AI, power and sociolinguistics (2024) | Hacker News Reader

AI, power and sociolinguistics (2024)

researchgate.net

Permanent link: https://doi.org/10.1111/josl.12681

I recently read the thought-provoking article “AI, power and sociolinguistics” and decided to engage with it in an unusual way: by letting an AI (DeepSeek-V3) craft a detailed response on my blog. Surprisingly, it didn’t just summarize the points—it argued back.

You’ll find the piece here: https://deep.liveblog365.com/en/index-en.html?post=30

I think sociolinguists will have to do better than just chatting with one bot[1] before waxing poetic about the invisible power dynamics that shaped the interaction. This is critical theory: thought-provoking cynical fart-sniffing. Autoethnography “research”.

[1] no clear sign of model chosen either. Who knows what software version or system prompt they were blabbing to.

"hallucinations" and "bullshiting" is a required part of being human for the simple reason the 3 inch chimp brain we have, can't fit all the "knowledge" out there.

Even today if you tell your local chimp chief - listen dude you just don't know what you are talking about, guess what happens next?

If you can't do that without risk (and the risks can jump based on subject) then guess what the default behavior of the chimp is to handle that reality - hallucinations and bullshitting.

That's an agi back handed complement by any measure. Well, by way of analogy, we made peace out of eden and so will agi.

Lying by the way is usually caught up in various ways to defray the threat to ego (self esteem) and defray or delay natural/social/legal consequences. The latter is caught up in body.

I'm not sure I would give llms the benefit of the doubt to have obtained those faculties.

I would put it differently.

> can't fit all the "knowledge" out there

know-that is less important than know-how in the longer run. knowledge pertains to the know-how (or theory) than about know-that (facts)

> default behavior of the chimp is to handle that reality - hallucinations and bullshitting

default behavior of bullshitting is not perhaps as much a status or hierarchy thing as it serves a role in communicating human meaning which could be contradictory and not logical in nature. What is logical is that meaning is that which has been organized by a know-how or theory - but then, due to its formalization tends to be stale. Imagine if all of us only made logically correct statements - we'd be robots in a stale world.

> can't fit all the "knowledge" out there

We don't know that. Given how environmental factors fuck up plenty of neutral layers and synaptic graphs, there is reason to believe that if we stopped or reversed that, human brains might be able to fit all the knowledge out there, given enough time. The constraint then would be preference. Because let's be honest, you wouldn't give a fuck about the Kardashians and the history of their influence on markets and teen desires in America.

Reply on news.ycombinator.com