WeeklyTalk #157
AI in software development – a blessing or a curse?
In this episode, we are joined by Timo to discuss AI in software engineering.
In this episode, we are joined by Timo to discuss AI in software engineering. Contrary to the current AI hype, Timo offers a more skeptical perspective. He argues that the effort invested in prompting could be better utilized through direct problem-solving and manual coding. Furthermore, he highlights a growing risk: the loss of the critical skills required to evaluate machine-generated output. This episode explores the fine line between AI as a helpful tool and AI as a detrimental crutch.
00:00:00 - Introduction: AI Euphoria vs. Critical Distance
00:00:13 - The Thesis: Prompting Time vs. Thinking Time
00:01:10 - Challenge: Conveying Context and Codebase
00:01:54 - Differences Between Experienced Developers and Juniors
00:02:49 - The Vicious Cycle of AI Debugging
00:03:51 - AI as an Idea Generator and the Problem of Outdated Training Data
00:04:30 - Practical Example: When AI Hallucinates Library Versions
00:09:25 - How AI Systems Retrieve Information (RAG & Agents)
00:10:24 - Criticism of AI for 'Unpleasant' Routine Tasks
00:11:52 - Code Reviews and the Detectability of AI-Generated Code
00:13:08 - The 'Calculator Effect' and the Shift in Skillsets
00:14:51 - Dangers of Dependencies and Model Bias
00:15:13 - Advice for Young Developers: Understand Instead of Copy
00:16:25 - Conclusion: Retain Responsibility for Architecture and Quality
February 12, 2026