This talk is from 2012, and it has aged better than most things said about technology since. AI has only made his central point sharper: change keeps outrunning what an organisation can learn.
The useful part is the language. Sloppy failure and smart failure are different things, and most organisations punish both the same way. Then they wonder why every AI pilot is designed to look successful rather than to teach anyone anything.
If you want people to run honest experiments, you need to be able to say which kind of failure you just had. Obeng gives you the words. It pairs well with Stop Planning, Start Probing.

