158
ChatGPT has a style over substance trick that seems to dupe people into thinking it's smart, researchers found
(www.businessinsider.com)
This is a most excellent place for technology news and articles.
When you underpay a bunch of gig workers to rate the outputs? Obviously it's going to write in a manner that best BS's a layperson.
Would be too expensive to hire experts in every field to train the AI to actually do good work. Imagine paying software engineers 100k plus benefits to vote on its code outputs, or getting Miss Manners to comment on its etiquette suggestions.