AI for Software Testing: The Ultimate Journey
summary
Claims that enterprise apps can self test using ML led to a moonshot project codenamed pterodactyl that auto-discovers an app, builds a state model, generates and reruns tests and learns over time but struggles with scale memory use and oracle/trainability gaps. An open source toolkit agent goes on GitHub, university partners (FIU, University of North Texas Dallas) contribute, bots classify app trustworthiness faster than 70 human testers, and some pieces work while hard problems remain.
AI-written summary · about
chapters (9)
- 0:00Start
- 0:12Keynote
- 10:10Going Back to the Future
- 25:43Training the System
- 34:05How Do We Test if an Application Is Trustworthy
- 37:30Award for Coach of the Year
- 50:08Which Skills Become More Important for Test Engineers To Have in the Future
- 51:01Critical Thinking
- 52:38Vision of 2024
related talks
01h 01mRise of the Machines: Can Artificial Intelligence Terminate Manual Testing?
17mAI-Powered Integration Testing
42mThe Path to Autonomous Testing
47mThe Path to Autonomous Testing
35mApp Testing with AI Tools – Make it Easy
44mAI Empowered Software Testing
41mNot Only Cars: “AI, Please Test My App”
44mHow AI is Exposing the Existential Crisis of Software Testing