[repolfx]

I'm not sure it has much in the way of implications.

There is no real profit to be made by generating realistic looking text. Spammers don't work that way, spammers haven't cared about realistic looking text for years. Nor have spam filters cared much about text for a long time, exactly because it's so easy to randomise. Anti-spam is not a good reason to hold back on language generation models, in my view.

As for HN, if bots can write posts as good as humans, great, why hold back?

[Cacti]

You’re fooling yourself if you think there are no significant uses of text generation. Fake news, propaganda, advertising, fake reviews, fake everything. Fabricated email from friends family and colleagues. Whole online communities fabricated out of whole cloth. It is a weapon, and a powerful one.

[repolfx]

No, it's useless and I speak from experience of dealing with spammers who forged mail from friends family and colleagues in the past.

People are not trivial automatons who can have their opinions rewritten on the fly by auto-generated text. If auto-generated text reaches into its giant grab-bag of learned expressions and produces something actually interesting or insightful, people might be interested in that new line of thinking, but if - like many of these examples - it's essentially rambling if coherent nonsense then it won't have any impact at all.

So I rather think it's you fooling yourself. You've been reading comments online for years without knowing who or what produced them. If you discovered half of them were artificial tomorrow, what difference would it make? The people around you are already judging arguments based on the content, not their volume or who wrote them.