I agree that using error monitoring service like Sentry, Rollbar, Bugsnag, etc. is better than a naked log for "exceptional" situations. They provide a lot of nice features including alerting, additional debugging info, the ability to mark bugs as fixed, etc.

I also agree that too many log statements can actually be counterproductive by providing a lot of noise.

But the idea that logs have no value is silly. Every developer writes bugs and will run into situations where the code is behaving differently than they thought in production. Being able to follow the logs to understand what happened is extremely valuable.