There's a standard gag about LLM chatbots: whenever I use AI to probe a topic I know something about, it makes numerous errors; by contrast, whenever I use it to explore topics I know little about, it knows so much more!
-
@peturdainn @thirstybear @kevlin "Gell-Mann amnesia"
@aes @peturdainn @kevlin That's the one! Thank you!
https://en.wikipedia.org/wiki/Michael_Crichton#Gell-Mann_amnesia_effect
[Edit: the link's sound and directs to an internal anchor in the article. The thumbnail reflects that the term was invented by Michael Crichton]
-
There's a standard gag about LLM chatbots: whenever I use AI to probe a topic I know something about, it makes numerous errors; by contrast, whenever I use it to explore topics I know little about, it knows so much more!
Unrelated: while workers report or are measured to have at best modest improvements to their work when using AI, CEOs and managers say AI has many benefits to the work of their organisation.
@kevlin oh no, it's the same one as "Musk sounded smart until he started talking about things I know"
-
@aes @peturdainn @kevlin That's the one! Thank you!
https://en.wikipedia.org/wiki/Michael_Crichton#Gell-Mann_amnesia_effect
[Edit: the link's sound and directs to an internal anchor in the article. The thumbnail reflects that the term was invented by Michael Crichton]
-
There's a standard gag about LLM chatbots: whenever I use AI to probe a topic I know something about, it makes numerous errors; by contrast, whenever I use it to explore topics I know little about, it knows so much more!
Unrelated: while workers report or are measured to have at best modest improvements to their work when using AI, CEOs and managers say AI has many benefits to the work of their organisation.
@kevlin LLMs deliver one concrete benefit to corporations: they consistenty drive down employee wage settlements (or rather, management use them as a tool to terrorise workers with the threat of unemployment).
-
J jonassmith@theforkiverse.com shared this topic
-
@kevlin LLMs deliver one concrete benefit to corporations: they consistenty drive down employee wage settlements (or rather, management use them as a tool to terrorise workers with the threat of unemployment).
@cstross Indeed
-
There's a standard gag about LLM chatbots: whenever I use AI to probe a topic I know something about, it makes numerous errors; by contrast, whenever I use it to explore topics I know little about, it knows so much more!
Unrelated: while workers report or are measured to have at best modest improvements to their work when using AI, CEOs and managers say AI has many benefits to the work of their organisation.
@kevlin I always found exactly the same about 'Which? Magazine'. It's easy to convince the ignorant, if you state your conclusions with sufficient confidence, and we all have areas of ignorance.
-
@kevlin I always found exactly the same about 'Which? Magazine'. It's easy to convince the ignorant, if you state your conclusions with sufficient confidence, and we all have areas of ignorance.
@woo We are always working with incomplete knowledge and are more susceptible to the persuasive force of confidence than we believe
-
@kevlin oh no, it's the same one as "Musk sounded smart until he started talking about things I know"
@kevlin @patterfloof
The Eberhard / Tarpenning Factor. -
There's a standard gag about LLM chatbots: whenever I use AI to probe a topic I know something about, it makes numerous errors; by contrast, whenever I use it to explore topics I know little about, it knows so much more!
Unrelated: while workers report or are measured to have at best modest improvements to their work when using AI, CEOs and managers say AI has many benefits to the work of their organisation.
@kevlin So... the conclusion is that we just need to eliminate all experts, and then the LLMs will achieve new heights of ... well, new heights, anyway.
-
@woo We are always working with incomplete knowledge and are more susceptible to the persuasive force of confidence than we believe
@kevlin I've grown to be deeply distrustful of very confident people because competent people have doubts.
-
@kevlin So... the conclusion is that we just need to eliminate all experts, and then the LLMs will achieve new heights of ... well, new heights, anyway.
@GerardThornley Or depths. Same axis, different sign.
-
J jwcph@helvede.net shared this topic
