Hi HN,
I've been thinking about the current state of AI development and see a familiar pattern emerging.
We're in the same situation mobile development was in 2007 – hundreds of fragmented frameworks and SDKs that don't interoperate. Today we have LangChain, OpenAI SDK, CrewAI, Vercel AI SDK, Mastra, and dozens more appearing weekly. Each has different APIs, integration patterns, and tool definitions.
I'm proposing Agent-Oriented Programming (AOP) as a solution: making AI agents first-class citizens in programming languages, similar to how Object-Oriented Programming elevated objects from structs.
Key concepts: agent keyword instead of class for AI constructs tool keyword instead of def for agent capabilities Native HTML elements like <agent> and <tool> New HTTP methods for agent operations (CHAT, TOOL, COMPOSE) Universal syntax across Python, TypeScript, Go, Rust Instead of this complexity: pythonclass MyAgent: def __init__(self): self.openai = OpenAI(api_key="...") self.search = SearchAPI(...)
You'd write:
pythonagent ResearchAgent():
def __init__(self, model="gpt-4", provider="openai"):
super().__init__(model=model, provider=provider)
This follows the same evolutionary pattern: Procedural → Object-Oriented → Agent-Oriented.
The proposal includes implementation roadmap, cross-language syntax examples, and how this could standardize the fragmented AI ecosystem.
Would love to hear HN's thoughts on this approach. Is this a natural evolution or am I overthinking the fragmentation problem?
https://www.aiop.dev/aop_intro.html |
No comments yet