SUI
There has been a lot of discussion around Web3 and AI for a while now, but what we are seeing today feels less like hype and more like a real shift.
Earlier, blockchain was relatively simple in its role. It mainly functioned as a system to store and transfer assets. But that is gradually changing. Systems are no longer just storing data they are beginning to understand user intent and act on it.
Data from the Sui ecosystem in 2026 makes this shift clear. Over 1.5 million AI agent compute nodes have already been deployed. These agents are not just sitting in the background they are actively interacting on-chain and executing payments. This changes the role of AI from being just a tool to becoming an active participant in the system.
This naturally raises a question:If AI can already handle execution, why should users still need to write code to build systems?
This is where User-Driven Development (UDD) becomes important.
But before understanding UDD, it is necessary to look at the core problems in Web3.
First ~ complexityBuilding in Web3 requires writing smart contracts, understanding security, managing backend systems, and handling data. For most users, this is simply too much.
Second ~ ideas cannot be executed easilyMany people have strong ideas, but without coding knowledge, those ideas never get built.
Third ~ heavy dependence on developersCreating anything new requires developers, which slows down innovation.
These issues together create a barrier that limits participation.
UDD introduces a simple shift:Instead of writing code, users express what they want.
In practice:
• Users give instructions in natural language
• AI interprets the intent
• The system generates on-chain logic
For example:“Trigger rewards automatically when user interaction reaches a certain level.”
What used to require detailed coding can now be defined in a single sentence.
But this leads to another important questionWhat actually makes this process work?
Without Dubhe Engine, UDD would remain just an idea.
Dubhe Engine is what makes this entire process practical.
Here is what it does:
• You provide instructions (like a conversation)• The system understands your intent• It generates smart contracts• Performs auditing• And deploys them
All of this happens without the user writing any code.
More importantly, Dubhe Engine is not just generating code it is handling the entire development workflow.
A natural concern is whether simplicity compromises safety.
This is where Dubhe’s design matters.
It is built on Move VM and the Dubhe Schema system, which ensures that AI-generated logic is structured rather than arbitrary.
This keeps the system predictable and controlled.
Numeron shows how this works in practice.
Users can define system behavior using simple instructions.
For example:“Evolve creatures based on transaction activity.”
AI interprets this, creates the logic, and deploys it.
This shows that building complex systems no longer requires programming expertise.
AI is already capable of handling many tasks independently.
In the future, multi-agent systems will likely emerge where:
• One agent handles security• One optimizes performance• One builds logic
Dubhe Engine can coordinate these agents and enable complete systems to be deployed in a very short time.
Web3 is going through a major shift.
Earlier, code was the main focus.Now, intent is becoming more important.
Dubhe Engine sits at the center of this shift by allowing users to turn their ideas directly into on-chain systems.
This could make Web3 far more open and accessible.
The question now is...how far can this model go?