Weber Wong
Founder and CEO of FLORA
Where to hear and read about Weber Wong1 episode or test
Weber Wong is the founder and CEO of FLORA, an AI creative tool on an infinite node canvas, built after stints in banking, venture and art school. The lesson most worth taking is to build from your own need as a practitioner, follow a first-principles vision, and measure depth of use rather than headcount.
product is like an artwork that is continually built upon for decades. There's no time in which it is ready.
How FLORA is Reinventing Creative Design with AI
Weber Wong: the story and the lessons
1Weber Wong's path from banking to FLORA
Wong moved from banking to venture to art school before building FLORA. As a banker he could model a business in Excel in two days. At NYU's ITP, getting one creative output through 10 tools took weeks.
He loved poetry and history and saw venture as studying change over time, applied to the future. At the VC firm no company felt creatively interesting enough to start, he did not feel like a "one of one" founder, and he missed art.
Real-time AI paintings at ITP became the first version of FLORA. On a full scholarship he built AI paintings from a webcam feed, cut at 20 frames a second and run through a fast image-to-image model at 50% strength. Other people asked how he did it, so he built a node-based tool for them.
Runway's founders came from ITP, which first showed him that technology could make art. He also saw early AI companies like Anthropic and Runway.
He chose to build a tool for every creative team, not an agency. Rising usage was costing him money, and his first idea was an agency where one creative does the work of 100. With no agency experience or customer connections, he aimed to sell the tool to every creative team.
He reasoned that the tool will replace much manual labor, so it is better to be enacting the change than reacting to it. FLORA started about two years before the recording and launched about nine months before.
2Why FLORA is a top-down creative canvas
Wong argues traditional creative tools are slow because they build bottom up. Adobe layers date to 1987 and video timelines to 1991, and Figma is a web version of the same model. AI creation runs the other way: ask for a rough version, then carve toward pixel perfection.
His question was why build a tool around one piece of media when a model can make it in seconds. The bigger opportunity is to connect the models and design the whole workflow.
The canvas felt obvious because he had seen creative systems built from nodes. ITP mixes technical and creative people, and tools like TouchDesigner and Blender nodes are common there for building reactive creative systems.
He admits it is not an easy interface to design, but says there is really no other choice for the most powerful tool.
Wong collapsed one node per model into text, image and video blocks. The alpha had a node for every model and was too complex and technical to be a generational product. FLORA rebuilt it from scratch for the full launch, so linking text to image brings in all text-to-image models.
"creative professionals do not care about Hyo Miniats O2. They care about having a great video." The simpler abstraction also suits later control and agent features.
FLORA aims to be the creative computing layer above the models. Wong expects the obsession with single model outputs to fade. He compares FLORA to Microsoft building BASIC, then DOS, Windows and Office on top of personal computers.
The current product is phase one of three, with control and agentic features planned. Open problems are making visual programs powerful yet easy, and sharing a workflow with a team and scaling it. He also says AI tools must eventually match the control of traditional software, and he plans to layer that in.
3How FLORA found product/market fit and its customers
Wong launched by naming professionals as the underserved group. The premise was that current AI creative tools are built by non-creatives for other non-creatives. Professionals went viral with it because someone was finally building for them.
At launch, AI-tool users and professionals barely overlapped. He sees a growing overlap of high-craft, AI-literate people with the taste to make something great.
He reads product/market fit in stages, starting with qualitative pull. Early signals were top agencies wanting into the alpha, and Pentagram generating hundreds of logo variations at once. True fit came when users codified whole processes as workflows and nodes generated per user skyrocketed, followed by smile-shaped retention curves.
With few early customers, signals from top teams come before quantitative data.
FLORA's north star is nodes generated, not weekly active users. Users who treat FLORA as a creative system have 100 to 200x the usage of those using it as a better Midjourney. Weekly he also checks signups, conversion to paid and weekly churn.
Caveat: he gave no revenue, user or retention figures, and the usage claims are his own.
Creative-system users are rare and high value. Wong estimates about 10,000 people or teams in the world work this way. About 1,000 use FLORA as a creative system and hundreds of thousands use it more shallowly.
Examples: a generative brand script does in two weeks a brand identity that normally takes six months and hundreds of thousands of dollars. Lionsgate tested 1,000 story beats as videos before greenlighting a film.
He narrowed to fashion, CPG and brand marketing where AI reaches production grade. Early interest came from everywhere. Brand and marketing is the biggest use, fashion and CPG second, film and VFX third.
Nike uses FLORA to visualize the shoe itself, not just its marketing. Wong still holds that any media can be made in FLORA, since it is all text, pixels and moving pixels.
4How Wong grows FLORA
Word of mouth from launches is the biggest growth driver. FLORA announces features on Twitter and has done nothing very sophisticated, yet signups grow steadily every week.
A creative partner program has users posting workflows and sharing work, building on FLORA's place in New York creative culture.
Wong hires customers' own creatives to teach new workflows. FLORA hired AI-savvy creatives from teams such as Pentagram to embed in other top creative teams, an idea borrowed from Harvey's legal engineers. He calls them the agency's agency.
