Original · Unique Research · 2026-03-11
Editor's note: This is the complete English rendition of Unique Research's historical founder profile and interview. First-person narration belongs to the original interviewer. Long Yi is a romanization of the founder's Chinese name. The account of his background, product capabilities, comparisons with Canva and Adobe Firefly, user feedback, pricing restrictions and plans for Veeso are retained as source or interviewee statements; no independent product test, user dataset or employment verification accompanies this edition. The Malaysian user's experience is an individual anecdote, not a general finding about national or ethnic preferences. The source's description of its own product's paywall concerns that product, not this free Substack article. Event attendance and year-end forecasts are historical, not current invitations or promises.
FOUNDER PROFILE
Long Yi: The Programmer Who Mined Bitcoin in High School Wants AI to Design for Everyone
He Removed the “Template Library” That Is Almost Standard in Design Tools
"
“Users should not have to adapt to templates. Templates should adapt to users' content.”
He mined Bitcoin in high school, worked as a programmer at a major technology company, and has been described as an entrepreneur who “relies on intuition.” Long Yi carries many labels.
But what impressed me most was his counterintuitive decision while building Seede AI: to remove the “template library” that is almost standard in traditional design tools.
“Every design tool has templates, and users are accustomed to starting with a template and modifying it. But what I actually observed was that most ordinary users became more confused after receiving a template.”
This was not rebellion. It reflected a programmer's deep understanding of “how technology can reconstruct trust.”
01 Bitcoin as an Awakening: How Technology Redefines the Rules
Long Yi mined Bitcoin while still in high school.
“To me, Bitcoin was not primarily a financial product. It was an exceptionally cool open-source project. Distributed ledgers and decentralized consensus—these concepts have become fashionable industry terms today, but at the time they planted a seed in my mind.”
That seed helped him realize early that the most valuable technologies are often not the most complicated ones, but those that can redefine the rules.
After graduating from university, he joined a major technology company to build technical platforms for the small circle of “internal engineers.” “I kept thinking: so many technical architectures are highly similar, yet they are rebuilt again and again because they serve different groups. That is an enormous waste.”
After leaving the big company, he saw a broader world: marketing professionals, operations teams, small merchants, and self-employed people. They lacked professional tools, but they too needed to express themselves.
“Equal empowerment means allowing people whom technology has long overlooked to use things that genuinely work well.”
02 His Wife's Everyday Complaints: From Programmer to AI Design Entrepreneur
The opportunity to build Seede AI grew out of his wife's everyday complaints.
“She is a typical marketing professional and frequently needs promotional materials, but every time it is painful. The design team needs to fit the work into its schedule, requirements have to be aligned repeatedly, and the template finally delivered is still difficult to modify. Neither side is satisfied, but the work has to get done.”
As Long Yi listened, he suddenly realized that this collaboration problem, in which “no one is wrong but everyone is uncomfortable,” was precisely AI's greatest opportunity.
It was not that design tools were insufficiently good. They had never been built for ordinary people in the first place.
Canva lowered the barrier, but it still essentially starts with templates. Adobe Firefly is powerful, but it remains a tool for professionals.
“Users do not want a tool; they want a result. A microbusiness owner does not understand design and simply needs a usable poster.”
03 Removing the Template Library: Intuition in the AI Era
Seede AI's most counterintuitive design choice is the absence of a template library.
“A template is not empty. It contains preset images, headings of different lengths, and several pieces of copy. Users have to fill in the content one item at a time and consider how to adjust it without damaging the original structure. In essence, they are still doing design work; the blank page has merely been replaced by a half-finished product.”
Long Yi's intuition told him that in the AI era, users should not have to adapt to templates; templates should adapt to users' content.
Seede AI takes this approach: users supply only the copy, and AI automatically understands its hierarchy, emphasis, and logic before generating the layout best suited to that content.
“Users do not need to know what the template looks like or think about how to fill it in. AI takes care of all of that.”
Why choose design?
“Because creativity and design are themselves highly intuitive fields. It is difficult to imagine a perfect design before we begin, but when we see one, we immediately know whether we like it or not. That kind of judgment is closer to an ordinary person's intuition than creation is. What we want to do is let AI complete the creation step while leaving the right to judge with the user.”
04 The Moat Is Not the Model but the Use Case
Frankly, startups really cannot compete with major companies on models and resources.
“But when it comes to making AI work in design for the public, our moat is the deep coupling of technology and use cases.
