The really sad thing is that people have discovered that "completely objective" doesn't sell very well.

Taking an extreme view sells much better, but provides much less useful information.

"Discovered"? It's always been this way. Controversy sells. If it bleeds, it leads. Etc...

I remember, when first doing academic publications, my supervisor told me controversy sells. It was true then, it's true now, it was probably true when newsletters were printed in cuneiform on stone tables.

But also note what the goals can be: a "computed" summary vs clever, intelligent, wise insights. Note that some texts are built so as to include both: a flow of core analysis with quality commentaries as asides.

No, not necessarily. From the events you go to the chains of events, the clusters, the trends, the teachings etc.

Similarly to experimental science, where you go from the protocols ("this happened there then") to the "laws".

The role of «decisions and bias» could be limited, with some similar quantitative, and maybe collaborative (aggregation of multi-agent contribution) approach that ranked the outputs through a computed importance.