where, you know, AI is is powering some of that?
00:23:52.880 — 00:25:27.370 · Speaker 1
I don't I don't think I want to be too prescriptive about what they find useful for it. I've seen some people find it really useful for consulting. I've seen other people find it really useful for building systems that they reuse over and over again. Of course, the use cases themselves are also very, very diverse.
Um, I think the ceiling of what we want it to be possible to do, even if people don't fully utilize it. So a single individual to have the same creative power as policies, um, because like, Um, right now, if you want to make a song and go viral, you have this software called Ableton. Um, any random kid in new Jersey in the suburbs can just make a trap beat, put it online, and go viral.
In the past, Mozart had to go find an orchestra. Right. Um, for visual creation. Like making a creative campaign, for instance. It's really high resource. How do I, as an individual, if you know my background is in poetry and I don't have any experience in all these different video image tools that are traditional, how do I make something that reflects my ideas and still gives me a lot of creative control?
Um, I think that's kind of like what we want to be able to achieve. And in addition to being able to being able to just make it, setting up repeatable systems and being like a creative system designer, where maybe you get some feedback that you want to change it a little bit. You can go in and tweak that, and it can be kind of a living system for some extent.
Um, that's kind of like roughly the direction that we want to go in.
00:25:28.530 — 00:25:42.690 · Speaker 2
I've got to ask you, because every time I've seen like node based and I've only recently been been introduced to certain technologies like Touch Designer, and if I'm not mistaken, they have a node based type of thing. If you played with that at all.
00:25:42.770 — 00:25:45.770 · Speaker 1
It's the most popular software in my grad program.
00:25:46.090 — 00:25:47.410 · Speaker 2
It's insane man.
00:25:47.410 — 00:25:49.210 · Speaker 1
It's cool, it's cool. Yeah.
00:25:49.530 — 00:26:06.610 · Speaker 2
Uh, what was the inspiration in thinking behind taking a node based approach versus this, like, infinite canvas? You know, uh, the kind of the I mean, you do have an infinite canvas, but like, the node based approach is just a very different way to think about it. And what what kind of influenced that?
00:26:07.850 — 00:27:47.180 · Speaker 1
A lot of things. So, um, touch Design was a big inspiration. And I think the interesting thing I found about it was that it had both the transformation and the, the visual output of that transformation in the same block. So you could change something and see it there and connect it to another thing. You can have entire interactive systems.
That was really cool. Um, a lot of people use it for Jane, so like, deejaying is disc jockeying. Vijay is a visual jock. Would like interact is visually see at concerts. Something about seeing something that reacts to you makes you feel good. It's kind of like looking at a mirror. So a lot of like artwork at ITP and stuff that I was doing, but with AI was around that.
Um, so I think that approach to like system design was always really interesting because at ITP, every time you make an artwork, we do that like you're the person you mentioned making like a rock app. Like they are making an interactive experience as art. Um, most creative tools are not systems. Um, there are sequential things that you do step by step.
It's control Z and control shift C. How do we layout control Z and control shift Z into a process? Because once you do that, it becomes a repeatable process and You get scale on your time and the amount of control you get for the amount of effort you have to put it like the cipher thing that you were setting up.
Um, I don't think a traditional creative tool you'd have to, like, do it manually over and over again, even though you already had the idea. So why can't you just as quickly as possible, get it there? And there's things we can improve for you to make that even faster. So I think the only way to accomplish that was a node based tool because it stored the transformations.
00:27:49.180 — 00:28:04.180 · Speaker 1
In some ways, I wish it wasn't that. I wish it was even more simple than that. And we thought we'd studied, like, dozens of different creative tools to figure out the right UX. It just so happens, like this is the right way. Um, so that's kind of like a lot of the thinking behind why we decided to go with that.
00:28:05.140 — 00:28:50.100 · Speaker 2
And that approach can be kind of intimidating for those of us who are so used to. I'm going to create my artboard, and then I'm going to go deep on the artboard, and then I'm going to zoom out and then maybe like, but with the it's, it's the going deep for a long period of time that I think is like part of the the thing that a lot of designers today are so used to.
