News for hackers. Stuff that matters.
Re:'Fiction is outperforming reality': how YouTube's algorit... by roenxi on Saturday February 03, 2018 @08:01PM (#16300634)
> There's an innocent recommendation: people find Trump more engaging.

100% this. There seems a to be a presupposition in the 'fake news' debate that somehow there are right and a wrong choices in politics and anyone making the 'wrong' choice must be re-educated by the powers that be.

I still recall the NSA scandal where it turned out they were indeed spying on everyone; that would undoubtedly have been branded 'fake news' pre-Snowden. One can easily imagine a world where a presidential candidate in favour of winding back the NSA apparatus was shut down, with videos talking about overreach being condemned as conspiracy theories.

The only solutions here are communication and engagement of people with different ideas. Trying to police the truth will not end well; somebody will end up corrupting the systems in ways that end badly.

Re:'Fiction is outperforming reality': how YouTube's algorit... by pulisse on Saturday February 03, 2018 @08:05PM (#16300642)
Trying to police the truth will not end well; somebody will end up corrupting the systems in ways that end badly.

YouTube's recommendation policy is currently, to a zeroth-order approximation, "push content that gets people riled up". Trying to get them to not follow that policy isn't policing truth; it's asking them to refrain from an affirmative act.