Hacker News Reader

He might have a point if a simple API-change impacts 170 test cases. This sounds suspiciously like "no functional tests but let's go for 100% coverage with unit tests" - not necessarily the efficient thing to do, from a management perspective.

"Manual QA then? That will be even slower than automated..."

That product owner might have a good point. Far too many engineers I know think 100% coverage is a reasonable goal and spend way too much time writing unit tests that don't add much value or worse add negative value.