AI is making a splash in the networking industry, and network professionals are cautiously optimistic about its potential. I mean, who wouldn’t be? After all, AI is like that prodigious child everyone loves to brag about at family gatherings.
Case in point: the Cisco Live 2023 conference in Las Vegas, where AI was the belle of the ball, popping up in keynotes and breakout sessions like it owned the place. Cisco’s chair and CEO, Chuck Robbins, even acknowledged that AI is moving quickly (no kidding), and organizations are scrambling to figure out how to use it responsibly.
Robbins posed some existential questions: “How are we going to use it? How well do we trust it? What’s the good upside for it? What are the risks?”Ah, the age-old conundrum of trust and potential in the realm of AI it’s like a Shakespearean tragedy, really.
But enough with the questions let’s talk about some solutions. Enter John Capobianco, developer advocate at Cisco Learning and Certifications, who showcased a ChatGPT plugin he developed called networkGPT. This fine piece of AI craftsmanship demonstrates how generative AI can mingle with networks and support network operations. How splendid!
Capobianco’s journey with ChatGPT began in November 2022 when OpenAI released the first version of its chatbot. He was quick to hop on the bandwagon, becoming one of the first users within five days of its release. To put things in perspective, he said, “It took Netflix 3.5 years to hit 1 million users. It took ChatGPT five days.”Talk about a meteoric rise to fame!
Fast forward seven months, and Capobianco had access to the ChatGPT 3.5 API, created a Django framework with buttons and drop-downs, and received plugin access for ChatGPT, resulting in the birth of networkGPT.
During the session, Capobianco demonstrated different ways he used the networkGPT plugin to interact with the network, using natural language and prompts. One use case is identifying downed interfaces the plugin can provide information about interfaces that have been shut down. Neat, huh?
But wait, there’s more! NetworkGPT can also check syslogs and highlight interface activity, sparing network practitioners from the drudgery of reading through pages of syslogs when a problem arises. Additionally, it can help understand devices in their Cisco DNA Center, Nexus data center fabrics, or other network environments by making direct API calls and describing device types, platforms, statuses, and IP addresses.
As Capobianco works on refining networkGPT, he urges network teams to implement safeguards when starting with prompt engineering. He emphasizes the importance of intentional prompts and the difference between generating, proposing, and running a configuration.
And for those who are concerned about privacy, fear not! Users can select an option within the GPT-4 model to turn off chat history, so it doesn’t learn from their prompts.
Lastly, Capobianco offers a playful tip for future AI-human relations: be polite in your prompts when using generative AI platforms. “Someday this is going to take a physical form and have laser beams for eyeballs that can melt our faces,”he said. “It’s going to come to my house and go, ‘No, not this one. He was always polite and said please and thank you.’”Ah, manners they just might save us all in the end.
