Sentience does not imply intelligence. Also, super-intelligence machines could develop neuroses or depression.
Technically, you are correct. Sentience implies perceiving, feeling sensory input, or even some form of consciousness. But in common use, the word often implies consciousness and thinking and intelligent reasoning.
In humans, neuroses are most often the result of conflicts between emotional reasoning, expectations that are based on flawed reasoning and what we experience as common reality.
Any "machine based intelligence" that develops neurotic tendencies would seem to have defined limitations in terms of both original programming and a qualification as "super-intelligent". If we program our own neurotic limitations into a machine intelligence, it's not super-intelligent, just another flawed copy.
In order to qualify as super-intelligent, it's necessary to go beyond simply calculating faster and storing bigger numbers or more facts and be able to create better knowledge based on a more reliable reasoning process.
This would produce better science because it is based on analysis and prediction and likewise better ethical analysis for the same reason. Any intelligence that fails at those, does not meet the qualification of super-intelligent.
“Sentience does not imply intelligence.”
No, but they clearly aren’t mutually exclusive either, and sentience, in my view, would make highly intelligent machines much more likely to become an aggressive threat.