Both of those visions of science are flawed, in very similar ways: Each seeks to maximize something without acknowledging the tradeoffs.

Trivially: If science is to avoid belief in untrue things, you can instantly complete science by not believing anything.

The "figure out what is likely to be true" at least isn't so trivially dismissable.

Science should maximize expected value. The difficulty in practice is that you often must understand something before you can know what utility this understanding will provide.