The first layer is the additional use-case data accumulated through the unglamorous, labor-intensive work.
General-purpose models generate an 'average result.' What is genuinely valuable is the trail of corrections users make through repeated adjustments and further editing: which elements are removed, which styles are emphasized, and how the final piece is revised into shape. Those actions are themselves data.
Seede AI is like a data flywheel: while users meet their needs, they leave behind process data from nonstandard use cases, which in turn helps us understand what they actually want. Major companies cannot obtain these 'correction trails from real workflows' merely by piling on compute.
The second layer is an end-to-end service loop.
We do not just generate designs. We can also make fine adjustments and even connect to printing services. Major companies often cannot justify this level of service detail in terms of return on investment, whereas it is precisely where small teams are most flexible and effective.
“So major companies have traffic and compute; we have depth in design use cases. A moat is not a wall, but a 'use-case tunnel' that keeps being dug deeper.”
05 Globalization: Differences in Aesthetics Are the Biggest Barrier
For Seede AI, the most important breakthrough in 2026 is globalization.
Its overseas version, Veeso, has already launched, but Long Yi knows exactly where the challenge lies.
“The greatest barrier to taking an AI design tool global is differences in aesthetics, followed by the product and interaction experiences that grow out of those aesthetics.
A design tool essentially produces a 'visual language,' and visual language is more implicit and harder to standardize than written language. What is described as 'grand' or 'refined' in a Chinese context may look 'crowded' or 'cold' in another culture. These underlying aesthetic differences work backward to determine which controls a product should offer, what its default settings should be, and what users expect from its interactions.”
Seede AI has already received some interesting feedback. One Malaysian user felt the style “suited his needs exceptionally well” because he was Malaysian Chinese and his aesthetic preferences were close to those in China. Users from Canada and France, however, clearly expected something different.
“This made us realize that aesthetics are divided not by country, but by cultural communities. Our next step is not simply to create a 'US version' or a 'Southeast Asian version.' It is to understand the underlying logic of these visual preferences so that AI can recognize and adapt to what looks good in different cultural contexts.”
06 The End State: From Tool to Platform
Seede AI's future end state is a platform, not a tool.
“In fact, we are already more than a tool. We are using AI to reconstruct collaboration in design. In the future, I hope to see a marketing lead put out the design requirements for a large event, with the work behind the scenes performed not by a traditional Agency but by an Agency made up of AI. That is a completely different collaboration network.”
Long Yi is also considering hardware products that would package the entire AI design loop, from requirements to delivery, as an integrated offering for users.
“We have never confined ourselves to the category of 'AI graphic-design tools.' Users do not want a tool; they want a complete solution for turning ideas into deliverables. As long as we keep providing better solutions and enough value, we can survive and grow.”
In Closing: Let Technology Empower Every Creator Equally
Long Yi says “letting technology empower every creator equally” has two meanings.
The first is leveling capabilities. In the past, good design required talent, training, and expensive software—a privilege held by a few. In the AI era, a street vendor in Indonesia and a designer in New York can use the same tools to turn their ideas into visuals. That is the first form of equality technology brings.
The second is aligning starting points. When we look globally, we find creators in many places who have no concept of “design tools” at all. For them, equality does not mean “using better tools,” but “having tools for the first time.” Giving visibility to these long-overlooked voices is a deeper form of equality.
At the end of the interview, Long Yi said:
“Technology can be universal, but products must adapt. Without adequate understanding of and respect for local culture, even the best technology cannot produce something that truly resonates. Globalization is not translating a product into different languages. It is using different languages to understand users anew.”
Perhaps that is Seede AI's answer: not teaching users around the world to use design tools, but teaching AI to understand users around the world.
Long Yi, founder & CEO of Seede AI, will attend the 2026 Unique Awards Hangzhou Summit Panel discussion: From Open-Source Roots to Global Growth—New Globalization Opportunities for AI Productivity Tools.
Selected Interview Q&A
Interviewee Snapshot
Q1: Introduce yourself in one sentence. How did your experience of “mining Bitcoin in high school” influence your later entrepreneurship?
A: To me, Bitcoin was not primarily a financial product. It was an exceptionally cool open-source project. Distributed ledgers and decentralized consensus—these concepts have become fashionable industry terms today, but at the time they planted a seed in my mind. Looking back, that may have been my earliest awakening to how technology can reconstruct trust and change the way people collaborate. Bitcoin made me realize that the most valuable technologies are often not the most complicated ones, but those that can redefine the rules.
