I’ll try to address your question, in my limited philosophical capacity, of whether or not an idea is “empirically good”, or if its quality depends on the perceptions of others. I’d proffer “demonstrably good” as perhaps a better term, but that is just my perception. ;)
The problem is one can start to slip into an argument based on semantics, such as observing that the word “good” is essentially a subjective term which could lead to thinking that it is likely all down to perception. But, one must consider context, practicality, and knock-on effects.
If in a particular context there is a desire for more or less of something, and an idea facilitates for more or less of this something, the method of facilitating this is very practical, and knock-on effects either don’t exist or are considered positive, then the idea is demonstrably good, within that context.
The problem is, you could have a context which has a demonstrably good idea within it which actually leads to something horrific. Think of all the demonstrably good ideas which came together to make thermonuclear weapons.
You might well have ideas which are, on balance beneficial, but does that make them absolutely “good”? We make judgments on what is and what isn’t important, and what is and isn’t beneficial. Life, human experience and cognition are highly complex; we have so many different ways of looking at things, rational or not, and many factors, physical and otherwise, to consider when a new idea is introduced. Context is king.
In the end, we are a collection of semi-rational primates with a bunch of needs and wants, and we are able to come up with ideas. Perhaps we can hit upon an idea which could only be considered a good one through empirical demonstration, but with all our different wants and needs, perhaps it will be near impossible to nail such an idea down. There are ideas which demonstrably work better than others, but maybe that is all we can hope for?