[tivert]

Except that's not a social network. If it was open about what it was, its nature would undermine the main reason for using a social network.

Probably the only people who'd bother to engage are some geeks who've confused themselves into thinking a social network is just about generating messages for consumption.

Also, such a fake social network might be able to replicate BS political slapfights, but I'd think it would fall on its face when touching on anything that's not repetitive or baroquely self-referential. It'd get old fast.

[midwesternerer]

Yeah, you aren't wrong, but how do you know that you are talking to an actual human right now and not a GPT powered hackernews bot? Like you are assuming I'm a real person, but you don't know it. You can't ever really know it.

[tivert]

> Yeah, you aren't wrong, but how do you know that you are talking to an actual human right now and not a GPT powered hackernews bot?

Because it's 2023, and GPT, while impressive in its way, still sucks. Even if you generated the text with GPT, you're still operating it and I'm interacting with you.

> Like you are assuming I'm a real person, but you don't know it. You can't ever really know it.

That's a little sophomoric. No one needs to know with perfect clarity. If bots ever get good enough and prevalent enough most social media users are fake, there will be a little lag as people figure it out. Then people who keep using social media will get mocked for wasting their time, and it will die.

[midwesternerer]

I'm just hoping that the robots dont hand wave away my arguments by making fun of them as "sophomoric".

[tivert]

> I'm just hoping that the robots dont hand wave away my arguments by making fun of them as "sophomoric".

I didn't hand-wave it away, I just labeled it correctly and addressed it.

A mind-blowing trump card of an "argument" it wasn't.

[midwesternerer]

Dude, your argument against it was 1) that a technology that has been public for six months isn't perfect 2) that smart people would just know that comments were not being made by humans.

I truly don't understand how you are interpreting this conversation.

[tivert]

> I truly don't understand how you are interpreting this conversation.

That makes sense, because you're pretty badly misunderstanding me (which your summary makes perfectly clear).

I mean, 1 was part of a direct answer to your own question "how do you know that you are talking to an actual human right now and not a GPT powered hackernews bot." Emphasis mine. Your 2 is just a total misunderstanding. Hint: I was talking about social knowledge, not "smart people just knowing" in any particular case.

[zamnos]

You should worry about everything you interact with, not just text on some small Internet forum. After all, you could be the star of your own reality-tv Truman Show, and everything you see is made up for the benefit of a TV show that you're the star of, but aren't privity to? How do you know your mom and dad aren't just hired actors along with all the people you interact with? Maybe I'm secretly being paid by producers of a TV show to write this comment to provoke you into discovering the truth of your existence for the season finale?

[midwesternerer]

I think the difference is that the cost of faking my family is extremely high, but the cost of spewing comments on the internet went to basically 0.

I think it is pretty widely accepted that corporations and governments already conduct astroturfing, but those operations were limited in that you had to open an office somewhere and pay people hourly to write propaganda. Now you only need an API key.

I'd be really surprised if there are not conversations happening right now at Facebook and its competitors about injecting generated comments into people's feeds to increase engagement.