"when we bring them to our customers and work with them, they do a better job than me even on showing them how to contextualize it for their creative process."
FLORA leads the category with events and an operating manual. It hosts talks at its office with people from Pentagram and A24, and published a Field Guide as an operating manual for the future creative team.
Teams now ask how to restructure their process for AI, so owning that education builds relationships and a possible moat.
Project-level sharing raised organic signups. FLORA added project sharing and saw more signups from organic sources, beyond what top-of-funnel marketing brought in.
He expects more growth loops and sharing to spread power users' knowledge inside teams, and plans to track loops.
Onboarding now teaches how to fish before showing the big fish. FLORA replaced onboarding built on big example workflows with one that shows how to connect text to image to video. It has gone well.
The plan is to show an impressive result first to motivate users, then teach the skill. That work is unfinished.
Wong tests cheap tweaks like green buttons even this early. Some buttons changed from white to green, and he says it has recently been helping, though he gave no figure.
FLORA runs few formal A/B tests so far but plans more.
5Wong's rules for vision, feedback and positioning
Wong takes user feedback only at the margins. FLORA has never pivoted and says it is one-third through its vision. The February first version had the right skeleton, and a request gets built when several users ask in the same week.
"the core innovations do not come from listening to all the noise and then trying to turn that into a bunch of small features." He admits this may differ from others in the space.
He asked what creatives need in 5 to 10 years, not what rivals skipped. Almost no one believed him at first, which he reads as the mark of a non-consensus view that also turns out right. He calls the category year one of a 10 to 15 year build.
He compares FLORA to Nike, founded by a runner for runners, and says no other tool in the space is built by creatives.
He makes the value prop specific to each kind of creative team. A modular tool can do many things, so it is hard to explain. FLORA learned early that product value must be communicated very specifically per team, a challenge he compares to Notion's.
His poetry background shapes this: poetry is about expressing yourself, copywriting is about expressing to someone else.
6Wong's gaps and what he changed
Wong hired a head of product because he had never built a product. A couple of months before the recording he hired someone who founded Design Lab and scaled it to $50 million in revenue. His experience with many tools and learners improved FLORA's onboarding.
Wong had only made interactive art, reading reactions on faces and tweaking code. He is strong on qualitative intuition and needed metrics discipline for a team of artists.
Art taught him to ship, but he got caught chasing perfection. At ITP he learned Unity in half a week, then made a film with it. Art has a point where it is ready, while a product is built on for decades.
FLORA got caught adding scope and trying to make things exactly perfect. It improved by getting things out, testing the interaction and iterating.
He says his biggest need is engineers. He wants 10 more staff-level engineers right now.
The conversation skips pricing, the cost of running many models, and how FLORA would hold its lead if Adobe or model makers copied the canvas.
How FLORA is Reinventing Creative Design with AI
FLORA's founder explains how building an AI creative tool for professionals, on an infinite node canvas and from his own needs as an artist, led to adoption by teams like Pentagram, Nike and Lionsgate. He covers the abstraction that simplified the product, forward deployed creatives as a growth channel, the metric that tracks real depth of use, and onboarding and sharing changes that helped.
Wong moved from banking to venture to art school before building FLORA. As a banker he could model a business in Excel in two days. At NYU's ITP, getting one creative output through 10 tools took weeks.
He loved poetry and history and saw venture as studying change over time, applied to the future. At the VC firm no company felt creatively interesting enough to start, he did not feel like a "one of one" founder, and he missed art.
Real-time AI paintings at ITP became the first version of FLORA. On a full scholarship he built AI paintings from a webcam feed, cut at 20 frames a second and run through a fast image-to-image model at 50% strength. Other people asked how he did it, so he built a node-based tool for them.
Runway's founders came from ITP, which first showed him that technology could make art. He also saw early AI companies like Anthropic and Runway.
He chose to build a tool for every creative team, not an agency. Rising usage was costing him money, and his first idea was an agency where one creative does the work of 100. With no agency experience or customer connections, he aimed to sell the tool to every creative team.
He reasoned that the tool will replace much manual labor, so it is better to be enacting the change than reacting to it. FLORA started about two years before the recording and launched about nine months before.
Wong argues traditional creative tools are slow because they build bottom up. Adobe layers date to 1987 and video timelines to 1991, and Figma is a web version of the same model. AI creation runs the other way: ask for a rough version, then carve toward pixel perfection.
His question was why build a tool around one piece of media when a model can make it in seconds. The bigger opportunity is to connect the models and design the whole workflow.
The canvas felt obvious because he had seen creative systems built from nodes. ITP mixes technical and creative people, and tools like TouchDesigner and Blender nodes are common there for building reactive creative systems.
He admits it is not an easy interface to design, but says there is really no other choice for the most powerful tool.
Wong collapsed one node per model into text, image and video blocks. The alpha had a node for every model and was too complex and technical to be a generational product. FLORA rebuilt it from scratch for the full launch, so linking text to image brings in all text-to-image models.
"creative professionals do not care about Hyo Miniats O2. They care about having a great video." The simpler abstraction also suits later control and agent features.
