#17 Testing in the AI era

In this episode, Toby and Krisztian welcome their first proper guest: Alan Richardson, a 30-year software veteran and testing specialist known as Evil Tester. They dig into testing in the AI era: how to test AI-generated code, whether TDD still makes sense with AI, why self-healing tests are a red flag, and how AI is opening up security and adversarial testing. Alan makes the case for architecture-first development as the key to getting good test output from AI agents.

https://techleaguepodcast.com/

LinkedIn: https://www.linkedin.com/company/techleague-podcast/

Instagram: https://www.instagram.com/techleaguepodcast/

Apple Podcasts: https://podcasts.apple.com/us/podcast/tech-league/id1852602975

Spotify: https://open.spotify.com/show/1zx8UIe2EjawuVU3I31fdP

Chapters

0:00 Introduction and guest intro: Alan Richardson (eviltester.com)

2:00 Why testing matters more in the AI code generation era

5:30 Architecture-first: good code leads to good tests

10:00 Does TDD work with AI? Why it mostly doesn't

14:30 Playwright and UI tests: the abstraction problem

23:00 Information theory and what testing actually is

27:00 Adversarial AI testing: using AI to exploit your own CVEs

33:00 Security scanning tools vs penetration testing with AI

38:30 Domain expertise still matters

43:00 Generalist vs specialist in the AI era

47:00 The junior developer pipeline problem

51:00 Will AI homogenise software and design?

54:00 Wrap-up

Links:

Evil Tester https://eviltester.com

Playwright: https://playwright.dev/

Agentic EQ: https://agentic-qe.dev/

Vite: https://vite.dev/

Claude : https://claude.com/

Snyk: https://snyk.io/

Aikido: https://www.aikido.dev

Hacker One: https://www.hackerone.com/

Wiz: https://www.wiz.io/