| Website | https://henryfarrell.net/wp/ |
| ORCID | https://orcid.org/0000-0003-0611-3949 |
@henryfarrell lays out a compelling metaphor where Google AI is the 2020s' equivalent to Olestra.
https://www.programmablemutter.com/p/google-ai-fails-the-taste-test
Good point from @henryfarrell : AI summarizations make bad search results *feel* much more as if they're Google's fault, even if the underlying way in which they're obtained is not vastly different than before.
https://www.programmablemutter.com/p/google-ai-fails-the-taste-test
A few weeks ago, I wrote about a paper on online toxicity by @henryfarrell & Cosma Shalizi. Henry had the wonderful generosity to respond and point out where I misunderstood the article.
In this post, I respond with clarifications on the idea of toxicity and the use of models.
I also reflect on the value of having a thoughtful, considered disagreement in public online — something that seems to have largely disappeared from my circles.
https://natematias.medium.com/disagreements-fast-and-slow-d0bc49ac9c3f