FLORA aims to be the creative computing layer above the models. Wong expects the obsession with single model outputs to fade. He compares FLORA to Microsoft building BASIC, then DOS, Windows and Office on top of personal computers.
The current product is phase one of three, with control and agentic features planned. Open problems are making visual programs powerful yet easy, and sharing a workflow with a team and scaling it. He also says AI tools must eventually match the control of traditional software, and he plans to layer that in.
Wong launched by naming professionals as the underserved group. The premise was that current AI creative tools are built by non-creatives for other non-creatives. Professionals went viral with it because someone was finally building for them.
At launch, AI-tool users and professionals barely overlapped. He sees a growing overlap of high-craft, AI-literate people with the taste to make something great.
He reads product/market fit in stages, starting with qualitative pull. Early signals were top agencies wanting into the alpha, and Pentagram generating hundreds of logo variations at once. True fit came when users codified whole processes as workflows and nodes generated per user skyrocketed, followed by smile-shaped retention curves.
With few early customers, signals from top teams come before quantitative data.
FLORA's north star is nodes generated, not weekly active users. Users who treat FLORA as a creative system have 100 to 200x the usage of those using it as a better Midjourney. Weekly he also checks signups, conversion to paid and weekly churn.
Caveat: he gave no revenue, user or retention figures, and the usage claims are his own.
Creative-system users are rare and high value. Wong estimates about 10,000 people or teams in the world work this way. About 1,000 use FLORA as a creative system and hundreds of thousands use it more shallowly.
Examples: a generative brand script does in two weeks a brand identity that normally takes six months and hundreds of thousands of dollars. Lionsgate tested 1,000 story beats as videos before greenlighting a film.
He narrowed to fashion, CPG and brand marketing where AI reaches production grade. Early interest came from everywhere. Brand and marketing is the biggest use, fashion and CPG second, film and VFX third.
Nike uses FLORA to visualize the shoe itself, not just its marketing. Wong still holds that any media can be made in FLORA, since it is all text, pixels and moving pixels.
Word of mouth from launches is the biggest growth driver. FLORA announces features on Twitter and has done nothing very sophisticated, yet signups grow steadily every week.
A creative partner program has users posting workflows and sharing work, building on FLORA's place in New York creative culture.
Wong hires customers' own creatives to teach new workflows. FLORA hired AI-savvy creatives from teams such as Pentagram to embed in other top creative teams, an idea borrowed from Harvey's legal engineers. He calls them the agency's agency.
"when we bring them to our customers and work with them, they do a better job than me even on showing them how to contextualize it for their creative process."
FLORA leads the category with events and an operating manual. It hosts talks at its office with people from Pentagram and A24, and published a Field Guide as an operating manual for the future creative team.
Teams now ask how to restructure their process for AI, so owning that education builds relationships and a possible moat.
Project-level sharing raised organic signups. FLORA added project sharing and saw more signups from organic sources, beyond what top-of-funnel marketing brought in.
He expects more growth loops and sharing to spread power users' knowledge inside teams, and plans to track loops.
Onboarding now teaches how to fish before showing the big fish. FLORA replaced onboarding built on big example workflows with one that shows how to connect text to image to video. It has gone well.
The plan is to show an impressive result first to motivate users, then teach the skill. That work is unfinished.
Wong tests cheap tweaks like green buttons even this early. Some buttons changed from white to green, and he says it has recently been helping, though he gave no figure.
FLORA runs few formal A/B tests so far but plans more.
Wong takes user feedback only at the margins. FLORA has never pivoted and says it is one-third through its vision. The February first version had the right skeleton, and a request gets built when several users ask in the same week.
"the core innovations do not come from listening to all the noise and then trying to turn that into a bunch of small features." He admits this may differ from others in the space.
He asked what creatives need in 5 to 10 years, not what rivals skipped. Almost no one believed him at first, which he reads as the mark of a non-consensus view that also turns out right. He calls the category year one of a 10 to 15 year build.
He compares FLORA to Nike, founded by a runner for runners, and says no other tool in the space is built by creatives.
He makes the value prop specific to each kind of creative team. A modular tool can do many things, so it is hard to explain. FLORA learned early that product value must be communicated very specifically per team, a challenge he compares to Notion's.
His poetry background shapes this: poetry is about expressing yourself, copywriting is about expressing to someone else.
Wong hired a head of product because he had never built a product. A couple of months before the recording he hired someone who founded Design Lab and scaled it to $50 million in revenue. His experience with many tools and learners improved FLORA's onboarding.
Wong had only made interactive art, reading reactions on faces and tweaking code. He is strong on qualitative intuition and needed metrics discipline for a team of artists.
Art taught him to ship, but he got caught chasing perfection. At ITP he learned Unity in half a week, then made a film with it. Art has a point where it is ready, while a product is built on for decades.
FLORA got caught adding scope and trying to make things exactly perfect. It improved by getting things out, testing the interaction and iterating.
He says his biggest need is engineers. He wants 10 more staff-level engineers right now.
The conversation skips pricing, the cost of running many models, and how FLORA would hold its lead if Adobe or model makers copied the canvas.