These last few months, I have begun to hate using the internet more and more. The reason is AI. Well, that may be too broad. The reason is chatbots in particular. In contrast to generative AI, which has gotten better and better, the chatbots have gotten worse and worse. Not only are they worse, the downstream effect of their adoption has been the total destruction of Google. (I’ve been banging this drum for a while now.)
When I’m working and need to do fast fact checks like “What year did X happen?” or “What is the title of person Y?” I do not get routed to fast answers in milliseconds. Instead I wait for Google or Claude or Chat-GPT to tinker for a few seconds and sometimes more than a few seconds. Then I need to scan read 500-800 words of superfluous text.
With more complicated questions like tracking down a specific study or essay or figure or quote, I often have to ask further questions because the chatbot has summarized a quote or essay rather than given me actual quotes. Oftentimes, I have to ask it for the sources themselves so I can go to the webpage and read what was actually written or said. Sometimes I have to double back again because the link has no reference to what the chatbot summarized.
The chatbot will offer a milquetoast explanation for why it is constantly wrong, constantly hallucinating, and impediment to my work rather than a help.