Vinton Cerf will step down as Google’s chief internet evangelist next week, ending one of the most influential careers in technology history.
Speaking via video feed at the Open Frontier conference hosted by the Laude Institute, Cerf was recognized by Dave Patterson, a professor at the University of California, Berkeley, best known for co-developing the RISC processor architecture.
“Vint has been at Google for over 20 years and is retiring in a week from today, so I think we should applaud his relatively successful career,” Patterson said, to cheers from the audience.
Google did not respond to a request for comment by the time of publication.
Cerf, 83, and his collaborator Robert Kahn are credited with the architects of the networking protocols that became the Internet as we know it today. His work in developing and popularizing TCP/IP (a basic set of rules that allows different computer networks to communicate with each other) beginning in the 1970s has been recognized with numerous honorary degrees, the Presidential Medal of Freedom, and the Turing Award, among others.
Since 2005, Cerf has served as Vice President and Chief Internet Evangelist at Google. (At this point, it’s safe to say that the Internet is fully ubiquitous, for better or worse.)
Cerf was speaking on a panel with other computer scientists known for their work on durable open source projects, including Patterson. François Chollet, creator of the Keras deep learning library and co-founder of Ndea. John Ousterhout is a computer scientist at Stanford University who developed the Tcl programming language and co-founder of Electric Cloud. and Matej Zaharia, co-founder and chief engineer at Databricks. They offered advice on what it takes to build a viable open source system. This advice is becoming increasingly important as founders bet on open infrastructure for the next wave of AI products.
Much of the discussion at the conference focused on the challenges of centralizing advanced models in a small number of resource-rich labs, as opposed to the decentralized world of the open Internet, which has made Cerf’s proprietary protocols more durable. But Cerf predicted that the rise of AI agents (software that can operate autonomously and work with other software) will lead tech companies to return to standardized protocols.
“Agent models for AI, where multiple agents from multiple sources interact with each other, force composability and require requirements for interoperability and standardization,” Cerf said.
If he’s right, the companies that define these interoperability standards early on could end up having a huge amount of influence over how the agent economy actually works. This movement is no different from the early Internet protocol wars.
While other panelists speculated that natural language communication between LLM agents would be sufficient, Cerf predicted that formal standards would be needed.
“I don’t think English is going to be the best choice. It has flexibility, but it also has ambiguity, and I think the accuracy of interactions between agents is going to be really important. Agents need to really make sure that the other agents understand what they’ve just agreed to do together,” Cerf said.
“Remember the old phone game where you wished you could whisper something in someone’s ear, but then by the time 10 people left, the message was completely different? Imagine a bunch of agents talking to each other in natural language. That’s kind of scary.”
When Mr. Patterson felt more upbeat, he recalled meeting Mr. Cerf, known for his three-piece suit wardrobe, as a graduate student in the 1970s.
“He was always the best-dressed computer scientist I’ve ever met,” Patterson said. “My memory of Vint is that he came as a graduate student in the ’70s wearing a shirt and tie.”
“That’s absolutely true,” Cerf said. “I used to wear vests, but for some reason I always wanted to stand out. Instead of growing my hair long and putting something in my nose, I thought it would be a good idea to dress differently.”
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