Whereas it would seem to me the tools like Flora and that node based approach are really trying to facilitate this new paradigm that, hey, we can get to a lot of variations very quickly and then you can go deep. What would you say is some of the mindset, mindset shifting that you're seeing that needs to happen for people to to best leverage something like Flora at it?
00:28:50.100 — 00:29:59.510 · Speaker 1
People are often like living in the future because, um, it's called like centricity recently possible. And um, they basically take the latest technologies and figure out how to make art with them. So system design is just natural. Um, and that's why we kind of started working on this quite early. I think over time, the rest of the world will start to see the power of these kind of node based, workflow oriented tools and slowly move over to that.
Um, I think even Figma, initially when they launched, there was actually some resistance to that of like, this is not how I decided to work. I want my own time. I don't want to be in the cloud all the time. And then over time, that becomes more commonplace. And the idea of design systems was actually quite fascinating.
It was like just from designing, I end up with the repeatable system, and I didn't even think about it. And that's that was one of the biggest, most powerful unlocks there. And that required a mental shift to get there, too. I think ultimately, if we do our job and make a very powerful tool and ensure that we make it as intuitive and as palatable as possible, um, I think we're going to help a lot of people move around to that sort of view that hopefully gets some more scale on their ideas and helps them get to the end results.
They want better.
00:30:00.790 — 00:31:06.190 · Speaker 2
Now, because you guys are kind of approaching this as not just a insert, you know, design niche tool here, and you're more trying to create this creative operating system which, like you said, kind of spans different disciplines. I look at tools like blender that are the I'm going to say sticker shock, but it's like I see the interface and I'm immediately like X, you know, like, oh, wow, this is going to take me a whole.
I was talking to Ben Frick about this and he says, you know, even to this day, he's not a pro of all the different ways that he can use a tool like blender. Those interfaces can be intimidating. And so, you know, if a designer gets into Flora and they see the node based approach, that's probably the first thing that sticks out.
But if you can get through that it's like, okay, things start to make sense here. But I've got to imagine that, that the tools like that can very easily start to build on top of that complexity, and that complexity starts to show itself in the interface. How are are you guys thinking about keeping that intuitive and not not scaring people away with the,
00:31:07.590 — 00:31:10.470 · Speaker 2
the, the, the nuclear control panel, so to speak.
00:31:10.910 — 00:32:56.560 · Speaker 1
So node based tools have baggage. Um, node based tools are typically not well designed. And, uh, the reason why is because node based tools are visual programing. And typically in the past they were designed where one node was one chunk of code. Um, and then the noodles and the things you connect in are like the parameters you can put into that chunk of code.
So node based tools were designed around the way code works. We design nodes around the way the human brain works. That's why we just have text image and video. Not like VAD code or whatnot. Um, so as a result people are used to node based tools being really complex. But in spite of us being a node based interface, we've made it very, very intuitive.
I think from a first principles perspective, the barrier to entry is actually lower than signal. If you took someone that did not know any creative tools at all because you just connect box together. Figma. There's actually a lot of things you have for lyric. I think I've heard a lot from people that like after they get into it, in spite of like the initial like, oh, this is node based and they start playing around with it.
They find that it's very, very intuitive. In the in the long term, if we create a really great creative tool experience. In spite of the reputation of node based tools, people will tell each other, hey, this was actually a really good experience. I was able to make something really great. And that's ultimately what what went out in the future.
I think early on, a lot of people didn't want to kind of onboard the figure either. It was like very clunky at first. They were kind of doing this long term bet on, oh, somehow being in the cloud will help me, even though this is less functional. On one of our advisors was the head of GTM, the Sigma. So she tells us a lot about the early days.
Um, but then over time it became more and more intuitive. People really loved it, and people switched over to that paradigm entirely. I think the only way that we make that happen is just by to turn into great creative tools.
00:32:56.600 — 00:33:22.450 · Speaker 2
There's a lot of people right now who are, you know, waving a fist at this, uh, you know, AI slop generation. And when we last spoke, you said you're building floors specifically to reduce the amount or to avoid kind of AI slop in it. I'm wondering, how do you do that kind of practically, especially as kind of more these AI generated content is like flooding the creative field?
