The Everything Feed - All Packet Pushers Pods NAN086: A Reality Check On AI for Network Operations
Feb 26, 2025
Phil Gervasi, a seasoned professional in networking and technology education, explores the reality of AI in network operations. He discusses the potential of Large Language Models (LLMs) while addressing their challenges, such as hallucinations. Gervasi shares insights on integrating AI with low-code platforms and the importance of vector databases. He highlights the evolving role of network engineers and emphasizes the balance needed between excitement and skepticism regarding AI's capabilities in enhancing operational efficiency.
AI Snips
Chapters
Transcript
Episode notes
LLM Hype Cycle
- LLMs are likely past peak hype and entering the trough of disillusionment, according to the Gartner hype cycle.
- This means real use cases and true integration into workflows will emerge after this phase.
From English Teacher to Network Engineer
- Phil Gervasi's career journey began in English literature and education, then transitioned into technology through a summer job setting up computer labs.
- This experience sparked his passion for technology and eventually led him to networking.
LLM's Lack True Intelligence
- Large language models (LLMs) lack genuine intelligence and rely on probabilistic mathematical models to predict text sequences.
- This means they aren't deterministic, leading to varied responses for identical prompts.