hhmx.de

Democracy Matters :verified:

Föderation EN So 16.06.2024 17:17:08

I love it. An academic paper with a straightforward title.

ChatGPT is bullshit
link.springer.com/article/10.1

Mustardon ~ IWRY 🦅

Föderation EN So 16.06.2024 17:41:59

@DemocracyMattersALot
LLMs are essentially the latest version of ELIZA with a huge database.

Bongolian

Föderation EN So 16.06.2024 19:21:38

@bitsnpieces @DemocracyMattersALot ELIZA also had the further quality of showing the shallowness of Rogerian psychology.

David in Tokyo

Föderation EN Mo 17.06.2024 02:30:45

@bitsnpieces @DemocracyMattersALot

LLMs aren't like ELIZA at all. ELIZA at least tried to do what it was trying to do, i.e. be a Rogerian shrink.

LLMs simply randomly generate text.

SteveO ✅ - For Harris 🇺🇸

Föderation EN So 16.06.2024 19:07:25

@DemocracyMattersALot It seems to me that AI should be regulated by government. AI large language models should be required to be trained and certified on basic morality.

Bongolian

Föderation EN So 16.06.2024 19:26:39

@steveo1049 @DemocracyMattersALot It would be difficult to come to consensus on what morality is, but we can't even get our politicians and judges to be certified on morality.

nicholas_saunders

Föderation EN Mo 17.06.2024 07:39:41

@steveo1049 @DemocracyMattersALot the devil's in the details on that.

And, there are benefits.

Thanasis Kinias

Föderation EN So 16.06.2024 19:47:46

@DemocracyMattersALot
and now I have ‘bullshit’ as a tag in Zotero

Beni Borukhov

Föderation EN So 16.06.2024 21:07:55

@DemocracyMattersALot Ah glad to see the paper that makes me proud to have studied linguistics cited.

Ek Ong Kaar K Michaud

Föderation EN So 16.06.2024 22:40:59

@DemocracyMattersALot Fantastic article. Thank you.

SpaceLifeForm

Föderation EN Mo 17.06.2024 02:41:11

@DemocracyMattersALot

ChatGPT certainly knows how to spew bullshit.

Timo Grün

Föderation EN Mo 17.06.2024 11:07:03

@DemocracyMattersALot
In high school I got an A- on a test on a book I hadn’t read. An intelligent teacher would have seen through it and given me an F. That’s how bullshit works. LLMs use different methods to generate the bullshit, but it’s the same bullshit, and there are too many people who can’t tell the difference or don’t care.