00:33:22.490 — 00:34:19.050 · Speaker 1
I think slop is a is a couple definitions, but you can be defined as something made with low effort, um, something made without a specific end goal in mind. Um, and in theory, you could make slop not using AI tools, but current AI interface is biased towards slaw. And the reason why is they make the barrier to entry, to making something very low, and they make it hard to refine on top of it.
For us, we were focused from the very, very beginning on creative control because creative professionals need that. It doesn't matter how fast AI is unless you get to something of creative value. Otherwise, I'd rather spend like eight hours on Adobe Illustrator to get it perfect, right? So that's kind of been our intent from the beginning.
That's why we say we're anti slut. Um, all of what we're designing for the interface. Um, one of our core principles is much creative control as possible. Um, so, like,
00:34:20.129 — 00:34:36.730 · Speaker 1
whether that's something that we're going to add soon. Like image compositing, right. Um, or just the way that we enable you to see many variations at once, so you can choose the one that's closest to what you want. That's naturally embedded in and how we design things, because we've all used creative tools that actually did give us creative control.
00:34:36.770 — 00:34:55.250 · Speaker 2
You said the phrase creative value. Um, and I find that that's an interesting way to define slop measuring based on its creative value. What how can you drill into that phrase a little bit more? What is what is the what is your metric or definition around creative value?
00:34:55.290 — 00:35:48.380 · Speaker 1
I think like especially last year, um, a lot of AI tools were used by consumers to have fun. Great. Um, I think now AI is being used more for professional settings where as a creative professional, we does this for a living. There's actually advantages to using AI now, which is great. Um, especially for like solo design agencies and things like that.
If it's something that you can make a living off of, um, maybe that has more creative value. So it's not really like making a trying to have an objective view on something that is subjective. It's just like, does this actually help a creative survive? Um, and like get their livelihood, get their idea out there?
Um, how do I make a power tool for creative professionals? Uh, I think maybe that's what I'm getting at. And slop is not directly juxtaposed to that, but, um, slop is more okay if you're just having fun.
00:35:48.420 — 00:35:49.139 · Speaker 2
I guess
00:35:50.260 — 00:36:28.380 · Speaker 2
I, I actually agree with that, too. I think. Slop, uh, and this might be a hot take. It's kind of a bad rap sometimes. Um, but I think it has its own. I think it has its own, uh, utility, which is play and learning, because I know that if I watch my I watch my 15 year old. Um, you know, just playing over the summer with these connecting windsurf with unity.
And he's like, got this work flow and he's building these different things. And that is a fantastic time to just generate some slop. You know, to just have some fun and to just see what, what you can build with and what what would the young Weber be doing with the tools if you were 15? Again.
00:36:29.100 — 00:37:13.830 · Speaker 1
I wouldn't be writing poetry. I would be just visualizing it because I think that'd be cooler. Um, one thing I actually hear a lot, which is interesting, is like, um, people that are like, oh yeah, I got really deep into AI because I've always wanted to draw, but I never had the chance to take a class or whatnot.
I probably fall in the same camp. I'd probably be doing something there, but. And yeah, on your point on slop, that's interesting. I mean, like, yeah, like one, one person slop in another person's play. And I think slop came out of this negative reaction from the professional creative class who felt like they didn't get something so dumb.
Um, and it was all like consumers saying, oh yeah, I just killed Hollywood or whatever. And it was like a negative reaction from their end. But, um, I don't think there's anything wrong with just, you know, having fun with it.
00:37:13.870 — 00:37:31.870 · Speaker 2
You had mentioned how, hey, like, if I have to go into illustrator and spend eight hours to create to get to that creative value, then, you know, I'll do that. What do you think AI is struggling with right now when it still struggles with, um, and how is Flora trying to kind of address some of that blind spot?
00:37:32.710 — 00:38:13.389 · Speaker 1
I mean, consistency. Um, I hear this from our customers. They'll be trying to get repeatable products on this, right? There's no way to have perfection, right? And creative tools. The traditional creative tools. Consistency was the default. We're just used to that now. Like randomness is the default.