Follow-up: You have been described as an entrepreneur who “relies on intuition.” How does that intuition manifest itself specifically in AI design tools?
A: I think “intuition” here is better understood as the accumulation of long-term observation. By putting myself in different people's shoes and repeatedly feeling things through and thinking them over, I develop a judgment that is difficult to articulate but can point the way. For example, at Seede AI we removed the “template library” that is almost standard in traditional design tools. Many people might see that decision as risky, but my intuition told me that users should not have to adapt to templates; templates should adapt to users' content.
Rapid Positioning and the Moat
Q2: What is Seede AI's core business model? How does it differ from Canva and Adobe Firefly?
A: In one sentence: Content in, design out—enter the key content and receive a deliverable design directly. Canva represents the previous era: start with a template, then fill in the content and change the layout yourself. Seede AI represents the AI era: start with the content, and AI automatically generates the complete design. As for Adobe Firefly, it remains a tool for professionals; the barrier has not come down.
Q3: Who are your typical users: professional designers or ordinary people who know nothing about design?
A: Our typical users are ordinary people who do not understand design: marketing and operations professionals, content bloggers, and many people from all walks of life who are trying to design something themselves for the first time through Seede AI. Professional designers also use us—not to explore creative ideas, but to organize layouts quickly and improve delivery efficiency.
Q4: What is Seede AI's hardest-to-replicate capability?
A: Frankly, startups really cannot compete with major companies on models and resources. But when it comes to making AI work in design for the public, our moat is the deep coupling of “technology × use cases.” The first layer is the additional use-case data accumulated through unglamorous, labor-intensive work. The second is an end-to-end service loop. We do not just generate designs; we can also fine-tune them and even connect to printing services.
Execution and Commercialization in 2026
Q5: What are your main monetization methods today?
A: We currently charge users by selling credits, while also allowing them to try the product for free. At present, we do not impose many paywall restrictions; we limit only some export functionality.
Q6: What is the most important breakthrough or problem to solve in 2026?
A: Our most important breakthrough this year is to get globalization right, enabling our overseas product Veeso to find local product-market fit (PMF) quickly.
From Open-Source Roots to Global Growth
Q7: Is the greatest barrier to globalizing an AI design tool language, differences in aesthetics, or payment habits?
A: The biggest barrier is differences in aesthetics, followed by the product and interaction experiences that grow out of them. Design tools essentially produce a “visual language,” and visual language is more implicit and harder to standardize than written language.
Q8: Is Seede AI's future end state a “tool” or a “platform”?
A: Definitely a platform. In fact, we are already more than a tool: we are using AI to reconstruct collaboration in design. In the future, I hope to see a marketing lead put out the design requirements for a large event, with the work behind the scenes performed not by a traditional Agency but by an Agency made up of AI.
Q9: What capability do AI productivity-tool companies most need to develop as they move from a “standalone tool” to “global growth”?
A: The core bottleneck is understanding local markets. Technology can be universal, but products must adapt. Without adequate understanding of and respect for local culture, even the best technology cannot produce something that truly resonates. Globalization is not translating a product into different languages. It is using different languages to understand users anew.
Comparing Perspectives
Q10: Li Guohao, who will appear in the same session, is an Oxford postdoctoral researcher and founder of the CAMEL open-source community. How do you view “academic backgrounds” vs “grassroots entrepreneurship” in AI?
A: I think they are essentially different starting lines, but the destination is the same. Our team leans toward engineering, but we are not entirely “grassroots.” The core team and I came from leading technology companies and have built products with users on the scale of ten million. We are not “grassroots founders who do not understand technology”; we are pragmatists who do.
Business Collaboration
Q11: Where does Seede AI see the most promising incremental opportunity in the next 12 months?
A: Inspired in part by OpenClaw, I am very optimistic about providing Agent-facing design services and interactive-canvas infrastructure within the huge emerging market for proactive Agents.
Q12: What kinds of partners do you most hope to connect with at the Unique Awards?
A: We pay close attention to model evolution and applications. Veeso launched in early March, and we very much hope to find international distribution channels that can help us reach our target customers.
Closing Prediction
Q13: In one sentence, what kind of company do you predict Seede AI will become by the end of 2026?
A: A global design-services company—not only working in graphics, not only serving people but also serving Agents, and spanning the online and physical worlds.
Materials drawn from guest interviews for the Unique Awards.