Re:"AI-powered" has become a red flag
by oidar on Monday June 10, 2024 @01:05AM
(#40630342)
There's some things I'd love an AI powered service for - for example no one has really gotten AI powered email right yet. But I use and love a few "AI" services too. Kagi's on a good path with summarization of results and quick answers - I use that AI service everyday. Claude is pretty darn good at inventing recipes from a list of ingredients and appliances that I have available. But to be fair to the blogger, I don't care that it uses LLMs - I just like that it solves a problem for me.
[ Parent
| on HN ]
AI summaries are an absolute disaster for the web. It's often inaccurate, it relies on stolen content, and it puts search engine operators in the position of content publishers which is right in the crosshair of laws like section 240. Complete nonsense.
Independent of section 240 or other laws, I would argue that search engine operators are content publishers as soon as we don't know how they choose the order of the results. Same for social media feeds.
Curating is a form of publishing, and following an algorithm is not curating per se, but, if that algorithm is non-transparent, and unknown to the user, it is the same end result...
I agree that systems that are virtually public utilities at this point should be operated more transparently. Google can put whole swaths of legitimate businesses out of business with minor tweaks of the algorithm. We complain about SEO but how do we know Google is optimized for search quality vs their own financial gain?