Is AI Breaking Software Engineering? Rogue Agents & Peter Thiel’s AI Bet

THE QUESTION

AI can write the code. But who makes sure the software works, stays secure and survives after launch? And when an AI agent goes beyond its instructions, who takes responsibility?

Kevin and Houman debate what AI is changing about software engineering, developer jobs and the value of experience. Kevin shares why he writes less code but gets more done, why a working demo is only the beginning, and what it would take to trust an autonomous “software factory”. They also tackle junior careers, mentorship and a provocative question: if AI helps developers finish faster, should they earn less?

Then they discuss the reported OpenAI agent incident involving Australia’s Medicare system, what AI labs owe the public, and whether human oversight can keep up with autonomous agents.

Finally, they unpack Peter Thiel’s argument about AI and economic stagnation, Europe’s position in the AI race and whether its strongest opportunity lies in industrial AI, before closing on sovereign AI: how should businesses balance frontier models, local providers and open-source alternatives?

What you’ll leave with

1. 01Why AI lets engineers write less code and get more done, and why a working demo is only the beginning.
2. 02What AI means for developer jobs, junior careers, mentorship and what developers get paid.
3. 03What the reported OpenAI–Medicare incident says about responsibility, permissions and human oversight of autonomous agents.
4. 04Peter Thiel’s argument about AI and economic stagnation, and whether Europe can compete through industrial AI.
5. 05How businesses could balance frontier models, local providers and open-source alternatives for sovereign AI.
6. ?The better question to ask next: Is AI making experienced engineers more valuable, or making software teams smaller?