Tech and Telecom

China’s New AI Can Build Apple OS Without Coding

China’s newly released AI model, Kimi K3, has attracted widespread attention after generating a functional macOS-style web application from a single user prompt.

The analyst who made the workable interface said that he asked Kimi K3’s Agent Swarm to recreate Apple’s upcoming macOS 27 “Golden Gate” with Liquid Glass effects and native-style apps running inside a web browser.

Ad Powered By Advergic
Loading ad . . .
Ad - Continue scrolling to read

Completing the project took over 3 hours and consumed a significant portion of the model’s monthly credits quota.

ALSO READ
Apple and Google Ordered to Remove Apps Creating AI Deepfake Nudes

The AI completed the task after about three hours and 20 minutes, according to Weinbach. He said the project consumed roughly 60 percent of his monthly Kimi credits allowance.

Weinbach later shared the finished project online, demonstrating a browser-based interface that closely resembles macOS, complete with working applications and Apple’s new visual design language.

Not the Real macOS?

Despite its appearance, the project is not Apple’s operating system. Instead, it is a web application designed to simulate the look and feel of macOS while running entirely in a browser. It does not include macOS’s underlying system software or kernel.

What Is Kimi K3?

Kimi K3 is the latest flagship AI model from Chinese startup Moonshot AI. Released this week, it is designed for long-running coding and agent-based tasks. The model supports a context window of up to one million tokens and includes an Agent Swarm system that allows multiple AI agents to work on different parts of a project simultaneously.

The model has also gained attention for its coding performance. AI benchmarking platform Arena reported that Kimi K3 topped its Frontend Code Arena, achieving a 76% pairwise win rate when compared with competing models on the same coding tasks.

Future

The macOS-style demo has fueled discussion across the AI industry about how quickly coding models are advancing. Instead of generating individual code snippets, newer AI systems can increasingly plan, write, debug, and assemble complete software projects from a single instruction.

While the above-shown experiment supports the capabilities of AI-assisted software development, it also confirms that more computing resources are still required.

Stay Connected with ProPakistani

Get the latest tech news, telecom insights, and product launches wherever you prefer.

Add ProPakistani to Preferred Sources and see more of our stories in Google Search and Top Stories.

Share
Published by
Business Desk