The Future of AI Creative Tools
Silicon Valley thinks creatives want one-shot AI novelty tools, but professional creatives actually need step-by-step control and repeatable workflows—a $42M insight from a VC-turned-artist who built FLORA by understanding what Pentagram, Nike, and Lionsgate actually need.
Read OriginalMy Notes (6)
The main change is that the creative means of production will have changed. It will be a lot easier to produce a lot of content and assets. That may restructure how creative teams are built and creative companies are constructed. In general, the value of rote manual creative labor will go down, but the value of taste will actually go up. If you have great taste, it suddenly becomes super easy to scale that.
Look at an agency—they get paid partially for rote manual labor, but also because they’re the best creatives in the world. I think value will accrue to those who are really great creatives, even though the initial narrative was that everyone can create now and the best creatives are going to lose their jobs. I have a fairly contrarian view based on the fact that we work with all these really great folks. When we put FLORA in the hands of the best creatives in the world, they make the best work we’ve ever seen. I expect they’ll continue to do very well.
We keep our roadmap pretty tight—we know what we want to build for the next two years and keep it under wraps until we release it. But I’ll say we’re working on deeper control—professional-grade control like you’d see in Figma or Adobe. A better visual programming experience to build more powerful workflows, bulk generation, things like that.
The design director said something interesting: too often shoe ideas win just because someone’s better at Photoshop. He wants to equalize the playing field and likes that FLORA can help a shoe designer see it all the way through.
One of our customers is an AI-forward agency offering a new type of creative service called a “Generative Brand Sprint.” He does an entire rebrand in two weeks versus six months, completely outcompeting his competitors at the same level of quality. Typically the brand identity process takes weeks of research and a lot of manual work. He has a fully autonomous FLORA workflow that codifies his process—he takes notes from the customer call, runs it through, and gets back to them in about an hour.
It’s no longer cool just to generate a cat video. What they’re looking for is control—how do you integrate what you’ve generated into a technical and creative workflow that makes sense, that’s compatible with other tools, in a repeatable and scalable fashion.
Fundamentally, they see creative problems as something you can one-shot. I talked to a model company CEO who said, “I think movies are a problem I can solve.” But the point of movies is not to solve problems—it’s to express and have control over every shot, the camera motion, to be really creative about every single scene. Creative work isn’t necessarily required—people do it because they want to and they want control over it.
Summary used for search
TLDR
• Most AI companies misunderstand creative work: they think movies are "problems to solve" when creatives want control over every shot, not one-shot prompt-to-pixels generation
• FLORA is a node-based canvas connecting generative models into repeatable workflows—Nike now lists "FLORA mastery" in job descriptions, agencies do 2-week rebrands vs 6-month timelines
• "Creative systems" are the future: autonomous workflows that codify taste and process, like Nike's sketch→360-render→campaign pipeline (previously 3 separate teams)
• Contrarian prediction: AI will increase the value of taste, not decrease it—the best creatives will do even better because they can scale their judgment
• Generative models are computers that need interface layers on top, just like personal computers needed operating systems and then creative tools like Adobe
In Detail
Weber Wong's thesis is that Silicon Valley fundamentally misunderstands professional creative work. Most AI companies see creative output as a problem to solve with better prompts—"fun pictures of your dog"—when professional creatives need step-by-step control over every aspect of their work. Wong compares generative models to personal computers: just as PCs needed operating systems (Microsoft) and then creative interfaces (Adobe) on top, generative models need proper interface layers. FLORA is that layer: a visual, node-based canvas where teams connect different AI models into repeatable workflows that respect the actual creative process (ideate → explore → decide → refine → produce → distribute).
The proof is in the customer list and use cases. Pentagram uses FLORA to explore hundreds of logo variations simultaneously instead of hand-crafting each one. Nike posted a job requiring "FLORA mastery" and built a workflow that takes shoe designs from sketch to 360-degree render to branded campaign—work that previously required three separate teams. An AI-forward agency offers "Generative Brand Sprints" that complete rebrands in two weeks versus six months by codifying their entire process into autonomous FLORA workflows. Lionsgate generates concept movies inside the platform. These aren't novelty use cases—they're production workflows at world-class creative organizations.
Wong's most contrarian insight is about what happens to creative professionals. While the narrative says AI democratizes creativity and threatens the best creatives, he argues the opposite: rote creative labor loses value, but taste becomes more valuable. When you give FLORA to the world's best creatives, they produce the best work anyone's seen—because they can now scale their judgment. The bottleneck for great storytellers isn't skill, it's time. AI tools that respect creative process don't replace the best directors, designers, and storytellers—they give them more years of output in the time they have left.