And I think that that still trips up a lot of folks. And it is better for use cases where randomness is nice, but for things that require like Robotic just over and over again. I want this exactly in the same place. It's still hard. And I think for us, the next step there is to build build in a lot of traditional
00:38:14.710 — 00:38:29.270 · Speaker 1
creative tooling techniques so you can get the best of both worlds. So I think that's definitely something that's solvable. The models are still continuing to get better as well, especially in video. Um, so I think that will become much more manageable over time.
00:38:29.310 — 00:38:51.870 · Speaker 2
Now, you've regularly referred back to your your deep involvement with like New York's creative tech community. And I and I find that the community has just become such a huge part of navigating kind of this next evolution of, of creative work. How has that influenced how Flora's developing and your own, like, leadership approach in this space?
00:38:51.910 — 00:39:07.479 · Speaker 1
I don't necessarily think of myself as a leader. I think other people also have their own opinions. These are just mine. So there's actually a couple different communities. The creative tech community is an interesting one. They are quite different from most other folks. They are well versed in technology but are
00:39:08.880 — 00:40:48.680 · Speaker 1
almost into technology. For the complete opposite readings of the starter pool. The starter pool love scale. You know they just want to grow ITP creative tech world at horror scale. They want to make one of one interesting projects. Um, we're an interesting kind of mix in between. Um, there's the agency world.
Um, I would describe them as very task oriented, highly effective at what they do. Um, but in some ways closer to the business world, because they do deliver business outcomes than the creative technologists who are more like artists. They're very romantic and, um, do things for the sake of it more. I think there are some creative technologists that just see it as like an ingredient for them to make interesting art.
That's probably where I was at like a year ago. Um, agencies primarily see it as a a way to drive greater efficiencies from the top down perspective. Um, and and let's see it directly for, you know, the creative possibilities. What they like about it, above all, is a lesson. Be extremely accurate in their concept, not necessarily in the production yet, but in their concept.
Instead of pulling together different references from like Pinterest or whatnot, they can just generate the exact reference they want by blending up the different images. That helps them a lot. And yeah, I would say in general, like we've moved from like early adopter people adopting AI because they think it'll give them an advantage in the future.
So early majority, uh, people are starting to adopt AI because it's actually useful for them. And that's pretty interesting. Um, but yeah, that's kind of what I'm seeing so far. There's a lot of other like, little dynamics here and there that are just fascinating. Uh, but I'll leave it at that for now.
00:40:48.840 — 00:41:03.290 · Speaker 2
I'm very curious to hear, can you tell me about an instance where maybe feedback from the communities that you're paying attention to, um, maybe influenced some sort of decision making or reshaped anything key at Flora.
00:41:03.450 — 00:44:01.780 · Speaker 1
I think one of the biggest ones is like, so we were building our alpha version in like September of last year, and, um, I was still in grad school. I only dropped out like four months ago. I was trying to balance it too. As a result, I was still very much in that mindset of like with just build the coolest possible thing.
That's like be even less use case focus than I am now. Um, and then. Yeah, and then someone from pentagram reached out for some reason, and then we sort of like, we're surprised that we accidentally stumbled into something that someone at pentagram would want to use. Honestly. And then, um, we realized, like, oh, like, there's a different way of looking at Flora than we really expected.
So we started meeting a couple of other folks from, like, world class creative teams, and that helped us understand there are just some slight things that we can adjust about this that, um, make it really useful for a broader audience. Um, and in many ways, they want the same things out of Flora that we do, which is just make great creative work.
Um, and that was part of the impetus behind us basically just stopping development on the alpha and rebuilding the full version of floor that we have now from scratch in like three months. And then we launched that and that one really viral. So I think that was an interesting, pivotal moment for us. There are other things that cause us to, um, shift to the new version of floor that we have right now.
I think there's still a lot of workflow tools. I take a more, um, prescriptive, prescriptive approach where there are more types of notes. Uh, the main realization that we had was that, yeah, I know these tools are hard. People don't want to, like, learn a new type of way of doing it. So we need to make it as simple as possible.
Oh, yeah. And one thing that you mentioned earlier that I forgot to respond to was like, why node based tools? I think there's a nuanced thing here where like generally there's a trade off of convenience that control for every creative tool. Sigma and Canva. Fairly convenient, but you can only do like interface design and marketing with unity.
