Are data lakes actually useful when implemented correctly?

I read a lot of opinions and complaining about data lakes on the net.
I don't fully understand what a data lake is - or a warehouse for that matter as I haven't worked with one yet. But my question is are they actually useful and if so - why are people complaining?

Are people complaining about data lakes because theirs is badly implemented and/or badly understood and/or badly integrated in the day-to-day work? Or are they complaining because as …

data data lakes datascience