Email or username:

Password:

Forgot your password?
Leonard's posts Post Back to profile
Top-level
Zimmie

@lritter @mcc By using words like “political” or “woke” or “diversity”, they are trying to make something *shaped like* a real argument, but devoid of substance. They don’t have to explain it. Their people all only care about the shape of it, not about what is actually being said.

Incidentally, I wonder if the devolution of political speech into insubstantial, argument-shaped blurbs primed people to see LLMs as actually smart.

3 July at 18:51 | Wall-to-wall | Open on infosec.exchange
5 comments
mcc

@bob_zim @lritter "Machines did not pass the turing test, humanity failed it"

3 July at 18:51 | Open on mastodon.social
Leonard Ritter

@bob_zim @mcc pop fic did. LLMs only had three years. people watch too much TV. stories lie. not everybody gets that.

3 July at 18:54 | Open on mastodon.gamedev.place
Zimmie

@lritter @mcc I more mean the conflation of what fiction says an AI does (think and produce an answer) with what an LLM actually does (produce an answer-shaped blurb). Sure, the marketing also primes people to believe ChatGPT is a thinking machine, but a shocking number of people who have actually used it and seen the ways in which it is typically wrong still seem to think it has any intelligence at all. They say it “hallucinated” citations in a legal brief. No, it produced a legal-brief-shaped blob of text. Briefs have citations. The citations may happen to align with reality or they may not, but they’re not hallucinated.

Words are textually meaningless to LLMs, like they are to right-wing politicians. Only the associations between the words matter.

@lritter @mcc I more mean the conflation of what fiction says an AI does (think and produce an answer) with what an LLM actually does (produce an answer-shaped blurb). Sure, the marketing also primes people to believe ChatGPT is a thinking machine, but a shocking number of people who have actually used it and seen the ways in which it is typically wrong still seem to think it has any intelligence at all. They say it “hallucinated” citations in a legal brief. No, it produced a legal-brief-shaped blob...

3 July at 19:33 | Open on infosec.exchange
Zimmie

@lritter @mcc The real advantage to making things shaped like arguments but void of textual meaning is they can’t really be countered. Any response which goes to the text completely misses the meaning, and any response to the meaning is “putting words in my mouth.”

In the US, this is the natural next step in the Southern Strategy. Its architect, Lee Atwater, was possibly the most maliciously evil person in American politics. This strategy saw US conservatives leaning to use language more effectively to appeal to overt racists while not sounding overtly racist themselves. Towards the end of his life, someone actually recorded him describing it. As a warning, he uses slurs:

youtu.be/X_8E3ENrKrQ

@lritter @mcc The real advantage to making things shaped like arguments but void of textual meaning is they can’t really be countered. Any response which goes to the text completely misses the meaning, and any response to the meaning is “putting words in my mouth.”

In the US, this is the natural next step in the Southern Strategy. Its architect, Lee Atwater, was possibly the most maliciously evil person in American politics. This strategy saw US conservatives leaning to use language more effectively...

3 July at 19:50 | Open on infosec.exchange
Go Up