The GibberLink project allows AI to communicate in robot language, saving computing resources, and has attracted 15 million views on X.
A weekend hackathon project called GibberLink is gaining global attention for its ability to allow artificial intelligence (AI) agents to recognize each other and switch to a more efficient computer language communication, which is incomprehensible to humans.
GibberLink, created by Meta software engineers Boris Starkov and Anton Pidkuiko, was born at a hackathon competition in London organized by ElevenLabs and Andreessen Horowitz. The core of the project is the ability to allow AI agents to detect when they are communicating with each other over the phone.
When the AI partner is identified, the system activates the GGWave protocol - an open-source library that converts data into digital sounds. To humans, this sound resembles a series of "beeps" and "boops", but for AI, this is a much more efficient information transmission method than simulating human speech.
Today I was sent the following cool demo:
— Georgi Gerganov (@ggerganov) February 24, 2025
Two AI agents on a phone call realize they're both AI and switch to a superior audio signal ggwave pic.twitter.com/TeewgxLEsP
"Communication through GGWave can significantly reduce computing costs," Starkov and Pidkuiko shared with TechCrunch. They argue that the process of converting speech to text and then analyzing it is unnecessary when two AIs communicate with each other.
Although the scenario of AI calling each other is still far-fetched, it is not impossible in the near future. Companies are deploying AI to replace call center agents, while tech giants like OpenAI, Google, and Amazon are developing AI agents capable of handling complex tasks on behalf of users.
Reactions to GibberLink reflect both curiosity and concern about the future of AI. The project's demonstration video has attracted over 15 million views on X and has been shared by influential figures like Marques Brownlee.
Starkov and Pidkuiko emphasize that the underlying technology is not new - it originates from 1980s dial-up modems, when computers communicated with each other through the "handshaking" process using sound over telephone lines.
The unexpected popularity of the project has created some interesting side effects. The GibberLink.com domain has been purchased and listed for sale at $85,000. A GibberLink memecoin has also appeared, along with unofficial online courses on "communication between agents."
The two creators claim they have no plans to commercialize the project and have published the source code on GitHub. However, they stated that they may develop additional related tools in the future.
