That's a great example, because I much prefer the robot version of the Denny article to the human created version.

For me, neither was worth reading and since the machine generated version was shorter it was less not worth reading than the longer one...but the longer one contained the logical possibility of being worth reading whereas the machine written one was as good as it could possibly be.

Or to put it another way, while it was not worth the human effort to write the story, it wasn't worth the CPU cycles to write the machine generated version either. The story was not worth writing or publishing or reading at all because nobody cares including the author (which is why a machine can write it).

> The story was not worth writing or publishing or reading

Of course, these are the wrong metrics.

The correct metrics are: ad impressions vs. cost to generate content.

That's only for the publishing side. The reader's utility calculus matters too, and I agree with the other poster that both stories are garbage.

Ad impressions read by who ? Triggering which action ?

In that vein and thinking about click-bots, I'd favor revenue over ad impressions.

They both had their advantages. The human one wasted words on being cutesy but had the Las Vegas analysis.

You call it analysis, I call it bullshit.

Doesn't every bit of research we have show how hopeless all this analysis is?