Hacker News Reader

Comment by mrits | Hacker News Reader

How is it much different that wikipedia or just about any other source of information?

GPT answers are dynamic not constant like wikipedia, scientific articles, information records. It's also individualized, thus others can't help spot it is wrong. Answers are subject to be filtered or manipulated according to a hidden set of rules by unknown corporate or government agendas that make it hide or provide incorrect and misleading information, for example, in antitrust investigations, internal governmental investigations, etc...

Agreed in genera but the volatile edit history of many Wikipedia articles would bely your definition of constant.

Also scientific articles being constant don't quite matter when the "status quo" of science is asymmetrically understood.

There are doctors today still pushing the "fat is unhealthy" myth because of the sugar industry. Despite the ubiquity of modern unbiased research papers and proof.

But when you ask a doctor for advice, you're essentially getting the wetware equivalent of a stochastic parrot, because this random doctor is giving you dynamic answers coming from his/her own completely non-standardized corpus of knowledge.

To be fair, at least Wikipedia offers a traceable history of those edits. Most, if not all sufficiently large LLMs, are effectively opaque boxes that even their creator can rarely explain how a specific answer was reached, much less how the billions of parameters changed in the network over time.

You can also peg a known good version of Wikipedia and host it locally. That said I can't imagine trying to audit the whole shebang.

"Answers are subject to be filtered or manipulated according to a hidden set of rules by unknown corporate or government agendas"

When was the last time you tried to make a wiki edit on a popular page?

The tedious parts are automated.

Reply on news.ycombinator.com