[afcool83]

Admirable idea and execution…but it does apply opposing evolutionary/economic pressure for AI-slop to become less detectable over time. AI will learn and adapt.

Metaphorically speaking, it’s the Borg we’re dealing with, not the Klingons. All Janeway did was slow the Borg’s progress.

[mapontosevenths]

Cory Doctorow wrote a story ~20 years ago about how the first sentient machines would be spam bots because their job is to pass as human, and anti-spam systems provide competitive evolutionary pressure.

He may not be too far off.

[tetris11]

This one?

https://craphound.com/overclocked/Cory_Doctorow_-_Overclocke...

[mapontosevenths]

I think that's the one. I was a bit off on the timing, it's not 20 yet. Great read either way.

From the story:

“Spam-filters, actually. Once they became self-modifying, spam-filters and spam-bots got into a war to see which could act more human, and since their failures invoked a human judgement about whether their material were convincingly human, it was like a trillion Turing-tests from which they could learn. From there came the first machine-intelligence algorithms, and then my kind.”

[alansaber]

It's actually rather difficult for SoTA models to shift tone without losing performance on various datasets, so not such a one-sided arms race.