In recent years, the rapid development of artificial intelligence technology has sparked an efficiency revolution, with overseas ChatGPT and domestic DeepSeek leading to a surge in the development of numerous AI applications, as well as a significant increase in sales of AI GPU processors and server hardware. As an operating system and software ecosystem, can OpenWrt make a difference in this fast-evolving era of artificial intelligence?
The answer is certainly yes, but OpenWrt demonstrates “cleverness” rather than “brute force”!
Leveraging Edge Computing, Focusing on Specific Scenarios
The software ecosystem of OpenWrt is more suitable for network device products, so it cannot compete with general-purpose AI GPUs in terms of hardware computing power. However, not all scenarios require such high computing power; focusing on specific scenarios and applications can actually be more cost-effective. For example, Qualcomm is set to release its A7 Elite platform in 2024 to support edge AI applications.
However, network devices are cost-sensitive, which poses a significant challenge for manufacturers. Finding a suitable balance between “experience” and “cost” still requires exploration.
AI Applications Based on Cloud Services
After all, we are in the network era. Since the performance of local devices is limited, the capabilities provided by cloud services are virtually limitless, allowing inexpensive hardware devices to also benefit from the AI boom. While you are still hesitating, others have already started taking action. For instance, there is an OpenWrt AI Assistant Application project on GitHub, available at:
https://github.com/utakamo/oasis
This software provides functionality for OpenWrt to interact with AI. The AI offers optimal settings for OpenWrt based on user input; the OpenWrt device itself analyzes the information provided by the AI, verifies its validity, and notifies the user. The supported AI large models include:
- OpenAI
- Ollama
- Anthropic (Work in Progress)
- Google Gemini (Work in Progress)
- OpenRouter (Work in Progress)
The project also includes a settings interface for large models, such as the following settings interface for OpenAI:
The Digital Foundation in the Age of AI
Even if it is not possible to implement a complete AI application, hardware products driven by OpenWrt can still shine in the AI era, as hardware products provide the most fundamental data for AI applications, thus serving as the “digital foundation” for AI applications. Therefore, on the device side of the OpenWrt ecosystem, we need to focus on:
-
Telemetry technology
-
Remote access technology for cloud services
-
Technologies for constructing and formatting raw AI data
By utilizing the above technologies, we can provide “data materials” for cloud-based AI applications while responding to AI-driven strategic execution actions.