The Fat Pipe - Most Popular Packet Pushers Pods HN797: What To Do When The Business Asks for “AI”
Sep 19, 2025
Phil Gervasi, a seasoned network engineer and co-founder of Solutional, discusses the implications of AI initiatives from C-suite executives for network engineers. He emphasizes starting with business objectives before pursuing AI solutions and highlights the critical role of data engineering. Phil also explores different AI models and algorithms, the importance of tailored network architecture, and emerging AI use cases. Additionally, he introduces the concept of NeoClouds and shares practical strategies for integrating AI into networking operations.
AI Snips
Chapters
Transcript
Episode notes
AI Is Broader Than LLMs
- AI in 2025 is an umbrella of workflows, not just LLM chatbots.
- ML, data pipelines, and models (structured or unstructured) all sit under modern AI.
Model vs Algorithm Clarified
- A model equals an algorithm trained on data and then used to predict or classify.
- Training embeds learned behavior; inferencing runs the trained model on new inputs.
Start With Business Outcomes
- Before building AI, ask what business problem you are solving and whether AI is necessary.
- Map technical KPIs to business KPIs and avoid chasing technology for its own sake.