GOTO - The Brightest Minds in Tech Beyond the Hype: Real Talk on AI-Assisted Development • Jessica Kerr & Diana Montalion
Nov 18, 2025
In this discussion, Jessica Kerr, an engineering manager at Honeycomb.io, and Diana Montalion, a systems architect and author of "Learning Systems Thinking," dive into the reality of AI-assisted development. They highlight how AI excels in tedious tasks like typing and error handling but struggles with nuanced developer thinking. The duo advocates for a balanced relationship with AI, emphasizing the importance of human oversight. They explore AI's effectiveness in rapid prototyping and its potential to enhance learning for new developers, all while urging careful management of AI-generated code risks.
AI Snips
Chapters
Books
Transcript
Episode notes
Work Within Perpetually Imperfect Systems
- Jessica Kerr frames software work as operating inside a perpetually imperfect distributed system rather than achieving perfect correctness.
- She emphasizes compensating, adding redundancies, and designing responses to failures over eliminating all faults.
Design Failure Responses First
- Prioritize how your system responds to failures instead of trying to prevent every possible error in advance.
- Loudly fail in telemetry for developers and give users only the information they need to act.
AI Shifts Constraints From Typing To Judgment
- AI-assisted coding excels at typing and producing boilerplate, shifting the constraint from typing time to judgment and design.
- The tool cannot do the deep thinking required for nuanced system decisions.