You can build entire worlds, but it takes years to master all. The first gen AI tools are generally hyper maps and convenience. But creative professionals do more, need more control. So we're focused on that. And there's generally like a trade offs. There's like an efficient frontier. Good design gets you to that efficient frontier.
And then you as a maker of a creative tool, choose how much of that trade offs do I want? That being said, there are ways that you can make the barrier to entry low, but the feeling of power. Hi. Um, we have just text, image and video. Um, that's a lot simpler than what we had before, which was like 20 different notes.
Um, but on top of just text, image and video, we can add a lot of additional stuff. We can add image compositing, video editing, a bunch of other stuff that we have in the pipeline. But when you first enter, all you have to do still is just connect three blocks together. Um, so you can kind of like it's kind of like an iceberg, right?
At first you only see so much, and then over time you can explore more and more. And once you get the underlying foundation. So it's not always a trade off to some extent, and I think that's been a big part of his philosophy while designing fora. And there's probably ways we can make it even more simpler just to get started.
00:44:01.820 — 00:44:51.380 · Speaker 2
And you're thinking about all these enterprise use cases, you're thinking about supporting the creative communities, and there's so much that you have to balance. So I want to add one more to your plate. I have two teenage I've got four teenage boys, but I'm thinking about two of them in this particular question.
One of them is, um, not really, uh, intrinsically motivated by traditional education, loves to experiment and build. Another one is, uh, we're about to go to our colleges, uh, in a few weeks, and we have things to look at. Very analytical. Um, both are interested in technology. You just said you dropped out.
Um, you know, because, uh, you were trying to. You were trying to go post-grad, and now it's like you have things to focus on. What do you tell this next generation of creatives trying to think about how to explore?
00:44:51.660 — 00:45:52.190 · Speaker 1
I think he should check out this thing called P5. I don't know if you've heard about p5 js at ITP. They didn't teach us how to code by like teaching us how to make like a back end. They taught us not a code by teaching us how to use JavaScript to make pretty animations where you control every single pixel on the screen using code.
And before that, I didn't have the motivation to code. And then that's kind of like what got me into it. And, um, I think it kind of teaches them the foundations of like, how do I actually literally control every single pixel on the screen? And once you learn that, you realize that all creative tools are just an abstraction layer above code.
If he goes deep enough in the unity, he'll have to learn like C-sharp scripts anyway. Um, and yeah, so I think he might find that interesting. I'll send you something. Does this guy called Dan Shiffman at the Coding Train teach us a lot of great videos on creative coding. That's how I learned it. There's also this undergrad program of ITP called EMA.
If he's interested in all that, it's at NYU. It's super interesting as well. Um, but yeah, I mean, I think like, it sounds like he's interested in a lot of cool things, which is sick.
00:45:52.230 — 00:47:47.960 · Speaker 2
So whoever said traditional creative tools gave consistency as a default. But with AI, randomness is the default. And that that flip, that sort of inversion is what's driving so many people. And kind of crazy. It's not that AI can't make good things, it's that AI can't make the same good thing twice. And for professional creatives, that is a deal breaker.
Flora's whole thesis is that you need systems, repeatable workflows that you can tweak and refine, not just a prompt box that gives you a new roll of the dice every time. And this framing of slop is definitely more generous than most. One person's slop is another person's play, and the negative reaction was in response to professional creatives getting mad when someone would say, Hollywood is dead in sharing a Midjourney render, but there's nothing wrong with having fun.
The problem comes when the tool is biased towards slop. It's easy to produce but hard to refine, and that's the trap that many AI tools have fallen into. It's a low barrier to entry, but an unclear path to excellence. And then there's Webber's origin story, where he went from poetry to investment banking to venture capital, and then off to grad school before building the tool that he needed.
And I think that's probably the least linear path I've heard in a while, but it kind of makes sense if you think of prompting speaking your idea into existence as a form of poetry, just with a different output. He said his product is 30% of the way there, and they've moved offices twice in one year, and he even dropped out of school to finish the job.
His big vision is to give one creative person the power of a Pixar studio. And the more I see from Florida, the more I'm starting to think he might be able to pull it off. We'll see. That's the episode. I'll see you next time.